pubs.acs.org/Langmuir © 2010 American Chemical Society
Hydrogen-Bond-Directed Giant Vesicles of Guanosine Derivatives in Water: Formation, Structure, and Stability Jun Sawayama, Isao Yoshikawa, and Koji Araki* Institute of Industrial Science, University of Tokyo, 4-6-1 Komaba, Meguro-ku, Tokyo 153 8505, Japan Received December 31, 2009. Revised Manuscript Received February 13, 2010 Hydrogen-bond-directed giant supramolecular vesicles (diameter 1.20 ( 0.30 μm (SD)) of an alkylsilylated deoxyguanosine derivative, 2a, were prepared faciley by mixing a small volume of a 2a/THF solution with water. The formation of 2-D inter-guanine hydrogen-bond networks of 2a within the vesicles was indicated by IR spectra. The vesicle solution was stable enough for more than 30 days, in a wide range of temperatures, and between pH 4 and 10 without showing lysis, fusion, precipitation, or leakage of the encapsulated fluorescent probe. In a typical micrometersized vesicle, a sufficiently large internal water phase for encapsulating water-soluble substances was surrounded by a multilamellar membrane 15-20 nm in thickness, which was composed of 6-9 layers of 2-D hydrogen-bond-directed sheet assemblies. AC-mode AFM observation of the vesicle on a silicon substrate further demonstrated the high stability and deformable properties of the vesicle membrane under vacuum or mechanical stress. The formation and properties of the vesicle membrane in water were analyzed from the viewpoint of the 2-D hydrogen-bond-directed sheet assemblies, and the scope of the design principle to use nonpolar soft segments as the shielding units of the hydrogen-bond networks in water is discussed.
Introduction Self-assembled supramolecular architectures on a nano- to micrometer scale have been the subject of intense studies in recent years.1 Since the self-assembling process is the interplay of various intermolecular interactions, precise design of the building blocks is required in order to fabricate well-defined architectures. In this context, multiple hydrogen bonds are useful attraction forces for designing an architecture in a predictable way because of their high stability and directionality.2 However, close packing is another strongly demanding factor in molecular aggregation, and a small difference in molecular shapes greatly affects the mode of molecular packing.3 We showed that the use of adjustable soft segments is an effective tool for circumventing this problem, and supramolecular flexible fibers and films of alkylsilylated nucleoside derivatives were successfully fabricated from the tape motifs of a one-dimensional (1-D) hydrogen-bond sequence and two-dimensional (2-D) hydrogen-bond-directed sheet assemblies, respectively, by the rational design of the soft segments.4 In a previous communication,5 we demonstrated that fine-tuning of the molecular design allowed for fabrication of hydrogen-bond-directed stable giant vesicles in water (Figure 1). Though hydrogen-bond interaction does not operate effectively *Corresponding author. E-mail:
[email protected]. (1) (a) Palmans, A. R. A.; Meijer, E. W. Angew. Chem., Int. Ed. 2007, 46, 8948– 8968. (b) Ariga, K.; Hill, J. P.; Lee, M. V.; Vinu, A.; Charvet, R.; Acharya, S. Sci. Technol. Adv. Mater. 2008, 9, 1–96. (c) Ryu, J.-H.; Hong, D.-J.; Lee, M. Chem. Commun. 2008, 9, 1043–1054. (d) Reinhoudt, D. N. Supramolecular Materials and Technologies; Wiley: Chichester, 1999. (2) (a) Lena, S.; Masiero, S.; Pieraccini, S.; Spada, G. P. Chem.;Eur. J. 2009, 15, 7792–7806. (b) Koblenz, T. S.; Wassenaar, J.; Reek, J. N. H. Chem. Soc. Rev. 2008, 37, 247–262. (c) Brinke, G. T.; Ruokolainen, J.; Ikkala, O. Adv. Polym. Sci. 2007, 207, 113–177. (3) Kitaigorodskii, A. I. Organic Chemical Crystallography; Consultants Bureau: New York, 1961. (4) (a) Araki, K.; Takasawa, R.; Yoshikawa, I. Chem. Commun. 2001, 18, 1826– 1827. (b) Yoshikawa, I.; Li, J.; Sakata, Y.; Araki, K. Angew. Chem., Int. Ed. 2004, 43, 100–103. (c) Araki, K.; Yoshikawa, I. Top. Curr. Chem. 2005, 256, 133–165. (5) Yoshikawa, I.; Sawayama, J.; Araki, K. Angew. Chem., Int. Ed. 2008, 47, 1038–1041.
8030 DOI: 10.1021/la904916d
in water, 2-D hydrogen-bond-directed sheet assemblies were successfully fabricated in water by fine-tuning of the molecular design, i.e., the increased shielding effect of the nonpolar soft segments on the hydrogen-bond networks and proper control of the hydrophilicity of the surfaces of the sheet assemblies by oxyethylene units (2a and 2b in Figure 1). Giant vesicles have attracted much attention from a fundamental perspective7 as well as for their potential applications in encapsulating an enzyme or gene into a large internal water phase.8 Though the process to fabricate giant vesicles of lipids and amphiphiles has been established, the relatively poor stability of their structures is a major drawback for practical applications. Polymerization of the side chains, association with polymeric supports,9 or the introduction of additional interlipid interactions is the main way to stabilize the giant lipid vesicles.10 Recently, amphiphilic block copolymer systems that form stable giant vesicles were reported and have been studied actively.11,12 Though the use of the sheet assemblies composed of a 2-D hydrogen-bond network seems to be an attractive approach to this end, little is known about giant vesicles having hydrogen-bond-directed membranes except for our preliminary report on the 2b vesicles.5 In our course of study on the giant vesicles of alkylsilylated guanosine (6) Takasawa, R.; Yoshikawa, I.; Araki, K. Org. Biomol. Chem. 2004, 2, 1125– 1132. (7) Jesorka, A.; Orwar, O. Annu. Rev. Anal. Chem. 2008, 1, 801–832. (8) (a) Walde, P.; Ichikawa, S. Biomol. Eng. 2001, 18, 143–177. (b) Vriezema, D. M.; Aragones; Elemans, M. C. J. A. A.; Cornelissen, W. J. J. L. M.; Rowan, A. E.; Nolte, R. J. M. Chem. Rev. 2005, 105, 1445–1490. (c) Kasuya, T.; Jung, J.; Kinoshita, R.; Goh, Y.; Matsuzaki, T.; Iijima, M.; Yoshimoto, N.; Tanizawa, K.; Kuroda, S. Methods Enzymol. 2009, 464, 147–166. (9) (a) Ringsdorf, H.; Schlarb, B.; Venzmer, J. Angew. Chem., Int. Ed. Engl. 1988, 27, 113–158. (b) Mueller, A.; O0 Brien, D. F. Chem. Rev. 2002, 102, 727–757. (10) Antonietti, M.; F€orster, S. Adv. Mater. 2003, 15, 1323–1333. (11) (a) Discher, B. M.; Won, Y.-Y.; Ege, D. S.; Lee, J. C.-M.; Bates, F. S.; Discher, D. E.; Hammer, D. A. Science 1999, 284, 1143–1146. (b) Discher, D. E.; Eisenberg, A. Science 2002, 297, 967–973. (12) (a) Malinova, V.; Malinova, V.; Belegrinou, S.; Ouboter, D. B.; Meier, W. P. Adv. Polym. Sci. 2010, 224, 113–165. (b) Dongen, S. F. M.; Hoog, H.-P. M.; Peters, R. J. R. W.; Nallani, M.; Nolte, R. J. M.; Hest, J. C. M. Chem. Rev. 2009, 109, 6212–6274. (c) Du, J.; O'Reilly, R. K. Soft Matter 2009, 5, 3544–3561.
Published on Web 03/01/2010
Langmuir 2010, 26(11), 8030–8035
Sawayama et al.
Article
Figure 1. (a) Structures of the alkylsilylated derivatives of guanosine, (b) their 2-D hydrogen-bonding pattern based on the crystal structure of 1a,6 and (c) the 2-D sheet assembly. Hydrogen bonds between the 1-D tape motifs are indicated by the broken red lines in (b).
derivatives, we found that the vesicles could be prepared faciley by the injection method when 2a instead of 2b was used, allowing us to conduct detailed and quantitative studies on the hydrogenbond-directed vesicles. In this article, we will report full details and deeper insight of the hydrogen-bond-directed giant vesicles.
Results Supramolecular Films. Supramolecular films of 2a and 2b were obtained by casting the 2a or 2b/THF solution (5 wt %) on a glass, Teflon, or calcium fluoride plate.5 The resultant solventcast films were translucent and flexible. Differential scanning calorimetric (DSC) measurements of the 2a and 2b films showed endothermic melting peaks at 99 and 66 °C, respectively, and no other peaks were observed prior to their melting temperatures. The hydrogen-bonding patterns in the cast films and the viscous liquid of 2a and 2b on the calcium fluoride plates were studied by temperature-controlled IR spectrometry (Figure 2). The IR spectra of the 2a and 2b films at room temperature showed lowershifted νCdO and νN-H bands at around 1695 and 3140-3330 cm-1, respectively, but no free νN-H band at around 3500 cm-1 was observed (Figure 2a,b).13 In the case of the 1-D tape motif, a νN-H band at around 3500 cm-1 was observed because only one of the amino hydrogens participated in the hydrogen-bond sequence (Figure 2g).6 Since additional intertape hydrogen bonds of free amino hydrogen leads to the formation of the 2-D hydrogen-bond network (Figure 1b), the lower-shifted νCdO band and absence of a free νN-H band are the typical features of 2-D hydrogen-bonded guanosine derivatives (Figure 2f).4 Therefore, the spectra of the films shown in Figure 2 suggested the formation of hydrogen-bond-directed 2-D sheet assemblies in the films. X-ray diffraction patterns of the 2a and 2b films presented in a previous communication (Figure 2 in ref 5) showed a sharp peak at 3.80° (2.32 nm) corresponding to the thickness of 2-D sheet assembly, confirming that the films have a lamellar-like stacked structure of the 2-D sheet assemblies. Above their melting temperatures, they became isotropic viscous liquids. A weak band due to a free νN-H band appeared at around 3500 cm-1, and the νCdO band at 1695 cm-1 became slightly broader (Figure 2c,d). Therefore, partial disruption of the hydrogen-bonding network took place above their melting temperatures though most of the guanine units were in the hydrogen-bonded state.
(13) Szczepaniak, K.; Szczepaniak, M. J. Mol. Struct. 1987, 156, 29–42.
Langmuir 2010, 26(11), 8030–8035
Figure 2. IR spectra of the cast films of (a) 2a and (b) 2b at room temperature, (c) 2a at 120 °C, (d) 2b at 80 °C, and (e) the ruptured vesicles of 2a. IR spectra of the crystals of (f) 1a having the 2-D sheet assembly4 and (g) 1b having the 1-D hydrogen-bonded tape motif6 are shown as references in order to facilitate the comparison.
Figure 3. Microscopic images of the vesicle solutions of (a) 2a prepared by the injection method and (b) 2b prepared by the thinfilm method and (c) optical (upper) and fluorescence (bottom) microscopic images of the eosin Y-entrapped 2a vesicle solution. Note that the optical and fluorescence microscopic images taken in the solution were not for the same vesicles.
Preparation of Vesicle Solutions. As reported previously, a translucent aqueous solution of micrometer-size 2b vesicles was prepared by the thin-film method.5 In this preparation, water was added to a thin film of 2b at the bottom of a test tube and subjected to sonication (see Experimental Section for detailed conditions). Though sonication at room temperature did not cause full dispersion of 2b in water, full and homogeneous dispersion of 2b was achieved by sonication at 80 °C for 5 min (Figure 3b). The concentration of 2b in water was in the range of 10-5 mol dm-3, but it did not easily disperse homogeneously DOI: 10.1021/la904916d
8031
Article
Sawayama et al. Table 1. Aqueous Vesicle Solutions of 2a Prepared by the Injection Methoda entry
1
2
3
4
5
6
-6
amounts of 2a/10 mol 0.25 1.00 2.00 2.50 3.00 5.00 added to the solutionb dispersed in solution 0.25 1.01 2.00 2.45 2.94 2.60 after filtration 0.16 0.12 0.10 0.05 0.06 0.10 vesiclesc diameter/μm 0.81 ( 0.10 1.20 ( 0.30 1.28 ( 0.34 1.20 ( 0.36 1.64 ( 0.64 0.14 ( 0.07 15 ( 3.5 40 ( 6.8 50 ( 7.5 51 ( 7.2 54 ( 4.2 number/1010 dm3 a 2a in 0.1 mL of THF was mixed with 10 mL of pure water and the amount of 2a dispersed in solution was determined after standing for a day. b Amounts of 2a in 0.1 mL of THF solution. c Determined from the optical microscopic image.
when the amount of the 2b film was increased in order to prepare a higher concentration of 2b dispersion. In the case of 2a, a similar procedure at room temperature or 80 °C failed to disperse the 2a film homogeneously in water. Since the melting temperatures of the 2a and 2b films are 99 and 66 °C, respectively, sonication at temperatures sufficiently higher than their melting temperatures seems to be the key to inducing homogeneous dispersion of guanosine derivatives in water by sonication. Higher hydrophilicity of 2b might also contribute to successful vesicle preparation by the thin-film method. We found that the simple injection method was suitable to prepare the 2a dispersion. As described in the previous communication, the stable vesicle dispersion was obtained for 2b by the injection method at 80 °C and under ultrasonication, but the size of the vesicle was less than 1 μm and practically no giant vesicle formation was observed. However, we found that the stable giant vesicles could be prepared by the injection method at room temperature without ultrasonication under appropriate conditions. Typically, the mixing of 0.1 mL of a 2a/THF solution (2.50 10-6 mol or 2.50 10-2 mol dm-3) with 10 mL of water at room temperature successfully gave a translucent and homogeneous dispersion. The amount of 2a in the aqueous solution a day after preparation was determined spectrophotometrically, which confirmed that the amount of 2a added as a THF solution was fully dispersed in the solution. An optical microscopic image of the solution showed the presence of micrometer-sized spherical vesicles (Figure 3a), and the average diameter of the vesicles was determined from the image to be 1.20 ( 0.30 μm (SD). A fluorescence microscopic image of the eosin Y-entrapped vesicles, after removal of external eosin Y by Sephadex G-25, indicated the presence of an internal water phase, further confirming the formation of the vesicular structure (Figure 3c). Filtration through a PTFE membrane filter having an average pore size of 1 μm indicated that small 2a particles or aggregates that passed through the filter were only 5 mol % of the total 2a in the solution. Therefore, most of 2a was present as micrometer-size vesicles in solution. As shown in Table 1, the amounts of 2a in THF (0.1 mL) affected the preparation of the vesicle solution (10 mL). Up to 3.00 10-6 mol of 2a in 0.1 mL of THF was fully dispersed in water. A further increase of 2a caused a precipitate formation, while a decrease of 2a below 1.00 10-6 mol increased the number of small particles of 2a that passed through the membrane filter. The results indicated that the use of a THF solution containing (2-3) 10-6 mol of 2a is the appropriate condition for preparation of the giant vesicles. The concentration of 2a in the resultant aqueous dispersion was in the range of 10-4 mol dm-3, which is an order of magnitude higher than that of 2b prepared by the thin-film method. An attempt to prepare the 2b dispersion by the injection method under various conditions resulted in the formation of precipitates, and no vesicle formation was observed at all. Thus, it was found that the 2a giant vesicle solution can be prepared faciley 8032 DOI: 10.1021/la904916d
Table 2. Long-Term Stability and Dispersibility of the Aqueous 2a Vesicle Solutions incubation period/day
1
7
30
a
100 (2) 98 (2) 98 (2) 2a in solution /% vesiclesb diameter/μm 1.24 ( 0.34 1.25 ( 0.30 1.24 ( 0.32 59 ( 5.2 58 ( 5.8 57 ( 5.5 number/1010 dm3 a Amounts of 2a dispersed in the solution relative to those added to the solution. Relative amounts of 2a in the solution after filtration through a PTFE membrane filter (average pore size 1 μm) are given in parentheses. b Determined from the optical microscopic image.
Figure 4. Optical (upper) and fluorescent (bottom) microscopic images of the eosin Y-entrapped 2a vesicle solution (a) after 30 days, (b) at pH 4, (c) at pH 10, (d) after heating at 100 °C for 3 h, and (e) after the slow freeze-thaw process. Each scale bar indicates 5 μm. Note that the optical and fluorescence microscopic images taken in the solution were not for the same vesicles.
by the simple injection method under the appropriate condition. Properties of Vesicle Solutions. To test the long-term stability of vesicle dispersion, a vesicle solution prepared under representative conditions (entry 4 in Table 1) was kept at room temperature for 7 and 30 days. As shown in Table 2, practically no change was observed in the amount of 2a or the number and size of the vesicles dispersed in the solution even after 30 days. Further, a fluorescence image of the solution clearly showed the green fluorescence of the encapsulated eosin Y of the vesicles even after 30 days, indicating that the integrity of the vesicle structure was retained (Figure 4a). The results confirmed that the 2a vesicles in water have sufficiently high stability and dispersibility. For testing the effect of pH on the vesicle solution, it was kept for a day after the external pH of the solution was adjusted to 3, 4, or 10 by the addition of a small amount of HCl or NaOH. When the pH of the solution was 3, a partial precipitate formation and concomitant decrease of 2a in the solution were observed. At pH 4 and 10, however, practically no change of the vesicle solution was observed at all (Table 3), and the green fluorescence of the encapsulated eosin Y was clearly observed (Figure 4b,c). Therefore, the vesicle solution was stable between pH 4 and 10. Langmuir 2010, 26(11), 8030–8035
Sawayama et al.
Article Table 3. Stability of the Aqueous 2a Vesicle Solutions under Various Conditions
conditions
100 °C
5 °C
pH 3
pH 4
pH 10
a
100 (5) 100 (6) 43 (3) 100 (3) 92 (4) 2a in solution /% vesiclesb diameter/μm 1.23 ( 0.25 1.22 ( 0.25 1.21 ( 0.32 1.22 ( 0.29 1.25 ( 0.33 58 ( 5.2 60 ( 5.3 28 ( 5.2 58 ( 5.8 55 ( 6.2 number/1010 dm3 a Amounts of 2a dispersed in the solution relative to those added to the solution. Relative amounts of 2a in the solution after filtration through a PTFE membrane filter (average pore size 1 μm) are given in parentheses. b Determined from the optical microscopic image.
The vesicle solution also showed high thermal stability. After the solution was heated at 100 °C for 3 h in a sealed tube (Figure 4d) or kept at 5 °C for overnight, the green fluorescence of the entrapped eosin Y and the microscopic image confirmed that the vesicles were intact. Therefore, the hydrogen-bonddirected supramolecular giant vesicles are sufficiently stable for a wide range of temperatures. In the case of the lipid membranes, their stability greatly decreases above their melting temperatures, and except for those found in thermoacidophilic Archaea, their melting temperatures seldom exceed 60 °C.14 In spite of the high stability of the hydrogen-bond-directed giant vesicles described above, they can be disrupted easily due to the reversible nature of the intermolecular hydrogen bonds. Upon the addition of a triple volume of ethanol, the translucent aqueous vesicle solution immediately turned to a clear transparent solution, and the vesicular structures in the optical microscopic image disappeared, indicating that ethanol induced a solution of 2a in which the hydrogen bonds were disrupted. Freezing of the vesicle solution revealed another interesting feature of the vesicles. When the vesicle solution was slowly frozen at -10 °C and thawed again at room temperature, a precipitate formation was noticed and the amount of 2a in the solution decreased considerably. Microscopic observation of the resultant solution showed the presence of micrometer-sized vesicle-like structures in spheroidal rather than spherical shape, but no fluorescence of the encapsulated eosin Y was observed at all in a fluorescence microscopic image (Figure 4e). These results suggested that a rupture of the vesicle took place by the slow freezing process, leading to leakage of the entrapped eosin Y in the internal water phase. In the case of rapid freezing at -196 °C and subsequent thawing, nearly half of the vesicles still retained the encapsulated eosin Y. Therefore, the rupture of the vesicle was not because of the low temperature but was due to the crystallization of the internal water as is the case in a living cell.15 It should be noted that the IR spectrum of the ruptured vesicles (Figure 2e) was identical to that of the solventcast film prepared from the THF solution (Figure 2a). The sample was prepared by deposition of the vesicle solution on a calcium fluoride plate after the slow freezing and subsequent removal of water in a vacuum. The results confirmed that the residual membranes of the ruptured vesicles were composed of 2-D hydrogen-bond-directed sheet assemblies. Further analysis of the ruptured vesicles by AFM is described in the next section. Vesicles on a Silicon Substrate. To analyze the properties of the vesicle further, a drop of the vesicle solution prepared under representative conditions (entry 4 in Table 1) was placed on a silicon substrate and dried in a vacuum. The eosin Y-encapsulated vesicle placed on the substrate showed green fluorescence even (14) (a) Komatsu, H.; Chong, P. L.-G. Biochemistry 1998, 37, 107–115. (b) Gliozzi, A.; Relini, A.; Chong, P. L.-G. J. Membr. Sci. 2002, 206, 131–147. (c) Benvegnu, T.; Brard, M.; Plusquellec, D. Curr. Opin. Colloid Interface Sci. 2004, 8, 469–479. (d) Jacquemet, A.; Barbeau, J.; Lemiegre, L.; Benvegnu, T. Biochimie 2009, 91, 711–717. (15) (a) Kanaseki, T.; Kawasaki, K.; Murata, M.; Ikeuchi, Y.; Ohnishi, S. J. Cell Biol. 1997, 137, 1041–1056. (b) Siminovitch, D.; Chapman, D. FEBS Lett. 1971, 16, 207–212 E. (c) van Winden, C. A.; Zhang, W.; Crommelin, D. J. A. Pharm. Res. 1997, 14, 1151–1160.
Langmuir 2010, 26(11), 8030–8035
Figure 5. The 2a vesicles on a silicon substrate after drying. Optical (upper) and fluorescent (bottom) microscopic images of the eosin Y-encapsulated vesicles (a) before and (b) after keeping additionally in a vacuum for 12 h, (c) 3-D AFM image of the vesicle, and AFM images (upper) and the height profiles (bottom) of the vesicle along the white dashed line (d) kept in a vacuum for 12 h, and (e) a vesicle observed by unidirectional scanning with an increased maximum force of 2.1 nN (the arrow indicates the scanning direction of the tip).
after removal of the external water in vacuo, indicating that the internal water phase was retained (Figure 5a). As shown in Figure 5c, AC-mode AFM observation showed slightly flattened vesicles on the substrate (height-to-diameter ratio was 0.49 ( 0.10 (SD)) and the average diameter was 1.26 ( 0.37 μm (SD). The size of the vesicle observed by AFM is in agreement with that in the aqueous solution estimated from the microscopic image. When the vesicle on the substrate was further kept in vacuo (2 10-3 Pa) for 12 h, it showed little change (Figure 5d). In the case of the eosin Y-encapsulated vesicle, the green fluorescence of eosin Y was observed after this vacuum treatment (Figure 5b). Therefore, the internal water phase was preserved even after this vacuum treatment, indicating the high resistance of the vesicle membrane against water permeation. When a AC-mode unidirectional scan was carried out with an increased maximum force of 2.1 nN,16 deformation and elongation of the vesicle along the scanning direction was observed without causing a rupture of the membrane even after repeated scans or for different vesicles (Figure 5e). Restoration of the original shape was confirmed by an AFM image scanned with a decreased force. Thus, the deformable properties of the vesicles were demonstrated, which suggests the high stability of the vesicle membrane under mechanical stress. As described in the previous section, the slow freeze-thaw process induced the rupture of the vesicles. The AFM image of the ruptured vesicle showed a much more flattened structure than the normal vesicles. A typical example is shown in Figure 6a, where the vesicle diameter is 1 μm but the height is only 34 nm, nearly (16) (a) Zhou, X.; Xu, H.; Fan, C.; Sun, J.; Zhang, Y.; Li, M.; Shen, W.; Hu, J. Chem. Lett. 2005, 34, 1488–1489. (b) Nnebe, I. M.; Schneider, J. W. Macromolecules 2006, 39, 3616–3621.
DOI: 10.1021/la904916d
8033
Article
Sawayama et al.
Figure 6. The 2a vesicles after the slow freeze-thaw process. (a) AFM image (upper) and (c) height profile (bottom) along the white dashed line of the vesicle on a silicon substrate and (b) height distribution measured by AFM.
1/20 of normal vesicles. The observed height distribution after the slow freeze-thaw process is shown in Figure 6b. Since leakage of the internal water phase was indicated by the fluorescence microscopic image after the slow freeze-thaw process (Figure 4e), the observed flattened structure is most likely the ruptured vesicle membranes without an internal water phase.
Discussion Shape and Structure of Vesicles. In this study, we showed that the hydrogen-bond-directed supramolecular vesicles of 2a were easily prepared by the injection method. Under the appropriate conditions, the amount of 2a added to water was almost completely dispersed in the solution, and more than 95 mol % of 2a was shown to be present as micrometer-size giant vesicles from the filtration experiment. The average diameter of the spherical vesicles was ∼1.2 μm (Table 1). The presence of an internal water phase was clearly indicated from the green fluorescence of the encapsulated eosin Y (Figure 3c), and the IR spectra of the ruptured vesicles confirmed the formation of the 2-D hydrogenbond network of 2a (Figure 2e). Therefore, in the micrometer-size vesicles, the internal water phase was surrounded and separated by the 2a membrane constructed by the 2-D hydrogen-bond networks. Since the unit area of 2a per molecule in the 2-D hydrogenbonded state and the bilayer thickness of the sheet assembly are shown, from the crystal structure of model compound 1a,17 to be 1.91 nm2 and the X-ray diffraction peak of the film to be 2.32 nm, the molecular volume of 2a in the hydrogen-bonded sheet assembly is 2.21 10-27 m3. Therefore, the relative volumes of the 2a membrane and the internal water phase in the vesicle can be roughly estimated to be 7 and 93%, respectively, based on the molecular volume of 2a, the amount of 2a in the solution, and the number and average volume of the vesicles. In the slow freeze-thaw experiment, the observed flat structures are most likely to be the ruptured membranes. Though the observed height shows relatively broader distribution, the peak distribution appeared at 30-40 nm (Figure 6b). Based on the assumption that the observed height represents the sum of the upper and bottom membranes of the vesicle without any remaining internal water phase after drying, the thickness of the membrane is presumed to be 15-20 nm for those in the peak (17) Sato, T.; Seko, M.; Takasawa, R.; Yoshikawa, I.; Araki, K. J. Mater. Chem. 2001, 11, 3018–3022.
8034 DOI: 10.1021/la904916d
Figure 7. IR spectra of (a) the 2a cast films and the THF solutions of 2a at (b) 50, (c) 25, (d) 10, and (e) 2.5 10-3 mol dm-3.
distribution, which corresponds to 6-9 layers of 2a sheet assemblies. From the thickness and diameter of the ruptured membrane, the internal water phase is estimated to comprise 84% of the total volume of the vesicle. Thus, two different estimations of the relative size of the internal water phase indicated that the major part of the vesicle is the internal water phase. This discussion allows for elucidation of the detailed structure of the vesicle as follows. The typical vesicle has a sufficiently large internal water phase for encapsulating various substances, which is surrounded by a multilamellar membrane 15-20 nm in thickness and composed of 6-9 layers of 2-D hydrogen-bond-directed sheet assemblies. Process of the Vesicle Formation. Under the appropriate conditions, the micrometer-size giant vesicles are easily prepared by mixing a small volume of a 2a/THF solution with water. However, a low concentration of 2a in THF resulted in the formation of small-size vesicles, while a high concentration caused precipitation. To understand the detailed process of vesicle formation, the IR spectra of 2a in THF solutions were measured (Figure 7). When the 2a concentration was 2.50 10-2 mol dm-3, the condition corresponding to entry 4 in Table 1, the spectra showed not only hydrogen-bonded νN-H at 3140 and 3330 cm-1 and νCdO at 1695 cm-1 but also free ν N-H at 3495 and 3576 cm-1 and free νCdO at 1730 cm-1. Therefore, the formation of a hydrogen-bonded assembly in THF prior to mixing with water was indicated (Figure 7c). However, the presence of free νN-H at 3495 and 3576 cm-1 and free νCdO at 1730 cm-1 was clearly observed in the case of 2a concentrations below the suitable concentration range for preparation of the giant vesicles, i.e., 0.25 or 1.00 10-2 mol dm-3 (Figure 7d,e).13 The results suggest that preorganization of 2a by the inter-guanine hydrogen bonds is the key factor for successful fabrication of hydrogen-bond-directed giant vesicles. Indeed, the use of protic ethanol that suppresses preorganization instead of THF resulted in the immediate formation of powdery precipitates by mixing with water, and no sign of Langmuir 2010, 26(11), 8030–8035
Sawayama et al.
Article
vesicle formation was indicated at all. At a higher concentration in THF (Figure 7b), 2a was in a fully hydrogen-bonded state. Excess formation of hydrogen-bond assemblies prior to mixing might not be a feasible condition for the preparation of the giant vesicles. Thus, preformation to an appropriate extent of the hydrogen-bond assemblies in THF is indicated to be essential for successful fabrication of micrometer-size supramolecular vesicles by the injection method. Since the 2-D hydrogen-bond network is sandwiched by hydrophobic shielding units in the sheet assembly, preformation of the assembly in THF prevents the disruption of the 2-D hydrogen-bond networks even after mixing with water. Thus, the design principle using the hydrophobic shielding unit is shown to be effective for fabrication of the hydrogen-bond-directed supramolecular assembly even in highly polar aqueous media. Stability of the Vesicles. The results in Table 2 show that the vesicle solution is stable enough for a sufficiently long period without lysis, fusion, or precipitation. The hydrophilic surface covered by oxyethylene units and the high stability of the hydrogen-bond-supported membrane may be the reason for the observed long-term stability of the vesicle solution. The vesicle also shows high stability in a wide range of temperatures and between pH 4 and 10. As evidenced by the temperature-controlled IR study of the cast film, the high thermal stability of the 2a hydrogen-bonding interaction might contribute to the high thermal stability of the 2a membrane. The functional group dissociating or associating in this pH range is guanosine 1-NH, whose pKa is 9.42.18 However, the guanine base unit is shielded by the hydrophobic units inside of the sheet assembly, and it is likely that the hydrogen-bonded guanine unit is insensitive to the external pH. Throughout these stability experiments, no leakage of the encapsulated eosin Y was observed at all, demonstrating the high ability to preserve substances encapsulated in the internal water phase. AC-mode AFM observation of the vesicle on the silicon substrate with increased maximum force further demonstrated the high stability and deformable properties of the vesicle membrane under mechanical stress. It is well recognized that the high deformability of erythrocytes under sheer stress is essential in blood circulation, and the presence of cytoskeletal protein networks attributes to maintain the stability and integrity of the cell membrane.19 In the case of the 2a vesicle membrane, the high stability of the 2-D hydrogen-bond network seems to contribute to its high stability and deformability. No apparent leakage of the internal water or entrapped eosin Y inside of the vesicle was observed even under vacuum, confirming the high resistance of the vesicle membrane against water permeation.
Conclusion Except for several cases, the hydrogen-bond interaction is not effective in water.20 In this study, we confirmed that the membrane of the 2a micrometer-sized vesicles in water consists of the 2-D interbase hydrogen bonds. Structure and stability of the vesicles in water were clarified more quantitatively, and the unique properties of the 2a membrane;high thermal stability, deformability and stability under mechanical stress, and high resistance to permeation of water and other substances;are demonstrated further. However, it is also indicated that the preformation to an appropriate extent of the hydrogen-bond (18) Saenger, W. Principles of Nucleic Acid Structure; Springer-Verlag: New York, 1984. (19) Safran, S. A.; Gov, N.; Nicolas, A.; Schwarz, U. S.; Tlusty, T. Physica A 2005, 352, 171–201. (20) (a) Onda, M.; Yoshihara, K.; Koyano, H.; Ariga, K.; Kunitake, T. J. Am. Chem. Soc. 1996, 118, 8524–8530. (b) Ariga, K.; Kunitake, T. Acc. Chem. Res. 1998, 31, 371–378.
Langmuir 2010, 26(11), 8030–8035
assemblies is essential in order to fabricate the giant 2a vesicles. Since the hydrogen-bond-directed giant vesicle is easily prepared and shows excellent stability in water, it might find a wide variety of applications as a novel microcapsule.
Experimental Section Chemicals. The reagents and solvents were purchased from Sigma-Aldrich Co., Tokyo Chemical Industry Co., Kanto Chemical Co., and Wako Pure Chemical Industries, and when necessary, the chemicals were dehydrated and distilled by routine procedures. The water used for vesicle preparation was purified by Millipore Milli-QLabo. The synthesis of compounds 2a and 2b is described in a previous paper.5 Preparation of Vesicle Solutions. The vesicle solutions were prepared either by the injection method or the thin-film method. The Injection Method. A small volume (0.1 mL) of 2a or 2b/ THF solution (1.0 10-1-5.0 10-1 mol dm-3) was injected into 10 mL of pure water (rt) in a glass sample tube (φ = 17 mm) by a microsyringe and stirred by a SIBATA test tube mixer, TTM-1. The Thin-Film Method. Before the film preparation, a glass sample tube (φ = 32 mm) was rinsed with aqueous NaOH solution (1 wt %), washed with water, and dried. A thin film was prepared at the bottom of the tube by slow evaporation of the 2a or 2b/THF solution (1.0 10-2 mol dm-3, 0.1 mL). After drying at 50 °C in vacuo for 3 h, 25 mL of hot pure water (80 °C) was slowly added. A translucent vesicle dispersion was obtained after sonication at 80 °C for 10 min by a Branson Sonifier Cell Disruptor, 185. Since a translucent aqueous vesicle dispersion becomes clear by the addition of ethanol, the concentrations of 2a and 2b in aqueous vesicle solutions were determined photometrically from absorbance at 255 nm after the addition of a triple volume of ethanol to the vesicle solution. To test for the presence of small particles or aggregates, the vesicle solutions were filtered off through a hydrophilic PTFE membrane filter (average pore size 1 μm, Advantec H100A025A), and the amount of 2a or 2b in the filtrate was determined photometrically. For the fluorescence microscopic measurement, the vesicle solution was prepared in the presence of eosin Y (1 10-3 mol dm-3), and eosin Y in the external aqueous phase was subsequently removed by gel permeation chromatography (Sephadex G-25). Measurement. The electronic absorption spectra were measured by a Shimazu UV-3100PC spectrometer. Differential scanning calorimetry (DSC) was carried out using a Perkin-Elmer Pyris 1. The optical and fluorescence microscopic images of the vesicle solution were obtained by a KEYENCE digital microscope system equipped with a VH-Z450 zoom lens and a VB-7010 CCD camera. The sample solution was placed between two glass slides separated by a thin spacer (11 μm). The number of the vesicles was determined by counting the image of the solution in 50 50 11 μm and averaged for 10 different images. The average diameter was determined by measuring 50 different vesicles. For AFM measurements, a small drop of the vesicle solutions was placed on a silicon substrate and the silicon substrate (after keeping in a vacuum overnight) was subjected to AC-mode AFM observation using a JEOL JSPM-4200 equipped with an Olympus silicon microcantilever, OMCL-AC160TS-C2 (radius = 6 nm, resonance frequency = 300 kHz, spring constant = 42 N/m). The force applied to the sample was controlled by amplitude of the cantilever oscillation. For the scan under default setting, amplitude of the cantilever oscillation was maintained near the level of the free oscillation in order to avoid application of strong force (noncontact mode). For the scan with increased force, amplitude of the cantilever oscillation was decreased to approximately half (corresponding to 2.1 nN).
Acknowledgment. This work was partly supported by a Grantin-Aid for Challenging Exploratory Research (no. 21655038) from the Japan Society for the Promotion of Science (JSPS). DOI: 10.1021/la904916d
8035
