Controlling the Morphology of Metal-Promoted Higher Ordered

Our strategy for higher order assembly of collagen peptides containing repeating POG units is based on the metal-triggered joining of neighboring trip...
0 downloads 0 Views 4MB Size
Download PDF
Article pubs.acs.org/Langmuir

Controlling the Morphology of Metal-Promoted Higher Ordered Assemblies of Collagen Peptides with Varied Core Lengths Marcos M. Pires, Jeeyeon Lee, Dawn Ernenwein, and Jean Chmielewski* Department of Chemistry, Purdue University, 560 Oval Drive, West Lafayette, Indiana 47907, United States ABSTRACT: Self-assembling peptides have become an important subclass of next-generation biomaterials. In particular, materials that mimic the properties of collagen have received considerable attention due to the unique properties of natural collagen. Previous peptide-based designs have been successful in generating structures with morphological properties that were primarily determined by the type of self-assembling mechanism. Herein we demonstrate the metal ion-promoted, supramolecular assembly of collagenbased peptide triple helices into distinct morphologies that are controlled by defining the number of Pro-Hyp-Gly repeating units. We synthesized and characterized collagen-based peptides that incorporated either 5, 7, 9, or 11 Pro-Hyp-Gly repeating units. We found that the number of repeating units, and the resulting stability of the collagen triple helix, is intimately linked with the types of assemblies formed. For instance, collagen peptides that did not form a stable triple helix, such as NCoH5, did not participate in supramolecular assembly with added metal ions. Collagen peptides that formed stable triple helices, such as NCoH11, resulted in microsaddle structures with metalpromoted assembly, whereas a highly cross-linked, three-dimensional mesh formed with NCoH7, albeit at a higher metal ion concentration. These data provide evidence that triple helix formation is required for efficient metal-triggered assembly to the observed microstructures.

■

INTRODUCTION Recent advances in the design of polymeric self-assembling systems have led to the development of micro- and macroscaled materials that possess new and practical functions, such as conducting surfaces, templates for catalysis, self-healing structural supports, scaffolds for biomineralization, and tunable drug-releasing vehicles.1 Physical attributes of biomaterials have been previously shown to alter the encapsulation efficiency and release rate of associated molecules, as well cellular adhesion and differentiation for tissue engineering purposes.2 Materials such as poly(L-lactide) (PLLA), poly(ethylene oxide) (PEO), and poly(lactide-co-glycolic acid) (PLGA) have proven to be suitable for a range of biomedical applications, including drug delivery and as cellular matrices.3 However, self-assembling peptide-based biomaterials have recently been developed to overcome some of the limitations observed with traditional synthetic materials, such as poor biocompatibility and biodegradation.4 Furthermore, peptides are highly desirable building blocks for self-assembled materials due to their inherent programmability and range of biophysical properties. For example, Woolfson and co-workers recently demonstrated control over the length and width of peptide-based fibers by tuning the central amino acids of a self-assembling coiled coil peptide, while the sticky ends remained unchanged.5 Indeed, novel and general systems that yield biomaterials with diverse morphologies may represent significant advances to the field of functional self-assembled materials. One promising class of biomaterials is peptide mimics of natural collagen. Collagen is the most abundant protein in © 2011 American Chemical Society

mammals, constituting approximately one-quarter of the total protein mass.6 It is found in many different connective tissues including cartilage, bone, tendons, skin, and blood vessels. The mechanical strength and biochemical diversity of collagen is derived from its hierarchical structure that begins with the primary amino acid sequence and extends to its macromolecular structure. In nature, collagen is found in a triple helical state with individual left-handed polyproline type II strands twisted into a right-handed super coil. Collagen-like peptides have been used in order to model natural collagen7 and as building blocks in the construction of synthetic collagenlike assemblies.8 We have focused on designing stimutiresponsive higher ordered assemblies using collagen-like peptides.4b Having generated micrometer-sized structures using different assembly strategies, we sought to gain further insight into the determinants of structural morphology by maintaining the assembly strategy and varying the size of the internal collagen-like region of the peptide. Herein, we present a modular metal-mediated assembly strategy for a collagenbased peptide that generates three distinct microstructures by simply varying the length of the collagen portion of the peptide. Special Issue: Bioinspired Assemblies and Interfaces Received: September 30, 2011 Revised: December 9, 2011 Published: December 13, 2011 1993

dx.doi.org/10.1021/la203848r | Langmuir 2012, 28, 1993−1997

Langmuir

Article

Figure 1. (A) Structure of the collagen-based peptides containing metal-binding ligands and differing number of repeating POG units. All four peptides contain an NTA unit at the N-terminus (red) and a di-His unit at the C-terminus (blue). (B) Schematic representation of the metalpromoted assembly of triple helical collagen-based peptides.

■

Circular Dichorism (CD) Spectroscopy. CD melting temperature experiments were recorded on a Jasco CD spectropolarimeter (Model J810) using a 0.1 cm path length quartz cell. The CD data obtained were converted from degrees of rotation to mean residue ellipticity by dividing by the appropriate path length, peptide concentration, and number of residues in the peptide. Thermal stability of peptides was determined by measuring the mean residue ellipticity at 225 nm of solutions containing specified peptides (200 μM) in 20 mM 3-(N-morpholino)propanesulfonic acid (MOPS; pH 7.4). For variable temperature experiments, temperature was varied from 4 to 80 °C at 6 °C/h with a 0.2 nm data pitch and a 2 nm bandwidth. The melting temperatures (Tm) was determined to be the steepest slope of the unfolding process. The values were calculated by finding the minimum of the first derivative of the mean residue ellipticity versus temperature. Dynamic Light Scattering (DLS). DLS measurements were performed on a Zeta Sizer Nanoseries (Malvern) with laser wavelength of 633 nm. The solutions were measured in 50 μL quartz cuvettes and were placed in a sample holder under ambient conditions (temperature of 22 °C). The intensity size distributions were obtained from the analysis of the correlation functions using a multiple spherical modes algorithm. Preincubated peptide solutions and buffer solutions were filtered (0.45 μm pore size) prior to preparing the samples of NCoH5 or NCoH7 (1 mM) with zinc chloride (0.4 mM or 1 mM) in MOPS buffer (20 mM, pH 7.4), which were then incubated at room temperature for 24 h. Scanning Electron Microscopy (SEM) Imaging. SEM images of collagen peptide assemblies were obtained using a FEI NOVA nanoSEM high-resolution FESEM (FEI Company, Hillsboro Oregon) using the Helix low vacuum detector (0.98T) with operating parameters of 10 kV. Solutions composed of preincubated peptides (1 mM) and the specified metal ion (1 mM) in MOPS buffer (20 mM, pH 7.4) were incubated at room temperature for 24 h. Following assembly, all solutions were spun at 10 000g for 3 min, and the supernatant was carefully removed. The precipitates were washed by resuspending in 50 μL of double distilled H2O, spinning at 10 000g for 3 min, and carefully removing the supernatant. This washing step was repeated three times. Assemblies were resuspended in distilled water, and droplets of the sample (5 μL) were air-dried onto glass coverslips. The dried samples were sputter-coated with AuPd (3 min) prior to imaging.

EXPERIMENTAL METHODS

Materials. Rink Amide ChemMatrix resin was purchased from Matrix Innovation, Inc. (Montreal, Canada). All Fmoc-protected amino acids and activating agents for peptide synthesis were purchased from Novabiochem (La Jolla, CA). All other chemicals were purchased from Sigma Chemical Co. (St. Louis, Missouri) and used without further purification. General Synthesis of Peptides. The syntheses of the three peptides used in this study were carried out as follows: a 10 mL peptide synthesis flask was loaded with 500 mg (0.22 mmol) of Rink Amide ChemMatrix resin. The resin was washed with CH2Cl2 (3 × 7 mL) and dimethylformamide (DMF; 3 × 7 mL). Next, Fmocprotected amino acids (5 equiv, 1.0 mmol) in DMF (5 mL) were added to the synthesis flask with HBTU (5 equiv, 1.0 mmol) and N,Ndiisopropylethylamine (DIEA; 10 equiv, 2.0 mmol), and the flask was agitated for 3 h. The resin was washed with DMF, CH2Cl2, CH3OH, CH2Cl2, and DMF (3 × 7 mL each). Piperidine (20% in DMF, 5 mL) was added to the synthesis flask, the flask was agitated for 20 min, and the piperidine solution was drained. The resin was washed with DMF, CH2Cl2, and CH3OH (2 × 7 mL each). These steps were repeated until all amino acids were coupled to the resin. For the final Fmoc deprotection, piperidine (20% in DMF, 7 mL) was added to the synthesis flask, and after 20 min the flask was drained. Next, a protected NTA (2-(bis-tert-butoxycarbonylmethyl-amino)-6-carboxyamino-hexanoic acid tert-butyl ester) unit was installed, which had been synthesized according to literature procedures.9 Protected NTA (5 equiv, 1.0 mmol) in DMF (5 mL) was added to the reaction flask with HBTU (5 equiv, 1.0 mmol) and DIEA (10 equiv, 2.0 mmol), and the flask was agitated for 2 h. The resin was washed with DMF, CH2Cl2, and CH3OH (2 × 7 mL each). Cleavage/deprotection conditions: The resin was washed with DMF, CH2Cl2, CH3OH (3 × 7 mL each). A trifluoroacetic acid (TFA) cocktail solution (95% TFA, 2.5% triisopropylsilane, 2.5% H2O, 10 mL) was added to the resin, and the mixture was agitated for 2 h. The resulting solution was filtered and concentrated in vacuo to remove the TFA. The residue was triturated in cold diethyl ether, and the white precipitate was pelleted by centrifugation and dissolved in H2O. The desired peptide was purified to homogeneity by reverse-phase high-performance liquid chromatography (HPLC) using a Vydac C18 column with an eluent system consisting of solvent A (CH3CN/0.1% TFA) and solvent B (H2O/0.1% TFA) using a 60 min gradient consisting of 0 to 30% A, and a flow rate of 10 mL/min (peptide was monitored at λ 214 nm). Each purified peptide was further characterized by MALDI-TOF mass spectrometry. NCoH5 [M+H]+: 1872.81 (calculated) 1873.38 (found), NCoH7 [M+H]+: 2407.58 (calculated) 2406.80 (found), NCoH9 [M+H]+: 2941.37 (calculated) 2941.66 (found), NCoH11 [M+H]+: 3475.54 (calculated) 3476.60 (found).

■

RESULTS AND DISCUSSION

Our strategy for higher order assembly of collagen peptides containing repeating POG units is based on the metal-triggered joining of neighboring triple helical peptides that contain complementary ligands at each termini: a nitrilotriacetic acid 1994

dx.doi.org/10.1021/la203848r | Langmuir 2012, 28, 1993−1997

Langmuir

Article

Figure 2. (A) Thermal denaturation of the collagen-based peptide triple helices was monitored at 225 nm between 0 and 80 °C using CD spectroscopy. (B) DLS data showing the size distributions of the particles obtained for NCoH5 peptide alone (red), with addition of ZnCl2 0.4 mM (green), or ZnCl2 1 mM (blue) in MOPS pH 7.4; NCoH7 peptide alone (black) or with ZnCl2 0.4 mM (gray) in MOPS pH 7.4.

aqueous solution.11 Moreover, electrostatic repulsion at the Nterminus of the triple helix due to the introduction of the NTA ligand is expected to further disfavor NCoH5 trimerization. Previously we had shown that NCoH9 assembled in the presence of various divalent metal ions to form microflorettes of reproducible size and shape (Figure 3A).8d Both the NTA

(NTA) ligand at the N-terminus and two His residues at the Cterminus (Figure 1). The premise of the design is the linear propagation of collagen peptide triple helices due to metal− ligand associations. We previously discovered that the addition of various metal ions to triple helices of NCoH9 (Figure 1A) resulted in the formation of micrometer-sized particles, termed microflorettes (Figure 3A).8d Mechanistic studies at lower temperatures pointed to an initial linear association resulting in the formation of stacked sheets that curled and further assembled into the spherical florette structures.8d One direction of this growth involves linear assembly via the designed NTA/ histidine ligand−metal system, while the additional growth phases may be derived from further lateral association of the triple helices. It is known from the literature that long repeating units of the Pro-Hyp-Gly (POG) motif in synthetic collagen peptides are sufficient to assist lateral association.10 We hypothesized that varying the number of POG repeating units within our design may lead to morphologically distinct assemblies due to differences in the melting temperature of the triple helical peptides and their unique lengths. To this end, we synthesized and characterized three collagen-based peptides (NCoH5, NCoH7, and NCoH11) each with a different number of POG repeating units, but containing the same metal binding ligands at each termini as in NCoH9 (Figure 1A). The synthesis of all peptides was accomplished using Fmocbased solid phase chemistry as previously described.8d The stability of the triple helices of each peptide was characterized using CD spectroscopy. We found that all peptides exhibited a maximum molar ellipticity at 225 nm at 4 °C, a characteristic of polyproline type II structures. Thermal stabilities for each peptide triple helix were obtained by monitoring the molar ellipticity at 225 nm with increasing temperatures. As expected, the peptide with the longest continuous POG repeat, NCoH11, possessed the highest melting temperature (Tm) of approximately 67 °C (Figure 2A). The melting temperatures of NCoH9 and NCoH7 were found to be 50 °C8d and 25 °C, respectively. The melting curves for NCoH7, NCoH9, and NCoH11 exhibited the typical cooperative unfolding of the triple helix observed in collagen-like peptides.7b However, the melting curve for NCoH5 was devoid of cooperativity, an indication that this collagen-based peptide does not form a stable triple helix in solution. This finding is consistent with previous studies that showed that at least six repeating units of POG are necessary for the formation of stable triple helices in

Figure 3. (A) SEM images of metal-assisted microflorette structures composed of 1 mM NCoH9 and 400 μM Zn(II). (B−D) SEM images of metal-assembled saddle structures composed of 1 mM NCoH11 with 400 μM of specified metal ions in pH 7.4 MOPS buffer at room temperature for 24 h. Scale bar = 10 μm for A and scale bar = 3 μm for all other panels.

and dihistidine metal binding ligands were shown to be crucial for assembly, as control peptides lacking these ligands did not associate in the presence of metal ions.8d We envisioned that we could control the morphology of the assembled materials by varying the number of POG repeats, the feature that was expected to promote lateral association within the collagen-like peptides. Since NCoH9 and NCoH11 should both have a similar triple helical content at room temperature (approximately ∼85−95%, respectively), comparison of the metalpromoted assembly may indicate the role that the length of the collagen-like core plays in directing metal-promoted peptide assembly. SEM was used to image the materials obtained from 1995

dx.doi.org/10.1021/la203848r | Langmuir 2012, 28, 1993−1997

Langmuir

Article

Figure 4. SEM images of metal-assembled saddle structures composed of 1 mM NCoH11 with 400 μM of specified metal ions in pH 7.4 MOPS buffer at 4 °C for 24 h. Scale bar = 5 μm.

dimensional matrices observed for natural collagen. These results indicate that the number of POG repeating units and the triple helical state of the collagen-based peptide play an important role in how the metal-promoted assembly proceeds and also dictates the morphology of the final structure. We previously hypothesized that triple helix formation of NCoH9 was a prerequisite for metal-mediated assembly since homotrimerization of the individual strands aligns the metal binding ligands at each separate termini.8d To evaluate this hypothesis, we further decreased the number of POG repeating units in order to determine the effect on the assembled morphology. Peptide NCoH5 is composed of only five repeating POG units and did not assemble into a triple helix at room temperature. Addition of Zn(II) (0.4 mM and 1.0 mM) to a solution containing NCoH5 did not result in an observable precipitate. Analysis of these solutions by DLS demonstrated only a slight difference in the observed particle sizes that were formed from NCoH5 and NCoH5 with added metal ions (∼2−3 nm each) (Figure 2B). Several additional conditions were used to attempt to trigger peptide assembly, including using various metal ions (Cu(II), Ni(II), Co(II)), increasing the concentration of the peptide to 2 mM, decreasing the temperature to 4 °C, varying the ratio of metals to the peptide from 0.4 to 4.0, and prolonging incubation times. However, none of these efforts yielded self-assembled materials as DLS analysis confirmed that no additional assembly occurred in solution following the addition of metal ions (data not shown). The finding that NCoH5 does not form a triple helix, and exists as a monomeric species in solution indicates that binding between the metal loaded NTA unit and the histidine units would occur as a monovalent interaction. It has previously been shown that the binding of a metal-loaded NTA unit and dihistidine is 1000-fold weaker as a monovalent interaction as compared to a trivalent interaction.12 These data provide evidence that trimerization of the individual collagen peptide strands is a critical step for the designed metal-mediated assembly.

solutions containing NCoH11 (1 mM) and Zn(II), Co(II) or Cu(II) (400 μM each). The particles formed with NCoH11 and all three metal ions are quite distinct from the NCoH9 microflorettes;8d open, saddle-like structures were formed that are approximately 3−5 μm in size and appear to form from layered sheets (Figure 3B-D). In an effort to probe the supramolecular assembly mechanism for saddle-like structure formation, we also evaluated the assembly of NCoH11 at 4 °C with each of the metal ions. In each case, we observed the same saddle-like morphology composed of stacked sheets, but with a smaller number of layered sheets as compared to the microstructures formed at room temperature (Figure 4). These data indicate the potential for a hierarchical assembly, therefore, initiated by a linear, metal mediated assembly of NCoH11, followed by lateral association into sheets and stacked sheets. The collagen peptides NCoH5 and NCoH7 lacked the ability to form, or completely adopt, a stable triple helical conformation at room temperature. We used these peptides, therefore, to evaluate the need for a triple helical conformation for metal-mediated assembly. For instance, approximately half of NCoH7 exists as a homotrimer at room temperature. We set out, therefore, to determine the morphology of assemblies composed of NCoH7 with various divalent metal ions. Initial experiments revealed that the same conditions (1 mM peptide and 400 μM Zn2+ or Cu2+) used to generate microflorettes with NCoH9 and microsaddles with NCoH11 did not produce a visible precipitate with NCoH7. DLS was used evaluate whether soluble assemblies were forming with added metal ion. These data indicated that only small, nanometer-scale assemblies (∼63 nm) formed with NCoH7 under these conditions, as has been observed with other triple helical collagen peptides8f (Figure 2B). However, when the concentration of the metal ion was increased (1 mM Zn2+ or Cu2+) a turbid solution was obtained with NCoH7. SEM imaging revealed that the assemblies were composed of highly crosslinked fibers (Figure 5) that are reminiscent of the three-

■

CONCLUSION In conclusion, we have demonstrated that a strategy incorporating metal-binding ligands at the termini of collagen-like peptides can be modified for metal-assisted, supramolecular assembly by varying the number of POG repeating units. We demonstrate that a range of morphologies, including microflorettes, stacked sheet microsaddles, and fiberlike meshes, can be generated by varying the length of the collagen-like portion of the peptides. In the case of the microsaddles, mechanistic studies point to the formation of stacked sheets, indicating a complex hierarchical assembly process that may include linear association of the collagen

Figure 5. SEM images of assemblies formed from the NCoH7 peptide (1 mM) with specified metals (1 mM) in pH 7.4 MOPS buffer at room temperature for 24 h. Scale bar = 2 μm. 1996

dx.doi.org/10.1021/la203848r | Langmuir 2012, 28, 1993−1997

Langmuir

Article

172. (b) Feng, Y. B.; Melacini, G.; Taulane, J. P.; Goodman, M. Biopolymers 1996, 6, 859−872. (12) Shea, K. J.; Hart, B. R. Macromolecules 2002, 16, 6192−6201.

peptide strands at an early stage. Furthermore, we provide evidence that triple helix formation is required for efficient metal-triggered assembly to the microstructures. The facile modification of morphology by tailoring the collagen-like core of the peptide provides the opportunity to develop a range of novel nano- and micrometer-sized structures for biomaterial applications.

■ ■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected].

ACKNOWLEDGMENTS This work was supported by the National Science Foundation (0848325-CHE). We are grateful to Debbie Sherman for assistance with SEM.

■

REFERENCES

(1) Langer, R.; Vacanti, J. P. Science 1993, 5110, 920−926. (2) (a) Stuart, M. A. C.; Huck, W. T. S.; Genzer, J.; Muller, M.; Ober, C.; Stamm, M.; Sukhorukov, G. B.; Szleifer, I.; Tsukruk, V. V.; Urban, M.; Winnik, F.; Zauscher, S.; Luzinov, I.; Minko, S. Nat. Mater. 2010, 2, 101−113. (b) Barron, A. E.; Zuckermann, R. N. Curr. Opin. Chem. Biol. 1999, 6, 681−687. (3) (a) Langer, R.; Tirrell, D. A. Nature 2004, 6982, 487−492. (b) Anderson, J. M.; Shive, M. S. Adv. Drug Deliver Rev. 1997, 1, 5−24. (4) (a) Stupp, S. I.; Cui, H. G.; Webber, M. J. Biopolymers 2010, 1, 1−18. (b) Chmielewski, J.; Przybyla, D. E. Biochemistry 2010, 21, 4411−4419. (c) Woolfson, D. N.; Bromley, E. H. C.; Channon, K.; Moutevelis, E. ACS Chem. Biol. 2008, 1, 38−50. (d) Woolfson, D. N.; Ryadnov, M. G. Curr. Opin. Chem. Biol. 2006, 6, 559−567. (e) Rajagopal, K.; Schneider, J. P. Curr. Opin. Struct. Biol. 2004, 4, 480−486. (f) Zhang, S. G. Nat. Biotechnol. 2003, 10, 1171−1178. (5) Woolfson, D. N.; Papapostolou, D.; Bromley, E. H. C.; Bano, C. J. Am. Chem. Soc. 2008, 15, 5124−5130. (6) (a) Brodsky, B.; Ramshaw, J. A. M. Matrix Biol. 1997, 15, 545− 554. (b) Prockop, D. J.; Kivirikko, K. I. Annu. Rev. Biochem. 1995, 403−434. (7) (a) Brodsky, B.; Persikov, A. V.; Ramshaw, J. A. M.; Kirkpatrick, A. Biochemistry 2000, 48, 14960−14967. (b) Shoulders, M. D.; Raines, R. T. Annu. Rev. Biochem. 2009, 929−958. (c) Goodman, M.; Kwak, J.; Jefferson, E. A.; Bhumralkar, M. Bioorgan. Med. Chem. 1999, 1, 153− 160. (8) (a) Horng, J. C.; Chen, C. C.; Hsu, W.; Kao, T. C. Biochemistry 2011, 13, 2381−2383. (b) Chmielewski, J.; Przybyla, D. E. J. Am. Chem. Soc. 2010, 23, 7866−7867. (c) Chmielewski, J.; Pires, M. M.; Przybyla, D. E. Angew. Chem., Int. Ed. 2009, 42, 7813−7817. (d) Chmielewski, J.; Pires, M. M. J. Am. Chem. Soc. 2009, 7, 2706− 2712. (e) Cejas, M. A.; Kinney, W. A.; Chen, C.; Vinter, J. G.; Almond, H. R. Jr.; Balss, K. M.; Maryanoff, C. A.; Schmidt, U.; Breslav, M.; Mahan, A.; Lacy, E.; Maryanoff, B. E. Proc. Natl. Acad. Sci. U.S.A. 2008, 25, 8513−8518. (f) Chmielewski, J.; Przybyla, D. E. J. Am. Chem. Soc. 2008, 38, 12610−12611. (g) Conticello, V. P.; Rele, S.; Song, Y. H.; Apkarian, R. P.; Qu, Z.; Chaikof, E. L. J. Am. Chem. Soc. 2007, 47, 14780−14787. (h) Kinney, W. A.; Cejas, M. A.; Chen, C.; Leo, G. C.; Tounge, B. A.; Vinter, J. G.; Joshi, P. P.; Maryanoff, B. E. J. Am. Chem. Soc. 2007, 8, 2202−2212. (i) Kotch, F. W.; Raines, R. T. Proc. Natl. Acad. Sci. U.S.A. 2006, 9, 3028−3033. (j) Paramonov, S. E.; Gauba, V.; Hartgerink, J. D. Macromolecules 2005, 18, 7555−7561. (k) Koide, T.; Homma, D. L.; Asada, S.; Kitagawa, K. Bioorg. Med. Chem. Lett. 2005, 23, 5230−5233. (9) Lata, S.; Reichel, A.; Brock, R.; Tampe, R.; Piehler, J. J. Am. Chem. Soc. 2005, 29, 10205−10215. (10) Brodsky, B.; Kar, K.; Amin, P.; Bryan, M. A.; Persikov, A. V.; Mohs, A.; Wang, Y. H. J. Biol. Chem. 2006, 44, 33283−33290. (11) (a) Tanihara, M.; Kishimoto, T.; Morihara, Y.; Osanai, M.; Ogata, S.; Kamitakahara, M.; Ohtsuki, C. Biopolymers 2005, 3, 163− 1997

dx.doi.org/10.1021/la203848r | Langmuir 2012, 28, 1993−1997