Supporting Information
© Wiley-VCH 2008
69451 Weinheim, Germany
Supporting Information
Directed Assembly of Sub-Nanometer Thin Organic Materials with Programmed-Size Nanopores.
Delia C. Danila, L. Todd Banner, Evguenia J. Karimova, Lyudmila Tsurkan, Xinyan Wang, and Eugene Pinkhassik
Materials and Methods
Materials. 1,2-Dilauroyl-sn-Glycero-3-Phosphocholine (DLPC) was purchased from Avanti Lipids, α-Dglucopyranose pentaacetate (GPA), α-D-glucopyranose pentabenzoate (GPB), methyl orange, tert-butylstyrene,
divinylbenzene, β-cyclodextrin, Reactive Blue 2™ dye, calcein, fluorescein isothiocyanate-dextran (average MW
4,000), esterase, Sepharose 4B, NaOH, and Triton X-100 were purchased from Sigma-Aldrich. Procion Red™ dye
was donated by DyStar Textilfarben GmbH.
Synthesis of Reactive Blue 2/β–Cyclodextrin conjugate. The synthesis was based on the previously reported
[1]
literature procedure for dyeing dextrans with procion dyes. A freshly prepared solution of β-cyclodextrin (500 mg,
0.4 mmol) in water (40 ml) was mixed with a freshly prepared solution of Reactive Blue 2 (550 mg, 0.4 mmol, 60%
of dye in reagent) in water (20 ml). After stirring the mixture for 5 minutes at ambient temperature, NaCl (2 g) was
added. 30 minutes later, Na2CO3 (100 mg) was added, and the reaction mixture was stirred overnight. The mixture
was purified using size exclusion column chromatography (Sephadex G-25) using DI water (pH 6.8) as eluent. The
blue colored fraction that eluted shortly after the column void volume was collected. The solvent was evaporated
and the sample was further dried under reduced pressure for 3 hours. The yield was 670 mg (87%). The sample
homogeneity was confirmed with HPLC and TLC. The HPLC analysis (Figure 1) was performed using a Waters
600 pump and Waters 2487 UV/Vis dual wavelength detector equipped with a semi-preparative flow cell. The
detector wavelengths were set at 200 nm and 610 nm. The column was a Nova-Pak HR C-18, 60Å, 19 x 300 mm).
The solvent was 75% water/25% acetonitrile and the flow rate was 10 mL/min. The sample of the product exhibited
a single peak at both wavelengths with retention time of 4.40 min. The reference sample of Reactive Blue dye
eluted at 3.85 min.
Figure 1. HPLC chromatograms of Reactive Blue 2 (lower) and the Reactive Blue 2/β-cyclodextrin conjugate
(upper) at 200 nm (blue) and 610 nm (red).
The TLC analysis was performed on silica gel using 50% ethanol/50% ethyl acetate as eluent. The sample
of the product exhibited a single spot with Rf 0.5. 1H NMR data were acquired on a Varian DirectDrive 500 MHz
spectrometer. 1H NMR (D2O): δ 8.51-8.24 (m, 4H), 8.11-7.62 (m, 5H), 7.51-7.06 (m, 3H), 5.05 (d, J=3.4 Hz, 7H),
3.96-3.52 (m, 42H). Mass spectra were acquired on a ThermoElectron LCQ Advantage LC-MS system.
Electrospray MS (1% trifluoroacetic acid in acetonitrile; [M+4H+-Na+]3+ 639.55 (calcd. 639.46). Structures of neutral
and ionized conjugate molecules are shown on Figure 2. Four amino groups are protonated in acidic conditions
(shown in red rectangles). Simultaneously, one sodium cation is lost to yield a negatively charged sulfonate (shown
in blue rectangle) that forms a zwitterion with the nearby cation.
Supporting Information
Figure 2. Structures of neutral (top) and ionized (bottom) Reactive Blue 2/β-cyclodextrin conjugate.
Mass spectra revealed no signals corresponding to products with 1:2 or higher cyclodextrin to dye ratio. 1:1
ratio of β-cyclodextrin and Reactive Blue 2 in the conjugate was further confirmed by UV/Vis spectroscopy from
absorbance of a weighed conjugate sample compared with absorbance of weighed sample of Reactive Blue 2.
Wang et al. found no difference in the extinction coefficient between free and bound dichlorine triazine dyes as well
as very little spectral shift.[2]
Figure 2 shows the cyclodextrin alkylated in its most reactive and more sterically accessible C(6)-OH
position.[3] A cyclodextrin derivative alkylated in the less reactive and more hindered C(2)-OH position may be
present as a minor product. It is highly unlikely that the C(3)-OH position of β-cyclodextrin is alkylated under these
conditions.[3] Although spectral data do not unambiguously distinguish among constitutional isomers of compound 3
(including isomers due to different positions of sulfonate group in the terminal benzene ring), this has no effect on
the functional performance. We always observed a single peak in HPLC analysis with different ratios of water and
acetonitrile as eluent. The smallest cross-section of all possible 1:1 Reactive Blue 2/β-cyclodextrin conjugates is
the same and is equal to that of β-cyclodextrin. From practical standpoint, there is no difference for the functional
performance as a size probe among regioisomers of 1:1 Reactive Blue 2/β-cyclodextrin conjugates or their
mixtures.
Liposome Preparation. Unilamellar liposomes were prepared by the extrusion of multilamellar vesicles formed
spontaneously upon hydration of a lipid film.[4,5] The lipid film was prepared by evaporating solvent from DLPC
solution (500 µl, 20 mg/ml in chloroform) to dryness in a stream of dry argon. To form pores, a solution of α-Dglucopyranose pentabenzoate (GPB) (500 µl, 8.7×10-4 M in ethyl ether) or α-D-glucopyranose pentaacetate (GPA)
(500 µl, 8.7×10-4 M in ethyl ether) was mixed with the solution of DLPC in chloroform prior to solvent removal. After
solvents were evaporated (5-10 munites), the lipid film was further dried in high vacuum at ambient temperature for
1 hour. Multilamellar vesicles were formed by hydration of lipid film with an aqueous solution containing size
probes. We used TRIS buffer (50 mM, pH 7.4 at 25 ºC), phosphate buffer (100 mM, pH 7.6 at 25 ºC) or DI water
Supporting Information
with identical results. Typically, the following concentrations were used: methyl orange, 100 mM, procion red 10
mM, and β-CD-Reactive Blue conjugate, 3.7 mM. To prepare size probe solutions, we dissolved either individual
dyes or a mixture of all three dyes in water or one of the buffers above and adjusted the pH to 7.0 for DI water or
the original pH of the buffer with NaOH. The unilamellar liposomes were formed by extrusion through 100 nm
polycarbonate membranes in a mini-extruder apparatus purchased from Avanti Lipids. The average size of the
liposomes was approximately 100 nm with most liposomes in the range of 70 to 130 nm as confirmed by the
transmission electron microscopy and dynamic light scattering.
Loading monomers into bilayer and measurement of entrapped monomers. To load monomers into the
bilayer interior, 4-tert-butylstyrene (6 µl) and para-divinylbenzene (5 µl) were added to 1 ml of liposome solution,
and the solution was stirred for 24 h at 4 ºC. To measure the amount of monomers incorporated into the bilayer,
free monomers were microseparated from the liposome solution in a Pasteur pipette followed by the overnight
extraction of the aqueous solution with 1 ml of 0.4 mg/mL tolune in hexane. An aliquot of the solvent was analyzed
by gas chromatography. Amounts of 4-tert-butylstyrene and para-divinylbenzene were quantified against internal
standard (toluene). The DLPC: 4-tert-butylstyrene: para-divinylbenzene ratio was found to be 1.0:0.46:0.43.
Preparation of organic nanocapsules. Liposomes containing monomers in the lipid bilayer were irradiated by UV
light in a home-built apparatus (two 4W UV lamps, λ= 254 nm) for four hours at ambient temperature. Triton X-100
(1 ml, 2% in water) was added to the solution to remove the outer shell of phospholipids. An aliquot of the solution
(50 µl) was separated on size-exclusion chromatography column (Sephrose-4B) to measure the amount of retained
dyes. NaOH (0.5 ml, 0.1 M) was added to the remaining solution to hydrolyze the pore-forming template. An aliquot
of the solution was taken for TEM analysis. Another aliquot (100 µl) was separated through a size-exclusion
chromatography column (Sepharose 4B) to remove released size probes. The nanocapsules fraction was collected
and analyzed with UV-vis spectroscopy (Agilent 8453 UV-vis spectrophotometer) to quantify the amounts of
entrapped colored size probes. The amounts of retained size probes before and after exposure to NaOH are
summarized in Table 1.
Table 1. Retention of size probes after the polymerization.
GPA
1
Before NaOH
After NaOH
[a]
not detectable.
78±11
[a]
ND
GPB
2
3
1
89±13
83±8
79±14
86±7
84±5
ND
2
[a]
3
91±12
ND
[a]
83±9
86±6
It appears that the templates are embedded in the crosslinked polymer matrix and the base hydrolysis is
indeed needed to open the pores.
Transmission electron microscopy (TEM). TEM images were acquired on a JEOL JEM1200EX II microscope.
Samples were negatively stained with phosphotungstic acid (pH=5.9) on a carbon grid.
Structural characterization
The formation of nanocapsules was confirmed unambiguously by TEM data. The following samples were
examined by TEM: 1) DLPC liposomes in water , 2) 1:1 mixture of liposomes and with Triton X-100 (2%), 3) pure
Triton X-100 (2%), 4) monomers polymerized in H2O, in the absence of the bilayer scaffold, 5) polymer
nanocapsules formed by the polymerization of 4-tert-butylstyrene with para-divinylbenzene loaded into bilayers of
DLPC liposomes followed by mixing with Triton X-100 to dissolve lipids, 6) same as (5) but containing GPA in the
bilayer and with an additional step of GPA hydrolysis with NaOH, 7) same as (6) but containing GPB in the bilayer.
Figure 3 shows nearly circular structures corresponding to DLPC liposomes. When liposomes are lysed
with Triton X-100, no such structures are visible in TEM (Figure 4). Smaller structures with sizes of 5-20 nm on
Figure 4 correspond to detergent micelles, which are also present in the TEM image of pure Triton X-100
(Figure 5). A control sample of co-polymerized 4-tert-butylstyrene and para-divinylbenzene in water in the absence
of liposomes (Figure 6) shows no structures in the 50-150 nm range. Figures 7-9 show samples prepared by the
polymerization of 4-tert-butylstyrene and para-divinylbenzene in the interior of DLPC bilayers followed by dissolving
lipids with Triton X-100. Samples on Figures 8 and 9 were additionally treated with NaOH to hydrolyze GPA and
GPB, respectively. Figures 7-9 reveal nearly circular structures that are similar in size to the original liposomes.
Absence of these structures in Figures 4-6 unambiguously proves formation of organic nanocapsules templated in
the interior of phospholipid bilayers.
Supporting Information
Figure 3. DLPC liposomes.
Figure 4. DLPC liposomes lysed with Triton X-100
Supporting Information
Figure 5. Triton X-100
Figure 6. tert-Butylstyrene and divinylbenzene polymerized in
water in the absence of liposomes.
Supporting Information
Figure 7. Organic nanocapsules formed from tert-butylstyrene
and divinylbenzene in the absence of pore-forming templates.
Figure 8. Organic nanocapsules formed from tert-butylstyrene
and divinylbenzene with glucose pentaacetate as a poreforming template.
Supporting Information
Figure 9. Organic nanocapsules formed from tert-butylstyrene
and divinylbenzene with glucose pentabenzoate as a poreforming template.
Release of fluorescent markers
The size-probe retention assay was corroborated by studying the release of self-quenching fluorescent
dyes from porous nanocapsules. Recently, Gorteau et al. used a similar method to evaluate self-assembled pores
in liposomes.[6] Capsules were prepared as described above with encapsulated calcein (10 mM in phosphate
buffer, 0.1 M, pH 7.6 at 25 ºC) or fluorescein isothiocyanate-dextran (FITC-D, average MW 4,000; 7.5 mM in
phosphate buffer, 0.1 M, pH 7.6 at 25 ºC). Non-encapsulated fluorescent markers were separated by sizeexclusion chromatography (Sepharose-4B). The solution of nanocapsules (0.5 ml) was mixed with NaOH (0.5 ml,
0.1 M). Aliquots were taken before mixing with NaOH (time 0) and 30 min after adding NaOH. The pH in the 30 min
sample was adjusted to 7.6 with 0.1 M HCl. Since calcein does not exhibit fluorescence at alkaline pH, we could
not continuously monitor the efflux of calcein from capsules. The emission was measured at 520 nm and was
calibrated by the lysis of a control sample of liposomes with encapsulated calcein or FITC-D (excess Triton X-100).
The emission as a function of time of exposure to NaOH is shown on Figure 10.
0.8
0.7
Emission
0.6
Calcein, GPA
Calcein, GPB
FITC-D, GPA
FITC-D, GPB
0.5
0.4
0.3
0.2
0.1
0
0
30
Time, min
Figure 10.
Calcein. The increase of emission is due to release of calcein from nanocapsules and dilution in the
exterior buffer. The size of calcein is close to the size of Procion Red dye used as the 1.1 nm size probe. We
Supporting Information
observed release of calcein from capsules prepared using GPB as pore-forming template. No fluorescence
increase was observed with capsules prepared with GPA as templates. This indicates that GPA did not produce
sufficiently large pores to release calcein, while GPB-templated pores were large enough for complete release of
calcein from nanocapsules.
FITC-D. We observed no increase of fluorescence with either GPA or GPB as templates. This indicates
that neither GPA nor GPB produce sufficiently large pores to release FITC-D.
These combined observations validate the size probe retention assay used here to measure the pore size.
They also support the role of base in opening pores.
References
[1]. W.F. Dudman, C.T. Bishop, 1968, Can. J. Chem. 46, 3079-3084.
[2]. S.-J. Wang, H.-M. Lin, X.-R. Wang, K.-P. Hsiung, Y.-C. Liu, 2007, Anal. Biochem. 361, 190-196.
[3]. C.J. Easton, S.F. Lincoln, Modified Cyclodextrins: Scaffolds and Templates for Supramolecular Chemistry,
Imperial College Press, 1999, pp. 43-70.
[4]. V. Torchilin, V. Weissig, Eds. Liposomes: a Practical Approach. (Oxford University Press, 2003).
[5]. F. Olson, C.A. Hunt, F.C. Szoka, W.J. Vail, D. Papahadjopoulos, Biochim. Biophys. Acta, 1979, 557, 9-23.
[6]. V. Gorteau, G. Bollot, J. Mareda, D. Pasini, D.-H. Tran, A. Lazar, A.W. Coleman, N. Sakai, S. Matile,
Bioorg. Med. Chem. 2005, 13, 5171-5180.