ARTICLE pubs.acs.org/Langmuir
Nanofibers Resulting from Cooperative Electrostatic and Hydrophobic Interactions between Peptides and Polyelectrolytes of Opposite Charge Matthias Ferstl,† Andrea Strasser,‡ Hans-Joachim Wittmann,§ Markus Drechsler,|| Matthias Rischer,^ J€urgen Engel,# and Achim Goepferich*,† Department of Pharmaceutical Technology, ‡Department of Pharmaceutical/Medicinal Chemistry II, and §Faculty of Chemistry and Pharmacy, University of Regensburg, 93053 Regensburg, Germany Macromolecular Chemistry II, University of Bayreuth, 95440 Bayreuth, Germany ^ IPhaTec, 60388 Frankfurt am Main, Germany # Pharmaceutical Development, Aeterna Zentaris GmbH, 60314 Frankfurt am Main, Germany
)
†
bS Supporting Information ABSTRACT: We investigated whether cationic peptides that contain hydrophobic side chains were able to stabilize themselves via hydrophobic interactions between neighboring peptide molecules upon electrostatic binding to oppositely charged polyelectrolytes. The interaction mechanism was examined through a model system consisting of the anionic polyelectrolyte alginate and the cationic decapeptide ozarelix. The interaction resulted in the formation of highly ordered complexes that were noticeable upon visual inspection. These complexes were then investigated by microscopic techniques and shown to exhibit a branched network structure. Cryogenic-temperature transmission electron microscopy (cryo-TEM) and negative staining TEM revealed that the molecular interactions between alginate and ozarelix led to the formation of nanofibers. The rodlike nanofibers had a diameter distribution of 48 nm. Isothermal titration calorimetry was used to determine the thermodynamic parameters of the alginateozarelix interaction. The binding constant was found to be on the order of 106 M1, indicating a high binding affinity. The interaction of the peptide with the polyelectrolyte triggered profound changes in the conformation of ozarelix, which was confirmed by UV spectroscopy and circular dichroism. On the basis of these experimental results, a theoretical modeling study of the alginateozarelix interaction was conducted to gain a better molecular-level understanding of the complex structure. It revealed that, upon binding of ozarelix to alginate, new intermolecular and intramolecular aromatic interactions between the ozarelix molecules occurred. These interactions changed the conformation of the peptide, a modification in which the aromatic side chains played a major role. Our results indicate that the cationic peptides interact with the polyanions via electrostatic interactions, but are additionally stabilized via hydrophobic interactions. This binding mode may serve as a powerful tool to extend the duration of drug release in hydrogel drug delivery systems.
’ INTRODUCTION In 2008, the Pharmaceutical Research and Manufactures of America report on biotechnology medicines in development listed 633 biotechnology drugs, known as biopharmaceutics, that were under development for more than 100 diseases.1 These included 254 cancer therapies, 162 drug targets for infectious diseases, 598 for the treatment of autoimmune diseases, and 34 for the treatment of HIV/AIDS and related conditions (some medicines are listed in more than one category). The majority of these drugs were within the subset of peptide and protein therapeutics. In contrast to this rapid progress in the development of biopharmaceutics, the formulation of peptide and protein drugs to efficient delivery systems has yet to show similar advancement. Successful delivery systems help to optimize pharmacokinetics, minimize degradation, prevent harmful systemic side effects, increase drug bioavailability, and reduce injection r 2011 American Chemical Society
frequencies; the current failure to develop effective formulations is a significant obstacle to the widespread success of these agents. Specifically, many of these formulation strategies fail due to the unique requirements and restrictions that peptides and proteins pose compared to conventional drug compounds.2,3 Hydrogels are a proposed platform that may assist in closing this technical hurdle.4 However, hydrogels' rapid drug release serves as a current limitation to their utility.5 Different strategies have been explored to reduce the mobility of proteins and peptides in hydrogels and to slow their release.6 One method relies on enhancing interactions in the drughydrogel network via the electrostatic interactions between ionic polymers and oppositely Received: June 15, 2011 Revised: October 13, 2011 Published: October 17, 2011 14450
dx.doi.org/10.1021/la202252m | Langmuir 2011, 27, 14450–14459
Langmuir
ARTICLE
’ EXPERIMENTAL SECTION
Figure 1. Chemical structure of alginate, which consists of β-D-mannuronic acid (M) and α-guluronic acid (G) residues (A). Primary structure of the decapeptide ozarelix, where the hydrophobic amino acid residues are blue and the hydrophilic residues are red (B).
charged proteins.79 Even though this strategy was somewhat successful, we felt that further optimization was possible. We therefore began investigating the possibility of additional molecular interaction mechanisms that could be used for binding protein and peptide drugs to the hydrogel network and decreasing their mobility. Peptide and protein drugs contain pronounced nonpolar regions, allowing them to interact with their own species or other molecules via hydrophobic interactions.10 We suspect that this additional force may further stabilize interactions and extend the duration of drug release.11 It was our goal to investigate whether peptides are able to undergo electrostatic binding to polymer strands and stabilize themselves further via hydrophobic interactions between neighboring peptide molecules. The principle of this interaction was investigated on the basis of a chosen model system consisting of an anionic polyelectrolyte and a positively charged peptide. We were specifically interested in the exact molecular binding mechanism of such an interaction in the resulting complexes. We selected sodium alginate as the model anionic polyelectrolyte due to its well-characterized molecular architecture. It is a linear binary copolymer composed of (1f4)-linked β-D-mannuronic acid (M) and α-L-guluronic acid (G) residues (Figure 1A). The molecule carries one negative charge per monomer. These monomers can appear in homopolymeric blocks of G residues (G-blocks) and M residues (M-blocks), as well as in blocks containing an alternating sequence of both M and G units (MGblocks). Alginate is also a hydrophilic, biocompatible12 polysaccharide and is therefore optimal for a wide variety of pharmaceutical applications.13,14 We chose the decapeptide ozarelix acetate, a fourth-generation luteinizing hormone-releasing hormone (LHRH) receptor antagonist, as the model peptide. It contains a single positive charge that stems from a guanidine group, which allows for electrostatic interaction with alginate (Figure 1B). Additionally, ozarelix contains a considerable number of hydrophobic amino acids, making it a good candidate to demonstrate, upon binding to alginate, our proposed stabilization effect via hydrophobic interactions.
Materials and Methods. Ozarelix acetate (O) (N-acetyl-3-(2naphthalenyl-)-D-alanyl-4-chloro-D-phenylalanyl-3-(3-pyridinyl)-D-alanylL-seryl-N-methyl-L-tyrosyl-N6-(aminocarbonyl)-D-lysyl-L-norleucyl-Larginyl-L-prolyl-D-alaninamide), molecular weight 1459.09, was kindly supplied by Aeterna Zentaris GmbH (Frankfurt am Main, Germany) and used as received. Ultrapure sodium alginate was purchased from Novamatrix (Sandvika, Norway). Two types of ultrapure sodium alginate (A) of low viscosity were used in this study: Pronova UP LVM, a high mannuronic acid content alginate (high-M alginate) composed of a G/M ratio of e1, and Pronova UP LVG, a high guluronic acid content (highG alginate) composed of a G/M ratio of g1.5. According to the manufacturer’s specifications, both alginates had an approximate molecular mass of 75200 kDa. Preparation of PolyelectrolytePeptide Complexes. Aqueous stock solutions of sodium alginate and ozarelix were prepared by dissolving the components in deionized water (Millipore SAS, Molsheim, France) under magnetic stirring at room temperature. After filtration of the resulting stock solutions through a 0.2 μm sterile poly(ether sulfone) syringe filter (Corning Inc., Corning, NY), 1 mL of the ozarelix solution was transferred to a 2 mL Eppendorf tube. A 1 mL volume of the respective alginate solution was then added dropwise. The solution was gently vortexed (Vortex-T Genie 2, Scientific Industries, Bohemia, NY) for 5 s and then shaken for 24 h on a shaker (Edmund B€uhler GmbH, Hechingen, Germany) at level 4. Light and Polarized Light Microscopy. The microstructure of the AO complex was analyzed via microscopic techniques. Isoelectric AO (1/6) samples (0.1% (m/v), 0.767 mM alginate/0.685 mM ozarelix) were prepared as described above. A 20 μL volume of a sample was transferred to a microscope slide and covered with a coverslip to ensure slow drying. The samples were dried overnight at room temperature. Microscopic pictures were acquired using a Leica DM IRB/E microscope (Leica GmbH, Wetzlar, Germany) equipped with a Nikon Digital Sight DS-U1 camera (Nikon Corp., Tokyo, Japan) and processed with the software program EclipseNet (Laboratory Imaging, Prague, Czech Republic). Light micrographs for the samples were taken with crossed polarized filters. TEM Imaging. For cryo-transmission electron microscopy (cryoTEM) studies, a sample droplet of 2 μL was put on a lacey carbon film covered copper grid (Science Services, Munich, Germany), which was hydrophilized by glow discharge for 15 s. Most of the liquid was then removed with blotting paper, leaving a thin film stretched over the lace holes. The specimens were instantly shock frozen by rapid immersion into liquid ethane and cooled to approximately 90 K by liquid nitrogen in a temperature-controlled freezing unit (Zeiss Cryobox, Zeiss NTS GmbH, Oberkochen, Germany). The temperature was monitored and kept constant in the chamber during all the sample preparation steps. After the specimens were frozen, the remaining ethane was removed using blotting paper. The specimen was inserted into a cryo transfer holder (CT3500, Gatan, Munich, Germany) and transferred to a Zeiss EM922 energy-filtered TEM (EFTEM) instrument (Zeiss NTS GmbH, Oberkochen, Germany). Examinations were carried out at temperatures around 90 K. The TEM instrument was operated at an acceleration voltage of 200 kV. Zero-loss-filtered images (ΔE = 0 eV) were taken under reduced dose conditions (1001000 e/nm2). All images were recorded digitally by a bottom-mounted charge-coupled device (CCD) camera system (Ultra Scan 1000, Gatan, Munich, Germany) and combined and processed with a digital imaging processing system (Digital Micrograph GMS 1.8, Gatan, Munich, Germany). The single components of the complex were investigated alone as a blank trial. Negative staining TEM specimens were prepared by depositing 2 μL of the sample solution onto a copper grid and afterward removing the excess with filter paper. Then 2 μL of the staining solution was added, 14451
dx.doi.org/10.1021/la202252m |Langmuir 2011, 27, 14450–14459
Langmuir
ARTICLE
Figure 2. Light microscopy picture of AO aggregates (alginate, 0.767 mM; ozarelix, 0.685 mM) showing a branched network structure (A). Higher magnification of the marked area (B). Picture of the AO aggregate and corresponding polarized light picture that exposed birefringence (C, D).
Figure 3. Cryo-TEM pictures of the AO sample (alginate, 0.767 mM; ozarelix, 0.685 mM) exhibiting the formation of nanofibers beside larger aggregates. and after ∼30 s, the excess was removed with filter paper. The negative stain was a phosphotungstic acid solution (1% (w/v), pH 7). After negative staining, the grids were allowed to dry at room temperature for several minutes, photographed with a Zeiss EM922 EFTEM instrument , and processed with ImageJ software (National Institutes of Health, Bethesda, MD). Isothermal Titration Calorimetry (ITC). The ITC experiments were carried out using a VP-ITC MicroCalorimeter (MicroCal, Northhampton, MA).15 For the ITC experiments, only deionized water was
used, without any buffer added. To exclude the possibility of bindinglinked (de)protonation, we measured the pH of the system, which was always 6. The pH of the system did not change when we dissolved the sodium alginate, the peptide, or a mixture of both. At this pH the carboxy groups of alginate (pKa value of 3.44) are completely deprotonated while the arginine group in the side chain of ozarelix (pKa value of 12.1) is protonated. All solutions were thoroughly degassed under vacuum for 15 min under gentle stirring prior to use. The sample cell was loaded with 1.436 mL of a 0.1 mM ozarelix solution. The syringe was filled with 14452
dx.doi.org/10.1021/la202252m |Langmuir 2011, 27, 14450–14459
Langmuir
ARTICLE
0.3 mL of a 1 mM alginate solution, with the molar concentration of alginate calculated on the basis of the molecular mass of the monomers. During the experiment the solution in the cell was stirred at 300 rpm to ensure rapid mixing. Each experiment consisted of 60 injections of 5 μL of titrant over 5 s with a 2 min time interval between injections, thereby allowing for complete equilibration. All measurements were carried out at 25 ( 0.1 C. As a control experiment, alginate was titrated into water to exclude the possibility of major dilution effects (Figure S1, Supporting Information). The calorimetric data were analyzed by nonlinear fitting using MicroCal Origin software (MicroCal). The thermodynamic parameters standard reaction enthalpy (ΔH), binding constant (K), and standard reaction entropy (ΔS) and the values for the binding stoichiometry (n) were calculated by using the one set of binding sites model for data fitting. Although this mode of analysis may not be ideal for a cooperative system, it was found to be a simple approach that produced robust results. Prior to fitting, the average value of the posttransition heat transfer at the end of the titration was subtracted from the data. With the determined thermodynamic parameters, the Gibbs standard reaction energy (ΔG) was calculated by using the following equation: ΔG ¼ ΔH TΔS
ð1Þ
Circular Dichroism (CD). The CD spectra were recorded with a Jasco-J-810 spectropolarimeter (Jasco GmbH, Gross-Umstadt, Germany) at 25 C under a constant flow of nitrogen gas during operation. Using a quartz cell with a 1 mm path length, spectra were recorded from 190 to 350 nm, with a bandwidth of 1 nm, a scan speed of 100 nm/min, and a response time of 1 s. All spectra were recorded in quintuplicate and averaged. The samples were prepared by mixing 1 mL of a 0.137 mM ozarelix solution with 1 mL of alginate solution at the concentration required for the desired AO ratio. Every spectrum was corrected for the spectrum of pure solvent. UVVis Spectroscopy. UV spectra were recorded using an Uvikon 941 spectrophotometer (Kontron Instruments, Basel, Switzerland). The scan interval was 190350 nm and the scanning speed 200 nm/ min. Quartz cuvettes of 1 cm path length were used. The sample preparation was identical to that of the CD experiments. Modeling. Construction of the Alginate and Decapeptide, Force Field Parameters, Simulation Software, and Simulation Parameters. Alginate and ozarelix were constructed with SYBYL 7.0 (Tripos, St. Louis, MO). The force field parameters were calculated using the PRODRG server,16 choosing the options “chirality on” and “full charges” and the “gromos96.1” force field. For water molecules, the single point charge (SPC) model was used.17 All minimizations and molecular dynamic simulations were performed with GROMACS 3.3.1.18 For all molecular dynamic simulations, we used the Berendsen pressure and temperature coupling.19 The time step was set to 1 fs. Rectangular boundary conditions were used, and the particle mesh Ewald method20 was applied to describe the electrostatic interactions. The distances for Coulomb and LennardJones cutoffs were set to 1 nm. The nonbonded atoms pair list was updated every 1 fs. Simulation of Ozarelix in Water. The positively charged decapeptide was embedded in a water box (2.98 nm 2.98 nm 2.98 nm) containing 819 water molecules and 1 chlorine anion to achieve electroneutrality. After minimization, a 10 ns molecular dynamics (MD) simulation without any constraints was performed. Construction and Simulation of the AlginateOzarelix Complex. To simplify the complex interaction between alginate and ozarelix in this first approach, our first model used a molar binding stoichiometry of AO = 1/1 and an alginate chain composed entirely of mannuronic acid monomers. Initially, two models for the alginateozarelix complex were constructed. In the first model, the alginate, containing 11 mannuronic acid monomers, was constructed linearly. Afterward, 11 decapeptides
Figure 4. (AC) Cryo-TEM micrographs of AO (1/1) nanofibers at higher dilution (alginate, 0.0767 mM; ozarelix, 0.0685 mM). The pictures reveal the diameter of the nanofibers in their native hydrated state of 48 nm. The lengths of the fibers vary considerably. (DF) Negative staining TEM pictures of the self-assembled nanofibers, where some of the fibers are arranged in ribbonlike parallel arrays. Samples were negatively stained with phosphotungstic acid. were positioned manually around the alginate in such a way that the positively charged arginine side chain could establish an electrostatic interaction with the negatively charged carboxyl moieties of the alginate. However, within this model, an appropriate closed peptide layer around the alginate could not be established. Thus, this model was not further investigated. In our second model, the alginate, again containing 11 mannuronic acid monomers, was modeled as a helix.21 The helix was constructed in such a way that the negatively charged carboxyl moieties of the alginate pointed outward. Afterward, 11 decapeptides were positioned manually around the alginate in such a way that the positively charged arginine side chain could establish an electrostatic interaction with the negatively charged carboxyl moieties of the alginate. To obtain a closed peptide layer around the alginate, a 10 ps MD simulation with position constraints onto the whole alginate (5000 kJ/(mol nm2)) and the arginine side chains (5000 kJ/(mol nm2)) of each peptide was performed in the gas phase. The resulting system was embedded in a water box (3.98 nm 6.02 nm 6.02 nm) containing 3962 water molecules. After minimization, a 10 ns MD simulation was performed. The simulation was divided into six cycles: Cycles 15 (each 1 ns) were performed with position restraints onto the whole alginate and the arginine side chains (cycle 1, 5000 kJ/(mol nm2); cycle 2, 2500 kJ/(mol nm2); cycle 14453
dx.doi.org/10.1021/la202252m |Langmuir 2011, 27, 14450–14459
Langmuir
ARTICLE
Figure 5. ITC data of ozarelixPronova UP LVM (A) and of ozarelixPronova UP LVG (B). The bottom panel shows the binding isotherm created by plotting the areas under the peaks in the top panel against the molar ratio of alginate added to ozarelix present in the cell.
Table 1. Thermodynamic Parameters for the Interaction of Ozarelix and Alginatea parameter
high-M alginate
high-G alginate
n
0.661 ( 0.02
0.851 ( 0.08
K (M1)
3.68 106 (
1.16 106 (
5.66 10 3650 ( 62
1.88 105 3367 ( 32
17.8 ( 0.2
16.44 ( 0.43
3
1
a
ΔH (cal 3 mol ) ΔS (cal 3 mol1 3 K1)
The data presented are the average of two experiments.
Figure 7. Section of the UVvis absorbance spectra of ozarelix (c = 0.00685 mM) in the presence of different alginate concentrations between 190 and 250 nm.
’ RESULTS AND DISCUSSION
Figure 6. CD spectra of ozarelix (0.0685 mM) in the presence of different alginate concentrations (0.460.0092 mM) showing that the peptide interactions vary with the peptide/alginate ratio. 3, 1000 kJ/(mol nm2); cycle 4, 500 kJ/(mol nm2); cycle 5, 100 kJ/(mol nm2)). In cycle 6 (5 ns), the simulation was performed without any position restraints.
Light and Polarized Microscopy. In the first set of experiments, isoelectric AO mixtures based on the positive charges of ozarelix and the negative charges of alginate were examined. This interaction resulted in a white flocculent precipitate. These aggregates were investigated in the solid state by light microscopy. The pictures revealed that the AO aggregates had a branched network structure (Figure 2AC). The aggregates varied in size and shape and formed a network with a backbone to which other parts were bound. When the structures were investigated under polarized light, the branched network showed birefringence phenomena (Figure 2D), indicating the presence of structures with defined higher order. 14454
dx.doi.org/10.1021/la202252m |Langmuir 2011, 27, 14450–14459
Langmuir
ARTICLE
Figure 8. Simulation snapshots of ozarelix in water (AC). Number of hydrogen bonds in ozarelix itself compared to ozarelix surrounded by water (D). Distance of aromaticaromatic interactions within the decapepetide (E). Calculation of the angle α between the aromatic planes of PhClPhOH and PyrNaph (F).
TEM Imaging. Transmission electron microscopy was used to elucidate the ultrastructure of the AO aggregates. To preserve the native hydrated state of the material, all samples were investigated by cryo-TEM. Alginate and ozarelix alone did not show any visible structure under the transmission electron microscope (data not shown). At low resolution AO mixtures formed pronounced nanofibers beside larger aggregates (Figure 3), which explained the branched network structure that was observed with light microscopy. To investigate the nanofibers at higher resolution, highly diluted solutions of the AO complexes (0.0767 mM alginate/0.0685 mM ozarelix) were examined. Figure 4A shows a typical cryo-TEM image of individual AO nanofibers, which appeared rigid and rodlike, with a diameter of approximately 48 nm. Their length varied considerably; some of them were larger than 100 nm (Figure 4B and C). Supplementary negative staining experiments confirmed the cryo-TEM results (Figure 4DF). Here, some of the fibers arranged themselves in ribbonlike parallel arrays, most
likely due to the drying process. This may explain the birefringence the dried complexes showed under the polarized light. Isothermal Titration Calorimetry. Overall, the AO complexes suggest that a combination of electrostatic interactions in combination with hydrophobic forces was the reason for the observed data. To further substantiate this hypothesis, isothermal titration calorimetry was used to assess the binding strength of the AO interaction. First, we used the method to discriminate between different kinds of alginates. To see whether the alginate composition had an influence on the AO interaction, the thermodynamic parameters of two different alginate types, with either high guluronic acid (high-G alginate) or high mannuronic acid (high-M alginate) content, were examined. The titration thermograms shown in Figure 5 reveal that there is a strong interaction between A and O. The respective thermodynamic data are shown in Table 1. The results indicate that the composition of the alginate may influence the interaction with the decapeptide, since the binding constant K for high-M 14455
dx.doi.org/10.1021/la202252m |Langmuir 2011, 27, 14450–14459
Langmuir
Figure 9. Schematic representation of the proposed principle of the alginateozarelix interaction. Upon the electrostatic interaction with alginate, ozarelix can stabilize itself via hydrophobic interactions between neighboring peptide molecules.
alginate was higher than for high-G alginate. The use of the onesite binding model, however, fails to yield a definitive answer. More thorough studies are necessary to confirm this result and elucidate the source of these differences. All binding constants were on the order of 106, indicating strong binding.22 Cyclodextrin complexes, for example, have binding constants on the order of 103.23 The free energy values (ΔG) calculated for high-M alginate and for high-G alginate were 8.96 and 8.27 kcal 3 mol1, respectively. Both these values are higher than the binding energy of the purely electrostatic interaction between the anionic sites of heparin and the cationic groups of synthetic peptides, which showed free energy values between 6 and 7 kcal mol1.24 This clearly demonstrates that the binding affinity in our system is increased by the additional hydrophobic interactions. The large negative values recorded for the Gibbs free energy (ΔG) demonstrate that the binding process of ozarelix and alginate is strong. The negative enthalpy (ΔH) and positive entropy (ΔS) values of the interaction suggest that the electrostatic interaction plays a major role in the binding reaction.25 This hypothesis was supported by a further ITC experiment where no reaction between alginate and ozarelix occurred in the presence of 10 mM sodium chloride under otherwise identical conditions (Figure S2, Supporting Information). However, the positive sign of ΔS suggests that the hydrophobic interactions were involved in this interaction as well. In all the experiments, we observed toward the end of the reaction, when complex formation was complete, endothermic heat effects. We hypothesize that the endothermic response of the system could be due to a redistribution of the peptide from more densely packed to nonoccupied alginate chains upon each injection. This effect will also occur at the beginning of the titration but is masked by exothermic processes. Circular Dichroism. On the basis of the ITC experiments, we hypothesized that electrostatic interactions in combination with hydrophobic forces are responsible for complex formation. The interaction of the hydrophobic portions of the peptide could be accompanied by a spatial rearrangement of the peptide itself. Therefore, a conformational change of the peptide would further strengthen the hypothesis that hydrophobic forces participate in the interaction. Since CD spectra in the UV region can be used to monitor such changes,26,27 we employed CD measurements to determine the effect of alginate binding on the conformation of the decapeptide. Alginate was not detectable by CD and yielded no signal besides baseline noise (Figure S3, Supporting
ARTICLE
Information). The CD spectra of ozarelix showed a negative peak around 227 nm. The complex of an AO = 1/1 mass ratio showed a positive peak at 235 nm with a shoulder at 221 nm and therefore presented a reading completely different reading from that of ozarelix (Figure S3, Supporting Information). This result indicates that there is indeed a substantial change in the conformation of ozarelix due to the interaction with the polysaccharide. As CD spectra of small peptides can be difficult to interpret,28 we chose to run a molecular modeling study to obtain a detailed picture of the structural changes of the peptide during binding (see below). In another set of experiments, the spectra of different AO concentration ratios ranging from an alginate to an ozarelix surplus were recorded to see whether the interactions vary with the AO ratio (Figure 6). The ozarelix concentration was kept constant, while the concentration of alginate was continuously decreased. The first three mixtures (0.460.0767 mM alginate) showed an identical profile. At an ozarelix surplus, the intensity of the spectra began to decrease. At low alginate concentrations, the corresponding spectra approached that of ozarelix alone. This indicated that with less polyelectrolyte available for the interaction, the influence on the conformation of the peptide decreased. UVVis Spectroscopy. To confirm the conformational change of ozarelix in the presence of alginate, we measured the UV spectra of AO mixtures. The interaction of alginate and ozarelix also led to a change in the UV absorbance (Figure S4, Supporting Information). While alginate yielded no signal, ozarelix showed two bands at 225 and 267 nm. Upon addition of alginate, the peak at 225 nm decreased. With decreasing alginate concentration, the absorbance at 225 nm aligned with that of ozarelix alone (Figure 7). This agrees with the result of the conformational changes suggested from circular dichroism data. Modeling. On the basis of the experimental results, a modeling simulation study was undertaken. This was done to both see whether the determined diameter for the nanofibers of 48 nm by cryo-TEM would agree with the theoretical calculated value and provide a more detailed view of the ozarelix molecule itself during its interaction with alginate. This model would then hopefully yield an explanation for the spectroscopic results, which showed a conformational change of ozarelix induced by alginate. We wanted to determine the structural components that are responsible for this conformational change. First, we investigated the conformation of ozarelix in water. Some results of the MD simulation of the decapeptide ozarelix in water are shown in Figure 8. During the 10 ns MD simulation, a high flexibility for the decapeptide was observed. In Figure 8AC, three representative snapshots of the decapeptide during the simulation in the water box are given. During the whole simulation, the peptide showed, on average, one hydrogen bond within the peptide itself, whereas an average of approximately 30 hydrogen bonds could be detected between the peptide and surrounding water molecules (Figure 8D). To determine stacked aromatic interactions, the distance between the mass center of two aromatic groups was calculated (Figure 8E). These calculations revealed that the distances were smaller than 0.5 nm in some phases of the simulation only for the interactions between phenyl chloride (PhCl) and hydroxphenyl (PhOH) and pyridine (Pyr) and naphthalene (Naph). For all other possible aromaticaromatic interactions, the distances were significantly larger than 0.5 nm. For further verification of stacked aromatic interaction, the angle 14456
dx.doi.org/10.1021/la202252m |Langmuir 2011, 27, 14450–14459
Langmuir
ARTICLE
Figure 10. Proposed model of the alginateozarelix complex schematically pictured (A) and as a snapshot (B). Snapshots of intramolecular (C) and intermolecular (D, E) interactions of the aromatic chromophores in ozarelix. Surface of the ozarelixalginate complex (red, hydrogen bond donor; blue, hydrogen bond acceptor) and water molecules located at the surface (F). Number of hydrogen bonds during the productive phase of molecular dynamics simulation between ozarelix and solvent (Osol), alginate and solvent (Asol), ozarelix and ozarelix (OO), and ozarelix and alginate (OA) within the whole simulation box (G). Snapshots of hydrogen bond networking: hydrogen bond network between two different ozarelix molecules (H), hydrogen bond network mediated by two water molecules within one ozarelix molecule (I), hydrogen bond network mediated by two water molecules between two different ozarelix molecules (J).
α between the corresponding aromatic planes was calculated (Figure 8F). For d e 0.5 nm and α ≈ 0 or α ≈ 180, a stacked aromatic interaction is approximately established. Snapshots of the interaction for the PhClPhOH or PyrNaph are given (Figure 8B,C). On the basis of this information, we simulated the structure of the alginateozarelix complex. The stoichiometric ratio n evaluated
by ITC hinted at a possible reaction model. To simplify the model, we assumed that one positive charge of ozarelix would bind to one negative charge of alginate. A schematic picture of the proposed interaction is presented in Figure 9. Some results of the MD simulation of the complex in water are shown in Figure 10. As already observed for the decapeptide in water, the decapeptides in the alginateozarelix complex showed a high flexibility 14457
dx.doi.org/10.1021/la202252m |Langmuir 2011, 27, 14450–14459
Langmuir during the molecular dynamic simulation. The alginate helix had a mean diameter of about 0.7 nm, and the whole alginateozarelix complex had a mean diameter from 4 to 6 nm (Figure 10A,B). This result agrees very well with the cryo-TEM results. The observed variability may be explained in part by the physical structure of the alginate molecule. The monomers of alginate are arranged heterogeneously in a series of blocks, consisting of M-, G-, and MG-blocks, leading to different space requirements.29 This was confirmed by an imaging study of alginate polymers which showed a width between 1.41 and 4.65 nm when examined by atomic force microscopy.30 During the molecular dynamics simulations, a systematic arrangement of the decapeptides could not be detected. Compared to ozarelix in water, however, significant differences in aromatic interactions could be observed (Figure 10CE). Within the alginatedecapeptide complex, aromatic intramolecular interactions were established within one decapeptide molecule. In Figure 10C, a snapshot with a stacked aromatic interaction between Naph and PhCl is shown. The distance between both aromatic moieties is around 0.45 nm. In addition, intermolecular aromatic interactions were established between two (Figure 10D) or even three (Figure 10E) decapeptides. Similar stacked aromatic interactions within one decapeptide or three decapeptides, or more often between two decapeptides, were observed during the whole simulation. These new aromatic interactions induced by the alginate presence might explain the detected differences in the CD and UVvis spectra of ozarelix alone in comparison to the ozarelixalginate complex. As shown in Figure 10F, water molecules not only are located on the surface of the alginateozarelix complex, but also occupy the channels within the complex. A mean number of 300 hydrogen bonds between the 11 ozarelix molecules and solvent are found at each time step of the simulation (Figure 10G, Osol). A mean number of 67 hydrogen bonds are found between the alginate and the solvent (Figure 10G, Asol), and a mean number of 9 direct hydrogen bonds are seen between ozarelix molecules (Figure 10G, OO). Between the ozarelix and alginate, only 4 hydrogen bonds were detected (Figure 10G, OA). In general, the hydrogen bond network stabilizes the ozarelixalginate complex due to direct hydrogen bond interactions between two different ozarelix molecules (Figure 10H) and water-mediated hydrogen bond interactions within one ozarelix (Figure 10I) and between two different ozarelix molecules (Figure 10J).
’ CONCLUSION The present study demonstrates that peptides of a defined structure are able to undergo electrostatic binding to anionic polyelectrolytes and stabilize themselves via hydrophobic interactions between neighboring peptide molecules. Using ozarelix and alginate as model compounds, we determined that this interaction led to the formation of structured aggregates. On the molecular level, these interactions led to the formation of nanofibers, where the polyelectrolyte serves as a backbone to which the peptide becomes bound. The diameter of the nanofibers was found to be 48 nm. Spectroscopic methods, including circular dichroism and UV spectroscopy, showed that the interaction of alginate and ozarelix induces fundamental changes to the molecular conformation of ozarelix. A modeling simulation of the interaction yielded a better understanding of the interaction within and between the decapepetide molecules. It was shown that, upon the electrostatic interaction with alginate,
ARTICLE
new intermolecular and intramolecular interactions of the ozarelix molecules occurred and altered the conformation of the peptide. Through the action of these additional forces, the peptide polyelectrolyte complex is further stabilized. This principle might be used in the future as a tool to increase the efficacy of drug hydrogel systems by enhancing the duration of drug release.
’ ASSOCIATED CONTENT
bS
Supporting Information. Additional figures. This material is available free of charge via the Internet at http://pubs.acs. org/.
’ AUTHOR INFORMATION Corresponding Author
*Fax: +49-941-943-4807. Phone: +49-941-943-4843. E-mail:
[email protected].
’ ACKNOWLEDGMENT We thank Mark Kalinich from the Department of ChemicalBiological Engineering at the Massachusetts Institute of Technology for proofreading the manuscript. ’ REFERENCES (1) Tauzin, B. Report: Biotechnology Medicines in Development; Pharmaceutical Research and Manufactures Association: Washington, DC, 2008. (2) Frokjaer, S.; Otzen, D. E. Nat. Rev. Drug. Discovery 2005, 4, 298–306. (3) Wang, W. Int. J. Pharm. 1999, 185 (2), 129–188. (4) Peppas, N. Curr. Opin. Colloid Interface Sci. 1997, 2, 531–537. (5) Hoare, T. R.; Kohane, D. S. Polymer 2008, 49 (8), 1993–2007. (6) Lin, C.-C.; Anseth, K. S. Pharm. Res. 2009, 26 (3), 631–643. (7) Sutter, M.; Siepmann, J.; Hennink, W. E.; Jiskoot, W. J. Controlled Release 2007, 119 (3), 301–312. (8) Branco, M. C.; Pochan, D. J.; Wagner, N. J.; Schneider, J. P. Biomaterials 2010, 31 (36), 9527–9534. (9) Tabata, Y.; Ikada, Y. Adv. Drug Delivery Rev. 1998, 31 (3), 287–301. (10) Cserhati, T.; Sz€ogyi, M. Peptides 1995, 16 (1), 165–173. (11) Vericat, C.; Vela, M. E.; Benitez, G.; Carro, P.; Salvarezza, R. C. Chem. Soc. Rev. 2010, 39, 1805–1834. (12) Jork, A.; Thuermer, F.; Cramer, H.; Zimmermann, G.; Gessner, P.; Haemel, K.; Hofmann, G.; Kuttler, B.; Hahn, H.-J.; JosimovicAlasevic, O.; Fritsch, K.-G.; Zimmermann, U. Appl. Microbiol. Biotechnol. 2000, 53 (2), 224–229. (13) Skaugrud, Ø.; Hagen, A.; Borgersen, B.; Dornish, M. Biotechnol. Genet. Eng. Rev. 1999, 16, 23–40. (14) Draget, K. I.; Taylor, C. Food Hydrocolloids 2011, 25 (2), 251–256. (15) Wiseman, T.; Williston, S.; Brandts, J. F.; Lin, L.-N. Anal. Biochem. 1989, 179 (1), 131–137. (16) Schuettelkopf, A. W.; van Aalten, D. M. F. Acta Crystallogr., D: Biol. Crystallogr. 2004, 60 (8), 1355–1363. (17) Berendsen, H. J. C.; Postma, J. P. M.; van Gunsteren, W. F.; Hermans, J. In Intermolecular Forces; Pullmann, B. E., Ed.; Reidel: Dordrecht, The Netherlands, 1981; p 331. (18) Van der Spoel, D.; Lindahl, E.; Hess, B.; van Buuren, A. R.; Apol, E.; Meulenhoff, P. F.; Tieleman, D. P.; Sijbers, A. L. T. M.; Feenstra, K. A.; van Drunen, R. Berendsen, H. J. C. GROMACS 3.3.1; Department of Biophysical Chemistry, University of Groningen: Groningen, The Netherlands, 2004 14458
dx.doi.org/10.1021/la202252m |Langmuir 2011, 27, 14450–14459
Langmuir
ARTICLE
(19) Berendsen, H. J. C.; Postma, J. P. M.; van Gunsteren, W. F.; DiNola, A.; Haak, J. R. J. Chem. Phys. 1984, 81 (8), 3684. (20) Essmann, U.; Perera, L.; Berkowitz, M. L.; Darden, T.; Lee, H.; Pedersen, L. G. J. Chem. Phys. 1995, 103 (19), 8577. (21) Mackie, W. Biochem. J. 1971, 125 (4), 89P. (22) Liu, Y.; Liu, Y.; Guo, R. J. Colloid Interface Sci. 2010, 351 (1), 180–189. (23) Waters, L. J.; Bedford, S.; Parkes, G. M. B.; Mitchell, J. C. Thermochim. Acta 2010, 511 (12), 102–106. (24) Wang, J.; Rabenstein, D. L. Biochim. Biophys. Acta, Gen. Subj. 2009, 1790 (12), 1689–1697. (25) Ross, P. D.; Subramanian, S. Biochemistry 1981, 20 (11), 3096–3102. (26) Turgeon, S. L.; Beaulieu, M.; Schmitt, C.; Sanchez, C. Curr. Opin. Colloid Interface Sci. 2003, 8 (45), 401–414. (27) Chan, C. J. Biol. Chem. 2004, 279 (37), 38749–38754. (28) Formaggio, F.; Peggion, C. A.; Crisma, M.; Kaptein, B.; Broxterman, Q. B.; Mazaleyrat, J.-P.; Wakselman, M.; Toniolo, C. Chirality 2004, 16, 388–397. (29) Chitnis, C. E.; Ohman, D. E. J. Bacteriol. 1990, 172 (6), 2894–2900. (30) Decho, A. W. Carbohydr. Res. 1999, 315 (34), 330–333.
14459
dx.doi.org/10.1021/la202252m |Langmuir 2011, 27, 14450–14459
