Soft Nanotubes Derivatized with Short PEG Chains for Thermally Controllable Extraction and Separation of Peptides.

By means of a two-step self-assembly process involving three components, including short poly(ethylene glycol) (PEG) chains, we produced two different types of molecular monolayer nanotubes: nanotubes densely functionalized with PEG chains on the outer surface and nanotubes densely functionalized with PEG chains in the nanochannel. Turbidity measurements and fluorescence spectroscopy with an environmentally responsive probe suggested that the PEG chains underwent dehydration when the nanotubes were heated above 44-57 °C and rehydration when they were cooled back to 25 °C. Dehydration of the exterior or interior PEG chains rendered them hydrophobic and thus able to effectively extract hydrophobic amino acids from the bulk solution. Rehydration of the PEG chains restored their hydrophilicity, thus allowing the extracted amino acids to be squeezed out into the bulk solutions. The nanotubes with exterior PEG chains exhibited selectivity for all of the hydrophobic amino acids, whereas the interior PEG chains were selective for hydrophobic amino acids with an aliphatic side chain over hydrophobic amino acids with an aromatic side chain. The higher selectivity of the latter system is attributable that the extraction and back-extraction processes involve encapsulation and transportation of the amino acids in the nanotube channel. As the result, the latter system was useful for separation of peptides that differed by only a single amino acid, whereas the former system showed no such separation ability.

Introduction

Poly(ethylene glycol)s (PEGs) have attracted much attention for biological and medical applications involving peptides and proteins. Specifically, owing to the high water solubility, low toxicity and antigenicity, and thermal responsivity of PEGs, they have been widely used to increase the water solubility of peptides and proteins,[1−3] improve their cellular internalization,[4] prolong their blood circulation time,[5,6] facilitate their condensation and separation[7−10] and crystallization,[11,12] control their adsorption,[13−16] suppress their aggregation,[17,18] and accelerate their refolding.[19−22] However, only polydisperse PEGs[23,24] with relatively high molecular weights have been used to date. Recent studies have suggested that the physicochemical properties of PEGs depend strongly not only on their molecular weights and topology[25−30] but also on whether or not they are confined in nanospaces.[31,32]

Noncovalently bonded polymer nanotubes with controllable cavity sizes and functionalizable surfaces[33−35] are important materials for encapsulation, storage, transport, and release of peptides and proteins.[36−38] For example, nanotubes formed by self-assembly of rationally designed amphiphilic molecules in water are useful for qualitative and quantitative analyses of proteins,[39−41] stabilization of proteins under harsh conditions,[42−44] acceleration of protein refolding,[45−47] and mimicking of proteins.[48]

Herein, we report the selective introduction of a dense layer of short PEG chains to the outer surface or the nanochannels of nanotubes. Thermal dehydration and rehydration of the PEG chains enabled us not only to extract hydrophobic amino acids and peptides from bulk solution but also to back-extract them. We discovered that nanotubes with interior PEG chains, but not nanotubes with exterior PEG chains, were useful for separation of peptides that differed by a single amino acid.

Results and Discussion

Construction of Nanotubes Functionalized with Short PEG Chains

As previously reported,[49] self-assembly of rationally designed asymmetric amphiphiles with a different head group at each end produces molecular monolayer nanotubes with different inner and outer surfaces. In this study, we synthesized an asymmetric amphiphile designated lipid 1 and two PEG derivatives (PEG8Ste and GlyGlyPEG8) as components for the construction of two types of nanotubes functionalized with short PEG chains. In one type, the PEG chains were located on the outer surface of the nanotubes, and in the other type, the chains were located on the inner surface (i.e., in the nanochannel) of the nanotubes (Figure ).

Figure 1

Molecular monolayer nanotubes: (left panel) exPEG8-nanotubes composed of lipid 1, GlyGlyEtOH, and PEG8Ste and (right panel) inPEG8-nanotubes composed of lipid 1, GlyGlyPEG8, and lipid 2. The yellow and green bands in the chemical structures shown at the top of each panel indicate areas of intermolecular hydrogen bonding and hydrophobic interactions, respectively. The graphic at the bottom of each panel illustrates the thermal dehydration and rehydration behavior of the nanotubes.

First, we prepared binary self-assemblies of lipid 1 and either GlyGlyEtOH or GlyGlyPEG8 by dispersing a mixture of lipid 1 (10 μmol) and GlyGlyEtOH (10 μmol) or GlyGlyPEG8 (10 μmol) in pure water (10 mL) under reflux conditions and then gradually cooling the hot solution to room temperature. Transmission electron microscopy (TEM) revealed that this process exclusively produced nanotubes—designated GlyGlyEtOH-nanotubes and GlyGlyPEG8-nanotubes—with an inner diameter of 7–9 nm and a wall thickness of 3–4 nm (Figure S1, Supporting Information). Although self-assembly of GlyGlyEtOH or GlyGlyPEG8 alone gave micelles (Figure S2, Supporting Information), we did not observe these morphologies in either of the binary self-assembly systems.

Second, we prepared nanotubes functionalized with short PEG chains on the outer surface, designated exPEG8-nanotubes, by heating the GlyGlyEtOH-nanotubes (composed of 10 μmol each of lipid 1 and GlyGlyEtOH) with PEG8Ste (10 μmol) at about 50 °C in 1:1 (v/v) water/methanol (10 mL). In addition, we prepared nanotubes with short PEG chains on the inner surface, designated inPEG8-nanotubes, by heating GlyGlyPEG8-nanotubes (composed of 10 μmol each of lipid 1 and GlyGlyPEG8) with lipid 2 (10 μmol) under the same conditions. After cooling the hot solutions, we used TEM to confirm that the morphology of the nanotubes had not changed and that no other structures had formed (Figure a,b). We note that lipid 2 alone self-assembles in 1:1 (v/v) water/methanol to form bilayer nanotubes with an inner diameter of about 70 nm and a wall thickness of about 70 nm,[50] and PEG8Ste alone formed nanofibers with widths of 100–500 nm under the same conditions (Figure S2, Supporting Information). On the other hand, coassembly (one-step self-assembly) of the three components gave mixtures including the desirable nanotubes, the intermediate nanotubes, and the self-assembled structures of each component (Figure S3, Supporting Information).

Figure 2

Transmission electron microscopy images of (a) exPEG8-nanotubes and (b) inPEG8-nanotubes. The nanotube channels were visualized by means of negative staining with 2 wt % phosphotungstate. Length distributions of (c) exPEG8-nanotubes and (d) inPEG8-nanotubes. Photographs of the aqueous dispersions of (e) exPEG8-nanotubes and (f) inPEG8-nanotubes.

Variable-temperature circular dichroism spectroscopy enabled us to estimate the gel-to-liquid crystalline phase transition temperature (Tg–l) of the nanotube monolayer membrane formed by chiral molecular packing[51] derived from the chirality of the d-glucose moieties in lipid 1 and lipid 2 (Figure S4, Supporting Information). The two-component nanotubes, that is, the GlyGlyEtOH-nanotubes and GlyGlyPEG8-nanotubes, had relatively low Tg–l values (around 45–55 °C) in water, which we ascribed to void spaces in the molecular packing arising from the lack of long alkyl chains in GlyGlyEtOH and GlyGlyPEG8 for hydrophobic intermolecular interactions.[52] In contrast, the Tg–l values of the three-component nanotubes, that is, the exPEG8-nanotubes and inPEG8-nanotubes, exceeded 100 °C, the suggestion being that the PEG8Ste and lipid 2 molecules, with their long alkyl chains, filled the void spaces within the molecular packing structure of the GlyGlyEtOH-nanotubes and GlyGlyPEG8-nanotubes, respectively (Figure ). Elemental analysis conducted by gas chromatography–mass spectroscopy showed that the lipid 1/GlyGlyEtOH/PEG8Ste and lipid 1/GlyGlyPEG8/lipid 2 molar ratios of the three-component nanotubes were 1.00:0.95:0.95 and 1.00:0.95:0.98, respectively.

IR spectroscopy confirmed the molecular packing of the three components within the monolayer membrane of the nanotubes. GlyGlyEtOH or GlyGlyPEG8 formed a polyglycine-II-type hydrogen bond network with the digylcine moiety of lipid 1 (Figure S5, Supporting Information). Neither PEG8Ste nor lipid 2 disordered the lateral chain packing of the oligomethylene spacer of lipid 1; the packing was assignable to a triclinic parallel (T∥) type, which was easily distinguishable from the orthorhombic perpendicular (O⊥) type that was observed for the PEG8Ste-nanofibers and the bilayer nanotubes self-assembled from lipid 2 (Figures S5 and S6, Supporting Information).

To support the location of the PEG chains functionalized in the nanotubes, we carried out time-lapse fluorescence microscopic observations for the single exPEG8-nanotube and inPEG8-nanotube bearing a fluorescent dye partly immobilized on the exterior PEG chains or the interior PEG chains upon addition of a quencher dye. The quenching based on fluorescence resonance energy transfer from the fluorescent dye of the exPEG8-nanotube to the quencher dye randomly and quickly occurred compared to that in the inPEG8-nanotube system (Figure S7, Supporting Information). This result indicates that the quencher dye easily accesses the fluorescent dye, proving that the PEG chains are located on the outer surface of the nanotubes. On the other hand, the quenching in the case of the inPEG8-nanotube started from both the open ends of the nanotube and gradually moved toward the center part (Figure S8, Supporting Information). This result indicates that the quencher dye penetrates into the nanotube from the outside and gradually quenches the fluorescent dye with flowing in the nanotube channel, proving that the PEG chains are located on the inner surface of the nanotubes. The fact that TEM observations indicated that the inner diameters and membrane wall thicknesses of the two types of three-component nanotubes (Figure a,b) were similar was ascribable to the low contrast of the images of the hydrated PEG chains.[53]

The peaks of the length distributions of the exPEG8-nanotubes and inPEG8-nanotubes estimated from those TEM images were around 350 and 200 nm, respectively (Figure c,d), which are comparable to the size distributions of both the nanotubes in the aqueous dispersions estimated from dynamic light scattering measurements (Figure S9, Supporting Information). Therefore, both the nanotubes well dispersed in water (Figure e,f) should keep the length distributions.

Thermal Dehydration and Rehydration of Exterior and Interior PEG Chains

Long linear PEG chains in water are known to undergo dehydration in response to elevated temperatures, as indicated by conformational changes of the C–C bonds from the gauche form at low temperatures to the anti form at high temperatures.[54−59] As a result of these changes, the solubility of the PEG chains decreases and the aqueous solutions become cloudy, although there are a few reports that concern such thermal dehydration; the examples were short linear PEG chains and oligo(ethylene glycol) chains.[60] By measuring the turbidity of aqueous solutions of the exPEG8-nanotubes, we were able to estimate that the dehydration temperature of the exterior PEG chains was 57 °C (Figure S10, Supporting Information). When the solution was cooled, the turbidity cleared, indicating that the solubility of the exterior chains was increased by rehydration. The rehydration temperature, 43 °C, was slightly lower than the dehydration temperature. The thermal dehydration/rehydration cycle could be repeated several times. Similar turbidity measurements for the PEG8Ste-fibers showed that the PEG chains undergo dehydration at 85 °C, whereas the dehydrated PEG chains never undergo rehydration (Figure S11, Supporting Information).

To investigate the dehydration/rehydration behavior of the inPEG8-nanotubes, we encapsulated an environmentally responsive probe, 8-anilinonaphthalene-1-sulfonate (1,8-ANS), in the nanotubes to detect dehydration and rehydration of the PEG chains lining the nanochannels.[32] We found that raising the temperature produced a remarkable increase in the intensity of the blue-shifted fluorescence band of the 1,8-ANS encapsulated in the inPEG8-nanotubes (Figure S10, Supporting Information). Drastic spectral changes were observed at 44 °C, the implication being that the environment of the nanochannels became relatively hydrophobic at that temperature. The enhancement of the hydrophobicity of the nanochannel is ascribable to thermal dehydration of the interior PEG chains.[32] Because Tg–l of the inPEG8-nanotubes was high (>100 °C), we could exclude the possibility that 1,8-ANS was embedded in the hydrophobic membrane wall.[61,62] Lowering the temperature restored the original fluorescence spectrum, indicating that the nanochannels regained their hydrophilicity upon rehydration of the interior PEG chains. The rehydration temperature was estimated to be 38 °C. The fluorescence spectroscopy also revealed that there is no evidence of thermal dehydration of the GlyGlyPEG8-micelles (Figure S11, Supporting Information).

The dehydration and rehydration temperatures of the inPEG8-nanotubes differed from those of the exPEG8-nanotubes, even though the PEG chains of the two types of nanotubes have similar molecular weights and chemical structures. This difference must be related to (1) the apparent higher modification density of the interior PEG chains as a result of the larger curvature of the inner surface and (2) the presence of confined water molecules with specific physical properties, such as higher viscosity and lower polarity, in the nanotube channels.[42,63]

Peptide Separation by Extraction and Back-Extraction

We utilized the thermal dehydration/rehydration behavior of the PEG chains of the exPEG8-nanotubes and inPEG8-nanotubes for extraction and back-extraction of amino acids. All 20 amino acids (10 μmol each) were separately added to aqueous dispersions (10 mL) of exPEG8-nanotubes (one of the compositions, PEG8Ste: 10 μmol) or inPEG8-nanotubes (one the compositions, GlyGlyPEG8: 10 μmol). The dispersion abilities of the exPEG8-nanotubes and inPEG8-nanotubes were independent of additions of the amino acids, based on no electrostatic attraction of the nonionic glucose OH groups and the PEG chains on the outer surface of the nanotubes with the amino acids. To dehydrate the PEG chains, we heated the mixtures at 70 °C for 30 min, which was long enough to reach extraction equilibrium. Then, the mixtures were quickly filtered through polycarbonate membranes with a pore size of 0.2 μm. The amounts of unextracted amino acids in the filtrates were determined by means of fluorescence spectroscopy with the labeling reagent 4-fluoro-7-nitrobenzofurazan.[64] Extraction ratios were estimated by subtraction of the concentrations determined by fluorescence spectroscopy from the initial concentrations. After extraction and filtration, the nanotubes were redispersed in water (10 mL) at 25 °C for rehydration of the PEG chains. Membrane filtration at various intervals and subsequent fluorescence spectroscopy of the filtrates enabled us to estimate release ratios, that is, ratios for back-extraction of the amino acids from the nanotubes into the bulk solution. Finally, destruction of the nanotubes with dimethyl sulfoxide (DMSO) released any remaining amino acids. We confirmed that total recovery ratios for the above-described procedure were 100 ± 5%.

At elevated temperature, the exPEG8-nanotubes extracted substantial amounts of Trp, Phe, Pro, Met, Ile, Leu, Val, Ala, Gly, and Tyr, all of which are classified as hydrophobic or neutral amino acids (Figure a). Because the exPEG8-nanotubes showed no extraction ability for any of the amino acids unless the temperature was elevated, we contend that the driving force for the extraction was hydrophobic interactions between the hydrophobic amino acids and the dehydrated PEG chains on the outer surface of the nanotubes.

Figure 3

Amino acid extraction ratios obtained with (a) exPEG8-nanotubes and (b) inPEG8-nanotubes.

The inPEG8-nanotubes also extracted several of the hydrophobic amino acids upon thermal dehydration of the PEG chains in the nanotube channels, but all of the extraction ratios were lower than the corresponding ratios for the exPEG8-nanotubes (Figure b). The inPEG8-nanotube selectively extracted amino acids with aliphatic side chains, such as Met, Ile, Leu, Val, and Ala, over amino acids with aromatic or cyclic aliphatic side chains, such as Trp, Phe, and Pro. The hydrophobic interactions between the interior PEG chains and the amino acids occurred via encapsulation of the amino acids from the bulk solution and subsequent transport of the encapsulated amino acids into the nanotube channel, whereas the exterior PEG chains on the exPEG8-nanotubes were in direct contact with the amino acids in the bulk solution. In the former case, the encapsulation and transportation processes influenced not only extraction efficiency but also extraction selectivity.

Lowering the temperature of the nanotube dispersions released the amino acids, allowing them to be back-extracted into the bulk solution (Figure ). Back-extraction was likely induced by disruption of the hydrophobic interactions between the PEG chains of the nanotubes and the amino acids as a result of rehydration of the PEG chains. The back-extraction rates in the exPEG8-nanotube system were remarkably higher than the rates in the inPEG8-nanotube system, and this difference reflects the different locations of the interactions between the amino acids and the nanotubes, that is, the nanotube outer surface versus the nanotube channel. The surface is entirely exposed to the bulk solution, whereas the channel is exposed to the bulk solution only at the two open ends of the nanotube. In the exPEG8-nanotube system, the back-extraction profiles for all of the amino acids were similar (Figure a), whereas in the inPEG8-nanotube system, the profiles strongly depended on the amino acid side chain (Figure b). In the inPEG8-nanotube system, the back-extraction rates of the amino acids increased in the order Pro < Trp < Gly < Ala < Met < Phe < Leu < Val < Ile, which corresponds well to the order of their hydropathy indexes.[65] The more hydrophobic amino acids were preferentially squeezed out of the hydrophilic nanotube channels upon thermal rehydration of the interior PEG chains.

Figure 4

Time dependence of amino acid back-extraction ratios obtained with (a) exPEG8-nanotubes and (b) inPEG8-nanotubes.

We expected that the extraction and back-extraction abilities of the PEG-derivatized nanotubes would make them useful for separation of peptides that shared similar amino acid sequences. To evaluate this possibility, we chose three angiotensin analogues as model peptides: Asp-Arg-Val-Tyr-Ile-His-Pro-Phe (angiotensin II), Asp-Arg-Val-Tyr-Val-His-Pro-Phe (Val5-angiotensin II), and Ala-Arg-Val-Tyr-Ile-His-Pro-Phe (angiotensin A), hereafter abbreviated as Pep(Asp-Ile), Pep(Asp-Val), and Pep(Ala-Ile), respectively, to emphasize the variations in their amino acid sequences. All three peptides (10 μmol each) were added to an aqueous dispersion (10 mL) of the exPEG8-nanotubes (one of the compositions, PEG8Ste: 300 μmol) or inPEG8-nanotubes (one of the compositions: GlyGlyPEG8: 300 μmol). The mixtures were heated at 70 °C for 30 min and then quickly filtered through 0.2 μm pore size polycarbonate membranes. We carried out one and three back-extraction batch processes for the exPEG8-nanotube and inPEG8-nanotube systems, respectively. In batch I, the nanotubes collected on the membrane were redispersed in water (10 mL) and then the dispersion was allowed to stand at 25 °C for 120 min. A sample of the dispersion was subjected to electrospray ionization-mass spectrometry (ESI-MS) for the qualitative analysis of the peptides. The remainder of the dispersion was used for two additional back-extraction processes (batches II and III, respectively), which involved repetition of the membrane filtration and redispersion steps (Figure S12, Supporting Information). ESI-MS was also carried out for samples from batches II and III.

In the exPEG8-nanotube system, the mass spectrum of the batch I sample clearly exhibits signals corresponding to each of the three peptides (Figure a); the spectrum is consistent with the similar back-extraction profiles for the three peptides (Figure S12, Supporting Information). These results indicate that the exPEG8-nanotubes showed no ability to separate the peptides. In contrast, in the inPEG8-nanotube system, the mass spectra of the batch II and III samples indicated that they consisted mainly of Pep(Asp-Ile) and Pep(Asp-Val), respectively (Figure b), whereas the spectrum of the batch I sample showed signals for both Pep(Ala-Ile) and Pep(Asp-Ile). The order in which the peptides appeared in the spectra was the same as the order of the back-extraction rates, Pep(Ala-Ile) > Pep(Asp-Ile) > Pep(Asp-Val), which was in turn consistent with the order of the hydropathy indexes of the amino acids that were varied in the three peptides (Ile > Ala ≫ hydrophilic Asp; Figure S12, Supporting Information). The inPEG8-nanotubes were able to separate these three peptides even though they differed by just one amino acid. This level of selectivity has not previously been reported for peptides, although conventional polymers, such as inverse micelles and amphiphilic monodisperse PEGs, have been used as extraction agents for group separation of peptides.[66,67] The system described herein can be expected to be applicable to sample pretreatments for mass spectrometric analysis of peptides, which usually requires the extraction and condensation of target analytes coexisting with large amounts of bioimpurities.[68,69]

Figure 5

Positive-mode electrospray ionization mass spectra obtained after back-extraction of peptides with (a) exPEG8-nanotubes and (b) inPEG8-nanotubes.

Conclusions

Herein, we have reported that PEG chains on the exterior surface of nanotubes and in nanotube channels can be thermally dehydrated at 57 and 44 °C, respectively. The resulting dehydrated nanotubes can be used to extract hydrophobic amino acids by means of hydrophobic interactions with the PEG chains. Cooling the nanotube dispersions to 25 °C led to release of the extracted amino acids from the nanotubes to the bulk solutions by disruption of the hydrophobic interactions as a result of rehydration of the PEG chains. Nanotubes with interior PEG chains were superior to nanotubes with exterior PEG chains in terms of extraction selectivity, owing to the encapsulation and transport of the amino acids in the nanotube channel. The extraction and back-extraction abilities of the interior PEG chains allowed us to separate peptides on the basis of a difference in a single amino acid in their sequences. This system can be expected to be widely applicable for pretreatments of peptide samples for peptidomics and proteomics analysis.

Experimental Section

Synthesis of GlyGlyEtOH

Z-GlyGly-OSu (Bachem) was condensed with ethanol amine in methanol at room temperature. After evaporation of the solvent, the residue was washed with 5% aqueous citric acid and 10% aqueous NaHCO3. Hydrogenation over Pd/C in methanol removed the benzyloxycarbonyl (Z) protecting group. Condensation of the resulting compound (NH2GlyGlyNHCH2CH2OH) and acetyl chloride in methanol and subsequent recrystallization from water gave the target compound (68% overall yield). 1H NMR (500 MHz, DMSO-d6, δ): 8.11 (t, 1H, NH), 8.05 (t, 1H, NH), 7.76 (t, 1H, NH), 4.67 (t, 1H, OH), 3.68 (d, 2H, −NHCH2C=O), 3.66 (d, 2H, −NHCH2C=O), 3.38 (q, 2H, −NHCH2CH2OH), 3.12 (q, 2H, −NHCH2CH2OH), 2.12 (s, 3H, −CH3). ESI-MS (m/z): 218.1 [M + H]+. Anal. calcd for C8H15N3O4: C 44.23, H 6.96, N 19.34. Found: C 44.28, H 6.95, N 19.37.

Synthesis of Lipid 1

18-[(2,3,4,6-Tetra-O-acetyl-N-β-d-glucopyranosyl)carbamoyl]octadecanoic acid[70] was coupled with NH2GlyGlyNHCH2CH2OH in the presence of 4-(4,6-dimethoxy-1,3,5-triazin-2-yl)-4-methyl morpholinium chloride in methanol at room temperature, and the coupled product was purified by recrystallization from ethanol. Removal of the acetyl protecting groups on the sugar moiety by methanolysis with NaOMe gave lipid 1 (45% overall yield). 1H NMR (500 MHz, DMSO-d6, δ): 8.11 (t, 1H, NH), 8.09 (d, 1H, NH), 7.80 (t, 1H, NH), 7.51 (t, 1H, NH), 4.96 (d, 1H, OH-4), 4.87 (d, 1H, OH-3), 4.81 (d, 1H, OH-2), 4.69 (t, 1H, H-1), 4.66 (br, 1H, OH), 4.47 (t, 1H, OH-6), 3.69 (d, 2H, −NHCH2C=O), 3.65 (d, 2H, −NHCH2C=O), 3.63 (m, 1H, H-6), 3.40 (m, 1H, H-6), 3.39 (q, 2H, −NHCH2CH2OH), 3.16 (m, 1H, H-4), 3.13 (q, 2H, −NHCH2CH2OH), 3.0 (m, 3H, H-2, H-3, H-5), 2.04 (m, 4H, −CH2C=O), 1.47 (m, 4H, −CH2CH2C=O), 1.24 (m, 28H, −CH2−). ESI-MS (m/z): 661.4 [M + H]+. Anal. calcd for C32H60N4O10: C 58.16, H 9.15, N 8.48. Found: C 58.18, H 9.10, N 8.47.

Synthesis of PEG8Ste

PEG8Ste was synthesized in 98% yield by a coupling reaction between stearoyl chloride and m-dPEG8-amine (Quanta Biodesign). 1H NMR (500 MHz, DMSO-d6, δ): 7.88 (t, 1H, NH), 3.59 (m, 24H, −OCH2CH2O−), 3.47 (m, 2H, −CH2−), 3.37 (m, 2H, −CH2−), 3.28 (s, 3H, −OCH3), 3.18 (m, 2H, −CH2−), 3.10 (m, 2H, −CH2−), 2.04 (m, 2H, −CH2C=O), 1.46 (m, 2H, −CH2−), 1.23 (m, 28H, −CH2CH2C=O), 0.85 (t, 3H, −CH3). ESI-MS (m/z): 650.5 [M + H]+. Anal. calcd for C35H71NO9: C 64.68, H 11.01, N 2.16. Found: C 64.58, H 11.10, N 2.07.

Syntheses of Lipid 2 and GlyGlyPEG8

Lipid 2 and GlyGlyPEG8 were synthesized as reported previously.[32,50]

Morphological Observations

Aqueous dispersions of the self-assembled structures were dropped onto a carbon grid, negatively stained with a phosphotungstate solution (2 wt %, pH adjusted to 7 with NaOH), and observed with a transmission electron microscope (H-7000; Hitachi) at 75 kV.

Molecular Packing Analysis

The self-assembled structures were lyophilized and analyzed with a Fourier transform IR spectrometer (FT-620; JASCO) operated at 4 cm–1 resolution and equipped with an unpolarized beam, an attenuated total reflection accessory system (Diamond MIRacle, horizontal attenuated total reflection accessory with a diamond crystal prism; PIKE Technologies), and a mercury cadmium telluride detector. The X-ray diffraction patterns of the structures were measured with a Rigaku diffractometer (type 4037) using graded d-space elliptical side-by-side multilayer optics, monochromated Cu Kx3b1; radiation (40 kV, 30 mA), and an imaging plate (R-Axis IV). The exposure time was 5 min with a 150 mm camera length. Circular dichroism spectra of aqueous dispersions of the nanotubes were measured with a spectropolarimeter (J-820; JASCO) equipped with a temperature control unit (PTC-423L; JASCO).

Preparation of inPEG8-Nanotubes Encapsulating 1,8-ANS

Lyophilized inPEG8-nanotubes prepared from 10 μmol each of lipid 1, GlyGlyPEG8, and lipid 2 were added to an aqueous solution of 1,8-ANS (50 μmol). After aging overnight, the mixture was filtered through a polycarbonate membrane with a pore size of 200 nm. The inPEG8-nanotubes on the membrane were washed several times with water to remove any 1,8-ANS on the outside of the nanotubes. After complete destruction of the inPEG8-nanotubes by heating in dimethyl sulfoxide, measurement of UV–vis spectra with a spectrophotometer (U-3300; Hitachi) equipped with a temperature control unit (BU150A; Yamato) allowed us to calculate the amount (2.7 μmol) of encapsulated 1,8-ANS. The fluorescence spectrum of the encapsulated 1,8-ANS was recorded with a spectrophotometer (F-4500; Hitachi) equipped with a DCI temperature control unit (Haake).