Polyelectrolyte Block Copolymers - American Chemical Society

CH-1015 Lausanne, Switzerland. Received January 28, 2005. In Final Form: March 18, 2005. This report describes the aqueous solution self-assembly of a...
0 downloads 0 Views 345KB Size
Download PDF
6582

Langmuir 2005, 21, 6582-6589

Nonspherical Assemblies Generated from Polystyrene-b-poly(L-lysine) Polyelectrolyte Block Copolymers Anke Lu¨bbert,† Valeria Castelletto,‡ Ian W. Hamley,‡ Harald Nuhn,§ Markus Scholl,§ Laurent Bourdillon,| Christine Wandrey,| and Harm-Anton Klok*,§ Max Planck Institute for Polymer Research, Ackermannweg 10, D-55128 Mainz, Germany, Centre for Self-Organising Molecular Systems, University of Leeds, Leeds LS2 9JT, United Kingdom, E Ä cole Polytechnique Fe´ de´ rale de Lausanne (EPFL), Institut des Mate´ riaux, Laboratoire des Polyme` res, Baˆ timent MX-D, CH-1015 Lausanne, Switzerland, and E Ä cole Polytechnique Fe´ de´ rale de Lausanne (EPFL), Institut des Sciences et Inge´ nierie Chimiques, Laboratoire de Me´ decine Re´ ge´ ne´ rative et de Pharmacobiologie, CH-1015 Lausanne, Switzerland Received January 28, 2005. In Final Form: March 18, 2005 This report describes the aqueous solution self-assembly of a series of polystyrenem-b-poly(L-lysine)n block copolymers (m ) 8-10; n ) 10-70). The polymers are prepared by ring-opening polymerization of -benzyloxycarbonyl-L-lysine N-carboxyanhydride using amine terminated polystyrene macroinitiators, followed by removal of the benzyloxycarbonyl side chain protecting groups. The critical micelle concentration of the block copolymers determined using the pyrene probe technique shows a parabolic dependence on peptide block length exhibiting a maximum at n ) ∼20 (m ) 8) or n ) ∼60 (m ) 10). The shape and size of the aggregates has been studied by dynamic and static light scattering, small-angle neutron scattering (SANS), and analytical ultracentrifugation (AUC). Surprisingly, Holtzer and Kratky analysis of the static light scattering results indicates the presence of nonspherical, presumably cylindrical objects independent of the poly(L -lysine)n block length. This is supported by SANS data, which can be fitted well by assuming cylindrical scattering objects. AUC analysis allows the molecular weight of the aggregates to be estimated as several million g/mol, corresponding to aggregation numbers of several 10s to 100s. These aggregation numbers agree with those that can be estimated from the length and diameter of the cylinders obtained from the scattering results.

Introduction Amphiphilic polyelectrolyte block copolymers are an interesting but complex class of macromolecules that combine features of polyelectrolytes, block copolymers, and surfactants.1,2 Depending on the total block copolymer molecular weight, the relative block lengths, and the chemical composition of the constituent blocks, solution self-assembly of amphiphilic block copolymers can lead to a variety of different organized structures, including spherical and rodlike micelles and vesicles. The selfassembly properties of both low molecular weight and high molecular weight amphiphiles are often described in terms of a critical packing parameter, which reflects the overall geometry of the molecules and compares the volume of the hydrophilic headgroup with that of the hydrophobic part of the molecule.3 Block copolymers composed of a hydrophobic synthetic block and a charged polypeptide block are a particularly interesting class of * Corresponding author fax: +41 21 693 5650; e-mail: [email protected]. † Max Planck Institute for Polymer Research. ‡ University of Leeds. § E Ä cole Polytechnique Fe´de´rale de Lausanne, Institut des Mate´riaux. | E Ä cole Polytechnique. Fe´de´rale de Lausanne, Institut des Sciences et Inge´nierie Chimiques. (1) Fo¨rster, S.; Abetz, V.; Mu¨ller, A. H. E. Adv. Polym. Sci. 2004, 166, 173-210. (2) Soo, P. L.; Eisenberg, A. J. Polym. Sci., Part B: Polym. Phys. 2004, 42, 923-938. (3) Israelachvili, J. N. Intermolecular and Surface Forces, 2nd ed.; Academic Press: 1992.

polyelectrolyte block copolymers. Because of the ability of the peptide segment to fold into ordered secondary structures, peptide-synthetic hybrid polyelectrolyte block copolymers can adopt unconventional topologies, such as, for example, rod-coil structures in case of R-helical peptide blocks, which may lead to novel and unexpected aggregate morphologies. An interesting example that illustrates the effect of block copolymer topology on aggregate morphology is the work by Nolte et al., who found that polystyreneb-poly(isocyanodipeptide) block copolymers can self-assemble into helical superstructures.4 Pioneering work in the area of synthetic-polypeptide hybrid block copolymers has been performed by Gallot and co-workers.5 These authors have prepared a variety of polystyrene and polybutadiene based peptide hybrid block copolymers and have studied structure formation of these materials, both in the solid state5 and in concentrated aqueous solution,6 where mesomorphic gels were found. The group of Kataoka has extensively studied hybrid block copolymers composed of poly(ethylene glycol) and poly(L-aspartic acid), poly(L-glutamic acid), and poly(L-lysine) and has particularly focused on exploring the potential of these materials for drug delivery and gene therapy applications.7 More recently, we and others have reported (4) Cornelissen, J. J. L. M.; Fischer, M.; Sommerdijk, N. A. J. M.; Nolte, R. J. M. Science 1998, 280, 1427-1430. (5) For a review, see: Gallot, B. Prog. Polym. Sci. 1996, 21, 10351088. (6) Billot, J.-P.; Douy, A.; Gallot, B. Makromol. Chem. 1976, 177, 1889-1893.

10.1021/la0502600 CCC: $30.25 © 2005 American Chemical Society Published on Web 06/02/2005

Polyelectrolyte Block Copolymers

Langmuir, Vol. 21, No. 14, 2005 6583 Scheme 1

Table 1. Isolated Yields and Molecular Characteristics of the PSmPLysn Block Copolymers sample 1a 1b 1c 4af 4bf 4cf 4dg 4eg 4fh 4gg 4hh 4ig

PS10 PS10 PS10 PS8Lys9 PS8Lys22 PS8Lys29 PS10Lys39 PS10Lys52 PS10Lys59 PS10Lys63 PS10Lys67 PS10Lys72

DPn,tha [-]

yieldb [%]

DPnc [-]

Mnd [g/ mol]

DPnd [-]

Mw/Mnd [-]

Mne [g/ mol]

DPn,Lyse [-]

Mw/Mne [-]

10 10 10 10 20 30 40 50 60 60 70 70

90 90 90 41 51 59 59 73 74 90 80 80

10 10 8 9 22 29 39 52 59 63 67 72

820 890 760 4500 6900 11 300 13 300 20 600 24 100 24 800 31 800 29 400

7 8 6 13 23 40 59 74 87 90 117 108

1.22 1.17 1.17 1.15 1.26 1.30 1.44 1.43 1.20 1.36 1.17 1.39

1500 4500 8600 11 200 11 900 17 600 17 100 14 200

3 17 36 48 52 79 76 62

1.41 1.52 1.52 1.54 1.40 1.35 1.48 1.40

a Calculated number-average degree of polymerization of the poly(L-lysine) block on the basis of the monomer/initiator ratio. b Isolated yield of the deprotected block copolymer. c Peptide chain length determined from 1H NMR spectroscopy (DMSO-d6) of the deprotected copolymers. d GPC data (DPn ) number-average degree of polymerization of the peptide block, Mn ) number-average molecular weight, and Mw/Mn ) polydispersity) of the side chain protected block copolymers using DMF as the mobile phase and polystyrene calibration standards. e GPC data (DPn ) number-average degree of polymerization of the peptide block, Mn ) number-average molecular weight, and Mw/Mn ) polydispersity) of the deprotected block copolymers in ACN/water/TFA as the mobile phase and poly(styrene sulfonate) calibration standards f From 1c as macroinitiator. g From 1b as macroinitiator. h From 1a as macroinitiator.

on the dilute aqueous solution self-assembly of polybutadiene-b-poly(L-glutamic acid) block copolymers.8,9 Depending on the block length ratio, these block copolymers were found to form either spherical micellar or vesicular aggregates. In this contribution, the dilute solution properties of a different class of peptide-synthetic hybrid block copolymers will be discussed. The block copolymers that are the subject of the present contribution are composed of a very short, hydrophobic polystyrene block (degree of polymerization ∼10) and a poly(L-lysine) segment containing ∼10-70 repeat units. Results and Discussion Synthesis and Molecular Characterization. The synthesis of the polystyrene-b-poly(L-lysine) (PSmPLysn) block copolymers is outlined in Scheme 1. In a first step, a primary amine terminated polystyrene macroinitiator is used to start the polymerization of N-benzyloxycarbonyl-L-lysine N-carboxyanhydride (Z-Lys NCA). Three batches of polystyrene macroinitiators have been used, which are designated 1a, 1b, and 1c. According to 1H NMR spectroscopy, the number-average degree of polymerization of 1a and 1b is ten while 1c contains on average eight styrene repeat units. The polypeptide chain length is adjusted by variation of the monomer to polystyrene macroinitiator ratio. For the synthesis of PSmPLysn block copolymers with n e 30, a polystyrene amine hydrochloride macroinitiator is used. The use of primary amine hydrochlorides instead of primary amines for the ring-opening (7) Kataoka, K.; Harada, A.; Nagasaki, Y. Adv. Drug Delivery Rev. 2001, 47, 113-131. (8) Che´cot, F.; Lecommandoux, S.; Gnanou, Y.; Klok, H.-A. Angew. Chem., Int. Ed. 2002, 41, 1340-1343. (9) Kukula, H.; Schlaad, H.; Antonietti, M.; Fo¨rster, S. J. Am. Chem. Soc. 2002, 124, 1658-1663.

polymerization of Z-Lys NCA has been reported to suppress the concurrent “activated monomer” pathway and lead to the formation of narrow polydispersity block copolymers.10 All polymerizations are performed in dry DMF under the exclusion of moisture for a period of five days. After that, the reaction mixture is precipitated in water and the polymer is extensively washed with diethyl ether. Then, in a second reaction step, the Z groups that protect the L-lysine side chains are removed by treating the block copolymers with 33 wt % HBr/AcOH in trifluoroacetic acid (TFA). Finally, the crude deprotected block copolymers are repeatedly dialyzed against water until the conductivity remains constant and are isolated via lyophilization. The resulting block copolymers are characterized by means of 1H NMR spectroscopy and GPC. The results of these analyses are summarized in Table 1. As expected for primary amine initiated NCA polymerizations, the peptide block lengths determined via 1H NMR spectroscopy are in good agreement with the calculated values on the basis of the monomer/initiator ratio. GPC analysis of the PSmPZLysn and the PSmPLysn block copolymers is performed using DMF and acetonitrile (ACN)/water/TFA (50/50/0.2, v/v/v) as the mobile phase, respectively. Figure 1 shows representative GPC traces for both series of block copolymers. For most samples, there is a discrepancy between the peptide block lengths determined from the GPC experiments conducted in DMF versus those obtained from GPC in aqueous solution. Also, the GPC block lengths generally deviate from those obtained from the 1H NMR spectra. These differences are not too surprising, since the GPC molecular weights were evaluated using polystyrene or poly(styrene sulfonate) standards, which are expected to show a hydrodynamic (10) Dimitrov, I.; Schlaad, H. Chem. Commun. 2003, 2944-2945.

6584

Langmuir, Vol. 21, No. 14, 2005

Lu¨ bbert et al.

Figure 2. Pyrene excitation spectra of aqueous solutions containing 5.8 × 10-7 M pyrene and different concentrations of PS10Lys52 (4e) (c [mg/mL]; λex ) 390 nm).

Figure 1. (a) GPC chromatograms (DMF, 0.1 M LiBr, 60 °C) of the PSmPZLysn block copolymers. Curves indicated as (- - - -), (s), and (- - -) represent block copolymers obtained with polystyrene macroinitiators 1a, 1b, and 1c, respectively. (b) GPC chromatograms of the PSmPLysn block copolymers (acetonitrile, water, TFA (50/50/0.2, v/v/v)). Curves indicated as (- - - -), (s), and (- - -) represent block copolymers obtained with polystyrene macroinitiators 1a, 1b, and 1c, respectively.

behavior that is different from the PSmPZLysn and the PSmPLysn block copolymers. The GPC experiments are useful, however, in that they show only a single peak for the experiments with the ACN/water/TFA mobile phase. This suggests that under these conditions aggregation does not take place and the block copolymer molecules elute as individual species. In the remainder of this article, polypeptide chain lengths refer to the numbers obtained from the 1H NMR spectra of the deprotected diblock copolymers. Critical Micelle Concentration. The critical micelle concentration (cmc) of the block copolymers has been determined using the well-known pyrene fluorescent probe technique.11 To this end, a series of aqueous solutions with block copolymer concentrations ranging from 1000 µg/ mL to 0.1 µg/mL and a constant pyrene concentration of 5.8 × 10-7 M is prepared. Excitation spectra reveal an increase in intensity and a shift of the pyrene (0,0) band from 334.4 to 339 nm with increasing block copolymer concentration. This is illustrated in Figure 2 for PS10PLys52 (4e). From these spectra, the cmc can be determined by plotting the ratio of fluorescence intensities at 339 and 334.5 nm (I339/I334.5) versus the logarithm of the block copolymer concentration, as illustrated in Figure 3 for PS10PLys52 (4e). From the graph in Figure 3, the cmc is taken as the intersection of the tangents to the two linear parts of the curve. The cmc’s of all block copolymers used in this study are listed in Table 2. Figure 4 plots the cmc’s listed in Table 2 versus the poly(L-lysine) block length. The data shown in Figure 4

can be divided into three groups on the basis of the polystyrene macroinitiator from which the corresponding PSmPLysn block copolymers are prepared. The cmc’s of the block copolymers prepared from 1b and 1c display a parabolic behavior and pass through a maximum with increasing peptide block length. The pronounced difference in micellization behavior between block copolymers based on 1b and 1c is striking and illustrates how sensitive solution self-assembly is to molecular composition. Only two block copolymers have been prepared using 1a as macroinitiator, but the cmc’s of these block copolymers are in good agreement with those of block copolymers based on 1b with similar poly(L-lysine) block lengths. A similar parabolic dependence of cmc on ionic block length has been observed for polystyrene-b-poly(sodium acrylate) block copolymers with polystyrene block lengths of 11 and 23 units.12 This effect has been ascribed to the polyelectrolyte behavior of the ionic block. For small ionic block lengths, the solubility of the block copolymer molecules will increase with increasing ionic block length, resulting in an increase in the cmc. For long ionic block lengths, however, the solvent quality of the aqueous medium decreases with increasing ionic block length, which leads to the observed decrease in the cmc. A maximum in the cmc of the block copolymers based on 1a and 1b is found for a poly(L-lysine) block length of log(NPLys) ≈ 1.7-1.8 (n ≈ 60). The maximum cmc for polystyrene-b-poly(sodium

(11) Wilhelm, M.; Zhao, C.-L.; Wang, Y.; Xu, R.; Winnik, M. A.; Mura, J.-L.; Riess, G.; Croucher, M. D. Macromolecules 1991, 24, 1033-1040.

(12) Astafieva, I.; Khougaz, K.; Eisenberg, A. Macromolecules 1995, 28, 7127-7134.

Figure 3. Plot of the intensity ratio (I339/I334.5) as a function of the polymer concentration of PS10Lys52 (4e).

Polyelectrolyte Block Copolymers

Langmuir, Vol. 21, No. 14, 2005 6585

Table 2. Critical Micelle Concentrations (cmc), and Scattering Data Obtained on the PSmPLysn Block Copolymers sample 4a 4b 4c 4d 4e 4f 4g 4h 4i

PS8Lys9 PS8Lys22 PS8Lys29 PS10Lys39 PS10Lys52 PS10Lys59 PS10Lys63 PS10Lys67 PS10Lys72

Mna [g/ mol]

Mw/Mnb [-]

cmcc 106 [mol/L]

RHd [nm]

Rcoree [nm]

φSANS,cal c.f [M]

φSANS,expt.g [M]

2819 5523 6979 9269 11 973 13 428 14 261 15 093 16 133

1.15 1.26 1.30 1.44 1.43 1.20 1.36 1.17 1.39

3.09 7.41 4.17 1.48 2.40 2.57 2.29 2.57 2.19

70.8 75.8 76 71 76.0 67.0 101.4 79.4 78.5

2.5 2.5 2.0 2.5 2.5 2.5

5.39 × 10-4 4.18 × 10-4 3.72 × 10-4 3.51 × 10-4 3.22 × 10-4 3.10 × 10-4

5.97 × 10-4 5.25 × 10-4 4.57 × 10-4 3.78 × 10-4 3.29 × 10-4 3.23 × 10-4

a Molecular weight of the block copolymers based on the composition determined by 1H NMR spectroscopy. b Polydispersity (M /M ) of w n the protected polymers from GPC in DMF against a polystyrene standard. c Determined using the pyrene probe technique. d Hydrodynamic radius from dynamic light scattering (c ) 40 µg/mL). e Core radius from SANS. f Sample concentration for the SANS experiments as calculated from the weighted amounts of polymer. g Sample concentration calculated from the SANS results.

Figure 4. Plot of the critical micelle concentration (cmc) versus lysine block length for the different PSmPLysn block copolymers. 9, b, and O represent data obtained from block copolymers prepared from polystyrene macroinitiator 1c, 1b, and 1a, respectively.

Figure 5. Zimm representation of light scattering data obtained on aqueous solutions of PS10PLys63 (4g) at two different block copolymer concentrations.

acrylate) block copolymers with a polystyrene block consisting of 11 repeat units is observed for similar values of log(NPNaA).12 Light Scattering. Additional evidence for the aggregation of the PSmPLysn block copolymers is obtained from static light scattering experiments. Figure 5 shows a Zimm-type representation of the scattering data obtained for salt-free aqueous solutions of PS10PLys63 (4g) at pH ∼ 6.5 and at two concentrations, that is, 5 and 40 µg/mL. The different slopes of the lines connecting both sets of

data points indicate aggregate formation at the highest concentration, which is above the cmc for this sample (Table 2). The concentration dependence of the slope of the plots in Figure 5, however, makes it impossible to determine the molecular weight and aggregation number of the block copolymer assemblies by means of static light scattering. Figure 5 plots (Kc/R)/K versus q2. Since the refractive index increment could not be determined, this representation does not require knowledge of K (which, among other constants, contains the reactive index increment). Although the refractive index increment is a function of the sample concentration, the resulting changes in K in the investigated concentration regime will not affect the slopes of the lines through the data points to such an extent that they will become parallel. Form factor analysis of the data obtained at a block copolymer concentration of 40 µg/mL suggests that the aggregates are not spherical but rather are rods, cylinders, or ellipsoids. Further quantitative support for the nonspherical shape of the block copolymer aggregates is obtained from Holtzer13 and Kratky analysis14 of the static light scattering data. Figure 6 shows Holtzer and Kratky plots of the data for PS10PLys63 (4g). Both representations of the experimental data are in qualitative agreement with the curve shapes that would be expected for rodlike objects. In addition to static light scattering, dynamic light scattering experiments have been performed on salt-free aqueous solutions of the PSmPLysn block copolymers with a sample concentration of ∼40 µg/mL. The hydrodynamic radii (RH) obtained from these experiments are summarized in Table 2. With the exception of PS10PLys63 (4g), for which RH ≈ 100 nm, the hydrodynamic radii of the aggregates formed by the PSmPLysn block copolymers are in the range 70-80 nm. Small-Angle Neutron Scattering. Solutions of the PSmPLysn block copolymers in D2O at a concentration c ) 0.5 wt % have been investigated by SANS. The scattering curves have been converted to scattering data on an absolute scale following standard procedures.15 The SANS intensity I(q), from a solution of homogeneous infinite solid cylinders of radius rc and length L is, in units of cm-1, given by16,17

π I(q) ) np4L∆F2A2 [J1(qrc)/(qrc)]2 + BG q

(1)

where np ) NAφ/n j is the number density of scatterers (with NA the Avogadro number, φ the molar concentration, and n j the micellar aggregation number), A is the area of the (13) See, for example: (a) Holtzer, A. J. Polym. Sci. 1955, 17, 432434. (b) Berth, G.; Dautzenberg, H.; Christensen, B. E.; Harding, S. E.; Rother, G.; Smidrød, O. Macromolecules 1996, 29, 3491-3498. (c) Sommer, C.; Cannavacciulo, L.; Egelhaaf, S. U.; Pedersen, J. S.; Schurtenberger, P. Prog. Colloid Polym. Sci. 2000, 115, 347-352.

6586

Langmuir, Vol. 21, No. 14, 2005

Lu¨ bbert et al.

Figure 7. (O) SANS data obtained for a 0.25 wt % solution of PS10PLys52 (4e) at 25 °C. The full line represents the fit to the experimental data obtained using eq 3.

Figure 6. Holtzer (a) and Kratky (b) representation of light scattering data obtained for an aqueous solution of PS10PLys63 (4g) at a block copolymer concentration of 40 µg/mL. The inserts show the characteristic curve shapes that are expected for objects of different shapes.

cylinder cross section, ∆F2 is the excess scattering length density, and BG is the incoherent background. Using the expression np ) NAφ/n j , writing the volume of the cylinder V ) AL ) n j v (v: volume of one molecule) and using A ) πrc2, eq 1 can be expressed as

I(q) ) 4vφNA∆F2π2[J1(qrc)2/q3] + BG

(2)

To model the experimental data, the intensity expression described in eq 2 is convoluted with the resolution function for wavelength spread (∆λ/λ ) 0.1) and collimation effects for the SANS instrument.18 The corrected intensity Ic(q) (in cm-1) corresponds to the following expression:

Ic(q) ) 2πvφNA∆F2 σ

∫0

∞

J1(q′rc)exp{-0.5(q′ - q)2/σ2} dq′ + q′3 BG (3)

where σ includes the collimation and wavelength spread effects.18 Equation 3 is fitted to the experimental SANS (14) See, for example: (a) Imai, M.; Kaji, K.; Kanaya, T.; Sakai, Y. Phys. Rev. B 1995, 52, 12696-12704. (b) Mizukoshi, N.; Norisuye, T. Polym. Bull. 1998, 40, 555-562. (c) Doniach, S. Chem. Rev. 2001, 101, 1763-1778. (15) Lindner, P. In Neutrons, X-rays and Light Scattering; Zemb, T., Ed.; Elsevier: Oxford, U.K., 2002.

curves. The fitting parameters included in this model are ∆F, φ, rc, and BG. As a representative example, Figure 7 shows the experimental SANS data obtained for PS10PLys52 (4e) fitted with eq 3. A satisfactory fit can be obtained using rc ) 25 Å. This is ascribed to scattering from the polystyrene core of the micelles. There is significant contrast between D2O and both polystyrene and poly(Llysine), and in turn the contrast between these two is comparatively small. Nevertheless, in the q range assessed, the scattering appears to be dominated by the contribution from the core, and a “bare core” approximation may be invoked.19 The SANS results for all PSmPLysn block copolymers are summarized in Table 2. The core radii seem reasonable for a polystyrene block composed of only ∼10 repeat units20 and also suggest that only the polystyrene core of the cylindrical particles contributes to the scattering. Therefore, the excess scattering length density is fixed to ∆F ) 4.95 × 1010 cm-2, which corresponds to the difference in scattering length density between polystyrene (F ) 1.43 × 1010 cm-2) and D2O (F ) 6.38 × 1010 cm-2). As expected, the background remains almost constant for all the SANS curves, providing an averaged value BG ∼ 0.005. The molar concentrations obtained from the experimental SANS data are listed in Table 2. These values are in reasonable agreement with the calculated values, which are included in the same table. The agreement between the sample concentrations deduced from the SANS measurements and those calculated from the weight concentrations also supports the reasonableness of the assumption of cylindrical particles. Differences are probably due to experimental errors in the absolute scale normalization. Analytical Ultracentrifugation. To obtain further insight into the size of the aggregates, PS10PLys39 (4d), PS10PLys52 (4e), and PS10PLys63 (4g) have been studied with analytical ultracentrifugation (AUC). These experiments were carried out at sample concentrations of 0.40, 0.45, and 0.50 mg/mL. As an example, Figure 8 presents the apparent sedimentation coefficient distribution g*(s) calculated with SEDFIT for both the interference and the absorbance scans obtained for PS10PLys63 at a sample concentration of 0.40 mg/mL. The applicability of the (16) Porod, G. In Small-Angle X-ray Scattering; Glatter, O., Kratky, O., Eds.; Academic Press: London, 1982. (17) Cotton, J. P. In Neutrons, X-rays and Light Scattering; Lindner, P., Zemb, T., Eds.; Elsevier: Amsterdam, 1991. (18) Pedersen, J. S.; Posselt, D.; Mortensen, K. J. Appl. Crystallogr. 1990, 23, 321-333. (19) Plesˇtil, J. J. Appl. Crystallogr. 2000, 33, 600-604. (20) Khougaz, K.; Astafieva, I.; Eisenberg, A. Macromolecules 1995, 28, 7135-7147.

Polyelectrolyte Block Copolymers

Langmuir, Vol. 21, No. 14, 2005 6587

Figure 8. Apparent sedimentation coefficient distribution g*(s) for a 0.4 mg/mL solution of Ps10PLys63 (4g) in 0.05 M NH4Br. Top: absorbance optics (260 nm), bottom: Rayleigh interference optics (T ) 20 °C, 35 000 rpm, s: Svedberg ) 10-13 s).

evaluation is confirmed by the acceptable fitting (rms < 0.01).21 A bimodal distribution is clearly visible with maxima in the same range for the two optics. PS10PLys63 has been studied in several runs using freshly prepared solutions and also as a function of time over a period of three weeks with one run each week, and in all cases a bimodal apparent sedimentation coefficient distribution is found. In all experiments, two maxima, or at least one maximum and a clear shoulder, are identified at s1 ) 6.26.5 S and s2 ) 9.3-10 S for the interference optics and s1 ) 6.5-6.8 and s2 ) 9.7-10 S for the UV optics, respectively, with no significant concentration influence in the concentration range investigated. Taking the ratio of the apparent sedimentation coefficients s1/s2 ≈ 0.67, to estimate the mass ratio of the two fractions according to s ∼ M2/3, the relationship for spheres, a mass ratio of M1/ M2 ≈ 0.58 is obtained. From the values of the sedimentation coefficients and the partial specific volume of 0.693 cm3/g, the mass of the PS10PLys63 aggregates can be estimated to be in the range of several millions g/mol. This would correspond to aggregation numbers ranging between ∼60 (1 × 106 g/mol) and ∼600 (10 × 106 g/mol). Interestingly, the bimodal sedimentation coefficient distribution is also found for the two other investigated samples at all three concentrations with a slight shift of the maximum because of the different molecular weight of the block copolymer. At present, we do not have an explanation for the bimodal apparent sedimentation coefficient distribution. However, it clearly suggests that aqueous solution selfassembly of the polystyrenem-b-poly(L-lysine)n block copolymers does not result in a single population but generates two species that coexist. Dynamic light scattering experiments under conditions identical to those used for AUC revealed only one population of aggregates with RH ≈ 100 nm. This is in good agreement with the values listed in Table 2, which are obtained from solutions of much lower concentration. Since the ratio of masses of the two populations of aggregates observed in AUC are rather close (M1/M2 ≈ 0.58), it may be that, because of the dominance of the contribution from large and heavy scatterers, the bimodal size distribution is not obtained from DLS. Consequently, the hydrodynamic radius of 100 (21) Schuck, P. Biophys. J. 2000, 78, 1606-1619.

Figure 9. Molecular graphics representation of the proposed structure of the PSmPLysn aggregates.

nm obtained from DLS refers to the aggregates with the higher sedimentation coefficient. Aggregate Structure. A schematic molecular graphics representation of the proposed structure of the aggregates is shown in Figure 9 for PS10PLys52 (4e). Qualitative indications for the cylindrical shape of the aggregates are obtained from the Holtzer and Kratky representations of the static light scattering data, as well as from the SANS data, which could be fitted well assuming cylindrical scattering objects. According to the SANS experiments, the radius of the hydrophobic core is 2.0-2.5 nm. The hydrodynamic radius of the aggregates is 70-100 nm as determined by DLS. The representation in Figure 9 suggests that it is not unreasonable to assume that each “slice” of the cylindrical aggregates is composed of ∼16 block copolymer molecules. Taking 2RH as a measure for the cylinder length and using Rg ) 2 nm for the peptide block of the block copolymer22 suggests an aggregation number of ∼600 for PS10PLys63 (4g), which is within the same range as estimated from AUC. While electron or scanning probe microscopy could provide further, maybe even unequivocal, evidence on the structure of the block copolymer aggregates, our efforts in using these techniques have, unfortunately, not been successful so far. Conclusions In this contribution, we have studied the aqueous solution self-assembly of a series of amphiphilic PSmPLysn polyelectrolyte block copolymers. At neutral pH, when the poly(L-lysine) block has a random coil conformation, the cmc of the block copolymers varied between 1.5 and 7.4 × 10-6 M. Interestingly, the cmc shows a parabolic dependence on the ionic block length. For short poly(Llysine) block lengths, the solubility (and cmc) increases (22) For an unfolded protein, it has been proposed Rg ) (27.22 + 5 × Lr)0.5 (Å), where Lr is the total number of residues in the protein (in our case, the number-average degree of polymerization of the poly(L-lysine) block). See: Kundrotas, P. J.; Karshikoff, A. Phys. Rev. E 2001, 65, 011901.

6588

Langmuir, Vol. 21, No. 14, 2005

with increasing block length. In contrast, for longer poly(L-lysine) blocks, the solvent quality of the aqueous medium (and the cmc) decreases with increasing block length. The shape and size of the aggregates has been studied by a combination of dynamic/static light scattering, SANS, and AUC experiments. Interestingly, the polyelectrolyte block copolymers form nonspherical structures in aqueous solution, irrespective of poly(L-lysine) block length. Holtzer and Kratky representations of the static light scattering data provide a first qualitative indication for the nonspherical, presumably cylindrical, shape of the aggregates. This is supported by the SANS data, which, although they were obtained in a much higher concentration regime, could also be fitted well assuming cylindrical scattering objects. AUC experiments, which were carried out under conditions that more closely resemble those used for the SANS studies than the light scattering experiments, indicate a bimodal apparent sedimentation coefficient distribution, which is not understood at the moment. The AUC experiments, however, allow the mass of the aggregates to be estimated on the order of several millions g/mol, which, when using the hydrodynamic radius obtained from DLS as a measure for cylinder length (L ≈ 2RH), corresponds to aggregation numbers ranging from several 10s to several 100 block copolymer molecules. Experimental Section Materials and Methods. N-Benzyloxycarbonyl- L-lysine was purchased from Fluka. All other chemicals and solvents were obtained from Sigma Aldrich Chemie GmbH or ACROS. N,N′Dimethylformamide (DMF) and ethyl acetate were dried for at least 24 h over molecular sieves (4 Å). The synthesis of the amino functionalized polystyrene macroinitiator and the N-benzyloxycarbonyl-L-lysine N-carboxyanhydride (Z-Lys NCA) has been described previously.23,24 Spectra/Por dialysis bags of regenerated cellulose (Carl-Roth GmbH) were used for the purification of the deprotected block copolymers. All aqueous solutions were prepared with bidistilled water. 1H- and 13C NMR spectra were recorded on Bruker AMX 300 and DPX 700 spectrometers. As internal standard, the residual proton or carbon signal of the deuterated solvent (DMSO) was used. Chemical shifts are reported in parts per million (ppm) and peak multiplicities are described using the following abbreviations: bs, broad singlet; m: multiplet; s: singlet; d: doublet; t: triplet; and q: quartet. Gel permeation chromatography (GPC) in DMF was performed at 60 °C using a setup consisting of a Waters 510 pump and a series of three styrene-divinylbenzene columns (Polymer Standard Service, Mainz, Germany; column 300 × 8 mm; 500, 105, and 106 Å pore sizes). As the mobile phase, a 0.1 M solution of LiBr in DMF was used and sample elution was monitored with UV/vis and refractive index (RI) detection. Elution times were converted into molecular weights using a calibration curve constructed with narrow polydispersity polystyrene standards. An Applied Biosystems Biocad Sprint workstation equipped with a Novema column (column 300 × 8 mm, pore size 300 Å, 10-µm particle size) was used to perform GPC experiments with a mobile phase consisting of a mixture of acetonitrile/water/trifluoracetic acid (50:50:0.2 v/v/v). Flow rate was 1.0 mL/min and sample elution was monitored with a dual wavelength UV/vis-detector at 210 and 280 nm. Molecular weights were estimated using a calibration curve that was prepared using narrow sodium poly(styrene sulfonate) standards. Fluorescence spectra were recorded using a SPEX USA Fluorolog 2 Type F212 spectrophotometer. Samples were prepared by first adding a known amount of pyrene in acetone in a volumetric flask and subsequent evaporation of the acetone. Then, an aliquot of an aqueous block copolymer (23) Klok, H.-A.; Langenwalter, J. F.; Lecommandoux, S. Macromolecules 2000, 33, 7819-7826. (24) Lecommandoux, S.; Achard, M.-F.; Langenwalter, J. F.; Klok, H.-A. Macromolecules 2001, 34, 9100-9111.

Lu¨ bbert et al. solution with the appropriate block copolymer concentration was added to give a solution containing 5.9 × 10-7 mol/L pyrene. The resulting solution was stirred at room temperature for 24 h to equilibrate. The fluorescence spectra were measured in a 10 × 10 mm quartz cuvette with a path length of 1 cm and a gap width of 0.5 mm. The excitation wavelength for the emission spectra is 339 nm and excitation spectra were recorded at 390 nm. The integration time per data point and the step width were 0.5 s and 0.5 nm, respectively. Light scattering experiments were carried out with an ALV 5000-Korrelator, ALV-SP81-Goniometer using a Krypton-Ion-Laser 647.1 nm (Spectra Physics Model Kr 2025) and an Avalanche Photodiode Module. Samples for light scattering were prepared in bidistilled water and then were directly filtered over 0.45-µm syringe filters. Subsequently, the filtered solutions were equilibrated for a period of 24 h at room temperature prior to the experiments. SANS experiments were conducted at the SANS-I instrument of the Swiss Spallation Neutron Source SINQ (Paul Scherrer Institute, Switzerland). The scattering data was recorded using a two-dimensional detector. A wavelength λ ) 5 Å and a sample-detector distance of 4.5 m were chosen to cover scattering vectors in the range q ) (0.014-0.183) Å-1 (|q| ) 4π sin θ/λ, where the scattering angle is 2θ). Solutions of 0.5 wt % PSmPLysn were prepared by mixing the appropriate amount of polymer with D2O and storing for a period of days at low temperature (∼5 °C). Samples were then poured into sealed 1-mm-thick standard quartz cuvettes. Measurements were carried out at 20 °C. Temperature control was achieved using a water bath. The data were treated by standard methods to obtain scattering data on an absolute scale.15 An Optima XL-I Analytical Ultracentrifuge (Beckman, Palo Alto, CA) equipped with scanning UV/vis and Rayleigh interference optics was used for sedimentation velocity studies at 35 000 rpm and 20 °C. Radial UV scans were recorded at a wavelength of 260 nm. The block copolymer was first dissolved in pure water (Millipore, >18 MΩ), then diluted 1:1 with 0.1 M NH4Br aqueous solution, and dialyzed against 0.05 M NH4Br aqueous solution. Sedimentation velocity experiments were performed with polymer sample concentrations of 0.5, 0.45, and 0.4 mg/mL diluted from dialyzed stock solution and solvent. The concentrations corresponded to fringe detection in the range of 1.2-0.8 and UV absorbance of 0.1-0.08 if a standard double sector cell was used. Sedimentation velocity scans were evaluated using SEDFIT.21

Procedures Preparation of the Polystyrene-b-poly(N-benzyloxycarbonyl-L-lysine) (PSmPZLysn) Block Copolymers. A dry Schlenk-tube fitted with a stir bar and a drying tube is charged with a calculated amount of Z-Lys NCA in DMF (c ≈ 0.2 g/mL). Then, the appropriate amount of a DMF solution of the amine functionalized polystyrene macroinitiator is added and the mixture is stirred at room temperature. After 5 days, the reaction mixture is precipitated in a 15-fold (v/v) excess of water. Polymerizations carried out with polystyrene amine hydrochloride as the macroinitiator were performed at 50 °C over a period of 3 days. The precipitated polymer is isolated by filtration, washed extensively with diethyl ether to remove unreacted monomer, and finally dried under vacuum. Polystyrene-b-poly(N  -benzyloxycarbonyl- L -lysine) (PSmPZLysn). 1H NMR (700 MHz, DMSO, 333 K): δ (ppm) ) 0.49-0.75 (m, 6 H, -CH3); 0.90-2.15 (m, (9 + 3 × 10 + 6N) H, CH3-CH + CH3-CH2 + aliph. oligostyrene + RCH-(CH2)3); 2.96 (s, 2N H, RCH(CH2)3-CH2); 3.65-4.38 (m, N H, RCH); 4.97 (m, 2N H, Z CH2); 6.28-7.38 (m, (5N + N) H, RCH-NH + arom. CH); 8.13 (bs, N H, RCH(CH2)4-NH). Preparation of the Polystyrene-b-poly(L-lysine) (PSmPLysn) Block Copolymers. The Z-protected precursor block copolymer is dissolved in as little as possible TFA and a 4-fold excess (per lysine repeat unit) of a 33 wt % solution of HBr in AcOH is added. After stirring for 1 h at room temperature, the reaction is terminated by the addition of diethyl ether, which causes precipitation of the block copolymer. The crude deprotected block copolymers are repeatedly dialyzed against bidistilled water until the conductivity remains constant and are finally isolated via lyophilization. Polystyrene-b-poly(L-lysine hydrobromide) (PSmPLysn). 1H NMR (700 MHz, DMSO, 333 K): δ (ppm) ) 0.48-0.75 (m, 6

Polyelectrolyte Block Copolymers H, CH3); 1.24-1.80 (m, (9 + 10 × 3 + 6N) H, CH3-CH + CH3CH2 + aliph. oligostyrene + RCH-(CH2)3); 2.77 (bs, 2N H, RCH (CH2)3-CH2); 4.28 (m, N H, RCH); 6.33-7.34 (m, 5N H, arom. CH); 7.92 (bs, N H, RCH-NH).

Acknowledgment. This work was partially supported by the Deutsche Forschungsgemeinschaft (H.-A.K.;

Langmuir, Vol. 21, No. 14, 2005 6589

KL1049/2-1-/2-5 (Emmy Noether Program) and KL1049/ 3-1, -/3-2) and EPSRC (I.W.H., V.C.). We thank Steven van Petegem for assistance in SANS experiments at the SINQ facility of the PSI (Villigen, Switzerland). LA0502600