Electrochemically Directed Assembly of Designer Coiled-Coil

Subscriber access provided by Gothenburg University Library

Article

Electrochemically-directed assembly of designer coiled-coil telechelic proteins Yinan Lin, Bo An, Mehran Bagheri, Qianrui Wang, Jim Harden, and David L Kaplan ACS Biomater. Sci. Eng., Just Accepted Manuscript • DOI: 10.1021/acsbiomaterials.7b00599 • Publication Date (Web): 28 Sep 2017 Downloaded from http://pubs.acs.org on September 29, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Biomaterials Science & Engineering is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Biomaterials Science & Engineering

Electrochemically-directed assembly of designer coiled-coil telechelic proteins

Yinan Lin,† Bo An,† Mehran Bagheri,‡ Qianrui Wang,† James L. Harden,‡,* and David L. Kaplan*,†

†

Department of Biomedical Engineering, Tufts University, Medford, Massachusetts 02155, United States

‡

Department of Physics, University of Ottawa, Ontario K1N 6N5, Canada

*

Corresponding Authors: [email protected] and [email protected]

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ABSTRACT: We report the design and characterization of a de novo electrogelation protein comprised of a central spider silk glue motif flanked by terminal pH-triggered coiled-coil domains. The coiled-coiled domains were designed to form intramolecular helix bundles below a sharply-defined pH-trigger point (~pH 5.3), while the spider silk glue protein, owing to its substantial Glu content, serves both as an anionic electrophoretic transport element at neutral and elevated pH and as a disordered linker chain between the associated helix bundles at reduced pH. We show that in an electrochemical cell, a solution of these telechelic proteins migrates toward the anode where the terminal coiled-coil domains are triggered to form coiled-coil assemblies that act as transient crosslinks for the e-gel state. Upon cessation the current, the coiled-coil domains become denatured and the e-gel transforms back into a fluid solution of polypeptides in a fully reversible manner. This simplified triblock protein design mimics many of the characteristics of more complex electrogelation proteins, such as silk fibroin. As such, provides some insight into possible general mechanisms of protein electro-gelation. Moreover, this general class of electrogelation proteins has potential for biomedical applications of electrochemically triggered gelation, such as externally-switchable delivery of therapeutic cell and drugs from a responsive matrix.

KEY WORDS: electrogelation, hydrogel, coiled-coils, stimuli-responsive, telechelic protein


Page 2 of 28

Page 3 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1. Introduction Desired dynamic responses to various environmental cues are key in the design and development of the next-generation intelligent biomaterials to address the increasingly sophisticated challenges facing the field of biomedical technologies.1 In recent years, chimeric protein-based hydrogels have emerged as promising materials for tailor-made, multitasking platforms in tissue engineering and regenerative medicine.2-4 Extensive efforts, therefore, have been devoted to expand the toolkit of stimuli-responsive peptide motifs, ranging from chemically and chemo-enzymatically active sites to supramolecular selfassembled nanostructures.5 In particular, biomaterials that can be designed to undergo a change of physical properties (microstructural or mechanical properties) in response to an external control field have be an area of emerging interest. Electrogelation, wherein a material makes a transition between a fluid state and a solid, gel state in response to an applied electric field or current, is an area of active research interest. Example systems include both synthetic polymers6-12 and biomacromolecules12-19. Notably, fully reversible electogelation in response to low (normally < 1 mA) applied currents has been observed in regenerated Bombyx mori (B. mori) fibroin silk solutions16-19. Experimental and simulation studies revealed that the formation and dissolution of silk electrogels (e-gels) are governed by the electrochemical gradient of hydrogen ions (H+) built up in the near-anode regions with magnitude ranging from a few microns to a few millimeters, controlled by the spatial distribution of pH of the electrolysis reaction. The resulting soft e-gels have novel adhesive properties for a wide range of materials, including glass, metals, plastics, and wood. Unlike the nominal physicochemical gelation processes for regenerated B. mori fibroin soutions, which involve essentially irreversible formation of nanocrystalline β-sheet domains, electrogelled silk is dominated by fully reversible transformations of the random coil fraction of the fibroin in a complex hierarchical assembly process which is as yet not fully understood. Motivated by the silk e-gel system and the extensive literature on pH-responsive secondary structure, we have developed a simple de novo electrogelation protein system comprised of a central spider silk glue motif flanked by terminal pH-triggered coiled-coil domains. In an electrochemical cell, a solution of these telechelic proteins migrates toward the anode where the terminal coiled-coil domains are triggered to form coiled-coil assemblies that act as transient crosslinks for the designer e-gel. Upon switching off the current, the coiled-coil domains become denatured and the e-gel transforms back into a fluid solution of polypeptide, mimicking the native silk e-gel behavior. In the following, we first describe the design and properties of the coiled-coil Cc and spider silk glue E domains of telechelic construct CcECc. Next, we describe the synthesis and characterization of the telechelic CcECc e-gel system, including the inducible transition between the e-gel and solution states, and the morphology and microrheology of the e-gel phase. We conclude with a discussion of our finding and the possible biomedical applications of the designer e-gel system

2. Materials and Methods Materials. Peptides (> 98 % purity, HPLC) were purchased from GL Biochem (Shanghai, China). Citric acid (≥ 99.5%), sodium citrate dihydrate (≥ 99%), sodium acetate (CH3COONa, ≥ 99%), acetic acid (CH3COOH, ≥ 99%), 4-(2-hydroxyethyl)piperazine-1-ethanesulfonic acid (HEPES, ≥ 99.5%), hydrochloric acid (HCl, 37%), tris(hydroxymethyl)aminomethane (Tris base, ≥ 99.8%), sodium phosphate dibasic (Na2HPO4, ≥ 99%), sodium bicarbonate (NaHCO3, ≥ 99.5%), and sodium hydroxide (NaOH, ≥ 98%) were purchased from Sigma-Aldrich (St. Louis, MO). Phosphate buffered saline (PBS, 10X, pH 7.4) and UltrapureTM water were obtained from Life Technologies (Carlsbad, CA). All chemicals were used without further purification.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Construction of expression plasmids. The amino acid sequence for the E domain, (EEPETPSPET)10, is based on the consensus repetitive sequence of aggregate spider glue 1 of Nephila clavipes (Uniprot entry B7SVM6); the sequence for the Cc coiled coil domain is: IAALEAENEALKAEIAELKAEIAAEKAE A 6x his-tag was included at the N-terminus of all recombinant proteins for immobilized nickel affinity column purification. A triple glycine linker (Gly-Gly-Gly) was included between the CC domain and E domain to allow extra flexibility. A Trp residue was added after the N-terminal triple glycine linker for accurate protein concentration measurement by UV absorbance at 280nm. An RGD integrin binding sequence is inserted between two of the ten E domain repeats for future use as a cell adhesion promoter. See Figure S1 for an overview. The DNA sequences of both protein domains were synthesized at Biomatik Corp (Wilmington, DE) and fused together through Gibson Assembly. The construct was subcloned into pColdIII expression vector (Takara Bio Inc) through the NdeI and XbaI restriction sites. All enzymes and reagents for cloning were purchased from New England Biolabs (Ipswich, MA). DNA sequencing to confirm fidelity was carried out at the Tufts Core Facility. Protein expression and purification. All constructs in the pColdIII vector were expressed in E. coli BL21 strain, grown in 20 ml LB medium with 100 µg/ml ampicillin overnight at 37°C. This starting culture was used to inoculate 1000 ml of LB-ampicillin media in a shaking flask and grown at 37oC to an OD600nm=1.0. To induce protein expression, 1 mM isopropyl-β-D-thiogalactopyranoside (IPTG, Thermo Scientific) was added to the culture and the temperature was lowered to 22°C. After 16 h induction, cells were harvested by centrifugation and resuspended in His-tag column binding buffer (20 mM sodium phosphate buffer pH 7.4, 500 mM NaCl, 5 mM imidazole) containing 0.25 mg/ml lysozyme and frozen at -80°C until purification. Purification of the His-tagged recombinant proteins was carried out by affinity chromatography with nickel ion binding and imidazole gradient elution on an ÄKTA pure 25L FPLC (GE Healthcare). Frozen cells were thawed and further lysed by sonication. Cellular debris was removed by centrifugation at 8,000 x g at 4°C. The supernatant containing the soluble target protein was injected onto binding buffer equilibrated 5ml HisTrap HP columns (GE Healthcare). The column was then washed sequentially with 3 column volumes of binding buffer, binding buffer plus 50 mM imidazole and binding buffer plus 20 mM of imidazole. The his-tagged protein was eluted by elution buffer (binding buffer plus 100 mM imidazole). Identification of recombinant CcECc proteins. Purified protein were dialyzed into ddH2O and the purity was checked by SDS-PAGE (NuPAGE® Bis-Tris 4-12% with MES-SDS buffer, Invitrogen), as shown in Figure S2. Protein concentration was measured on a NanoDrop (Thermo Scientific) based on UV absorbance at 280nm. The molecular weights were confirmed via matrix assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometry (Figure S3) on an Applied Biosystems Voyager De-Pro MALDI (Applied Biosystems, Foster City, CA, USA) at the Tufts Core Facility. The protein solution was eventually lyophilized and sent for amino acid analysis (Figure S4) at Molecular Structure Facility of University of California, Davis. The 1H NMR spectra (Figure S4) were recorded on a Bruker (Billerica, MA, USA) Advance 500 MHz spectrometer. Potentiometric titration of CcECc in water. The carboxylate content of the purified recombinant protein was determined by the conventional conductometric titration method. To 20 mg of freeze-dried protein, 60 mL deionized water was added. A well-dispersed slurry was obtained after the mixture was stirred for 30 min. A 0.1 M HCl solution was then used to adjust the sample pH to the range between 2.5 and 3.0. Subsequently, a 0.05 M NaOH solution was added to the sample at a rate of 0.1 mL/min till the pH reached 11. The entire process was monitored by a pH-Stat titration system. The titration curve of deionized water was acquired in the same way. Preparation of buffer stocks. Different buffers were used to study how the pH affects the conformation of the peptides and the recombinant proteins: six 1 M citric acid/sodium citrate buffers of pH 3.0, 4.0, 4.5, 5.5, 6.0 and 6.5, one 1 M sodium acetate/acetic acid buffer of pH 5.1, one 1 M HEPES buffer of pH


Page 4 of 28

Page 5 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


7.0, one 0.1 M PBS buffer of pH 7.4, three 1 M Tris/HCl buffers of pH 8.0, 8.5 and 9.0, three 0.5 M phosphate buffers of pH 9.5, 10 and 10.5, and one 1 M sodium bicarbonate/sodium hydroxide buffer of pH 11.0. The peptide/protein concentrations were determined by the absorbance measured at 214 nm and 280 nm, using the molar extinction coefficients: ɛ280 = 5,690 cm−1M−1 per Trp residue and ɛ214 = 2,200 cm−1M−1 per peptide bond.20 Circular Dichroism (CD) spectroscopy. The CD measurements were carried out in 1 mm path length quartz cuvettes using an Aviv model 62DS spectrophotometer equipped with a Peltier temperature controller (Aviv Biomedical, Lakewood, NJ). The purified Cc, E and CcECc polypeptides were dissolved in buffers (~10 mM ionic strength) at concentrations ~30 µM, 12 µM, and 10 µM, respectively, and allowed to equilibrate at 20°C for 30 min prior to CD experiments. Measurements at each pH were obtained at 20°C in the 190 ~ 260 nm range (but only data from 190 ~ 250 nm is shown) at a resolution of 0.5 nm and at a scanning speed of 50 nm/min. Baselines recorded using the same buffer and cuvette were subtracted from the data. Baselines recorded using the same buffer and cuvette were subtracted from the data. The trigger or switch pH, denoted as pHTrigger, where a helix-to-random conformational transition occurred, was determined for each peptide from the pH dependence of the mean residue ellipticity ratios at 208 nm and at 222 nm. Differential Scanning Calorimetry (DSC). DSC experiments (Figure 2C) were performed on a NANO DSC II Model 6100 (Calorimetry Sciences Corp, Lindon, UT). Cc peptide samples were dissolved in citrate buffer pH 3.5 to a concentration of 5 mM. The solutions were dialyzed against the same buffer before measurement to collect the dialyzed buffer as reference in the experiment. Sample solutions were loaded at 0°C into the cell and heated at a rate of 0.1°C/min till 100°C. Data points were plotted in MicroCal Origin 6.0. Tm and enthalpy were calculated based on peak of curve and area under the curve respectively. Size exclusion chromatography. The pH-triggerable aggregation behaviors of the Cc peptide was examined by size exclusion chromatography (Figure 2) using an ÄKTAmicro chromatography system (GE Healthcare, Uppsala, Sweden) equipped with a Yarra SEC-3000 column (Phenomenex, Torrance, CA) in PBS (pH 7.4) and sodium acetate buffer (pH 4.0), respectively, at 4°C. Standard calibration of molecular weights was done with a Gel Filtration Calibration Standard sample (Bio-Rad, Hercules, CA). Determination of critical micelle concentration using UV-visible spectroscopy. The optical density at 400 nm was measured for a series of aqueous CcECc solutions/suspensions of varying concentrations, at pH 3.5 and 10.0, respectively, with an Aviv 14DS UV-visible spectrophotometer equipped with a Peltier temperature controller (Aviv Biomedical, Lakewood, NJ). Zetasizer measurements of micellar nanoparticles of CcECc. Both the size distribution and zeta potential of the CcECc micellar aggregates (0.7 mg/mL at varying pHs) were determined by a dynamic light scattering (DLS) setup using a Beckman Coulter Delsa Nano analyzer (Beckman Coulter Inc., Danvers, MA, USA). Fabrication of milli-fluidic devices. Prototype devices were designed and fabricated with optically transparent materials, polydimethylsiloxane (PDMS) and glass, to allow visible light to pass through for in situ light microscopy (Figure S5). Each device consisted of two PDMS films (Sylgard 184, Dow Corning Corp., Midland, MI) flanked by two glass slides 5-6 mm apart from each other. Before device assembly, gold electrodes (20 nm of chromium adhesion layer followed by 200 nm of gold) were patterned both on the top and side of the glass slides using magnetron sputtering (Nanomaster NSC3000). The PDMS films were then plasma treated together with the two glass slides for 30 sec, and put into contact to produce a covalently bonded and fully sealed milli-fluidic chip.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

In-situ, real-time optical monitoring of electrogelation of CcECc. The CcECc solution (~1.0 mg/mL in water) was injected into the milli-fluidic channel. A few drops of a universal pH indicator (Acros Organics, New Jersey, USA) were mixed with the CcECc solution for real-time imaging of the pH changes during the electrogelation process. The growth of CcECc e-gel was induced at the solution/positive electrode interface by applying a constant voltage across the parallel electrodes patterned on opposite interior walls of the square-shaped mini-channel using a DC power supply (Agilent E3612A DC power supply, Agilent Technologies, Inc., Englewood, CO). The resulting advancement of the interface between the optically clear solution and the opaque e-gel was monitored in situ by a Leica DMIL inverted microscope (Leica Microsystems, Wetzlar, Germany). The increase in thickness of the thin layer e-gel was measured using the ImageJ software (National Institutes of Health, USA). The pHdependent color changes were later calibrated with standard buffer solutions. Electrogelation of CcECc by a different experimental geometry. A 100 µL aliquot of CcECc solution (~1.0 mg/mL) was placed in a reservoir composed of a 9 mm diameter hole in a PDMS sheet (spacer) and a piece of aluminum foil (Al, bottom, positive electrode). Another piece of Al, acting as the negative electrode, was placed at the top in contact with the protein solution. The CcECc electrogelation was initiated near the bottom electrode surface after the application of a constant voltage of 12 V, generating an electric field of 4 V/mm across the two electrodes separated by 3 mm. Microparticle tracking microrheology. Brownian motion of green fluorescent tracer particles (5 µm diameter, polystyrene, Spherotech, Libertyville, IL) in water and in CcECc (1 mg/mL) was respectively monitored with a Leica DFC340FX digital color camera mounted on a Leica DMIL inverted microscope (Leica Microsystems, Wetzlar, Germany) at an exposure time of 43 ms and a rate of 1 frame per second. The microbeads were introduced into CcECc at a v/v ratio of 1:1000. The liquid was placed in a reservoir composed of a 9 mm diameter hole in PDMS and a glass cover slide. The observation of electrogelled CcECc was enabled by using a partially gold-coated cover glass at the bottom as the anode. Image data were collected immediately after turning off the potential difference at a location a few millimeters away from the edge of the gold coating. Data analysis was based on previous described methods21. Calculations were performed in Matlab using publically available particle tracking code from Georgetown Univeristy (http://site.physics.georgetown.edu/matlab/index.html). Cryogenic transmission electron microscopy (Cryo-TEM). To prepare the cryo-samples, 10~20 µL of CcECc (0.9~1.0 mg/mL in water) was spread on C-flat holey carbon grids (Electron Microscopy Sciences, Hatfield, PA) held by a pair of plunging forceps. A CcECc electrogel was formed on the TEM grid by using a platinum wire (0.5 mm in diameter, Sigma-Aldrich, St. Louis, MO) as the cathode. The forceps with either the CcECc solution or e-gel were quickly transferred to a semi-automated plunge freezing instrument (Gatan Cryoplunge Cp3, Gatan, Pleasanton, CA) equipped with a chamber of 100% humidity. Each of the samples was subsequently blotted for 2 sec at 4 °C, and flash-plunged into liquid ethane. The prepared cryo-grids were then transferred into the CT3500 cryo-transfer system (Gatan, Pleasanton, CA) in liquid nitrogen. Micrographs of CcECc were collected at -183 °C using a Tecnai F20 TEM (FEI, Hillsboro, Oregon) with a field-emission gun at 200 kV and a calibrated Ultrascan CCD camera (Gatan, Pleasanton, CA). Molecular dynamics simulations. Initial helical configurations of the Cc peptide sequence and of putative coiled-coil oligomers of it were generated using a python-package developed in our group. Coiled-coils with hydrophobic cores of three types were investigated: the classical coiled-coil type comprised of the a and d hydrophobic residue faces (denoted HP), hydrophobic cores comprised of alanine residues from the f positions (denoted ALA), and mixed cores comprised of a-d hydrophobic residue faces in contact with alanine resudues from f positions (denoted ALA-HP). In all cases, the relative orientations of Cc peptide monomers in these coiled-coil assemblies were chosen to favor intermolecular electrostatic interactions between charged residues in the e and g heptad positions. These


Page 6 of 28

Page 7 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


assemblies were solvated in TIP3P water22 with 10 mM NaCl in a box with periodic boundary conditions and boundaries that extend 15 angstroms from the polypeptide assembly surface in each direction, to avoid first neighbor simulation box interactions. All simulations were carried out with a time-step of 2 fs using NAMD 2.923 with the CHARMM22 force field24 for the polypeptides. Protonated glutamic acid residues emulated low pH solution conditions. A Langevin thermostat and barostat were employed under NPT conditions at 300 K and 1 atm. The cut-off for van der Waals interactions was 12 angstroms and electrostatic interactions were computed using the particle-mesh Ewald algorithm. At least 100 ns equilibration was performed for all configurations studied, without any restraints. The most stable conformations in the last 20 ns were adopted for Free energy calculations, as described below. The VMD package was used to produce all biomolecular images25 Free energy calculations. Free energy difference calculations were carried out using the Alchemical Free Energy Perturbation26 method, as implemented in the NAMD software package. In Free Energy Perturbation (FEP) methods, the free energy difference between two states, A and B, is calculated from

where H(λ) = (1 − λ)HA + λ HB, with HA and HB representing the Hamiltonian of the initial (λ =0) and final (λ =1) states. The sum is therefore over a set of N discrete steps in λ from λ =0 to λ=1. Using thermodynamic integration27, the free energy difference between two states, A and B, is obtained by integrating H(λ) with respect to λ, from λ =0 to λ=1.λWe used this method to estimate the free energy of association for dimeric, trimeric, and tetrameric coiled-coil aggregates by comparing the associated state to fully solvated, dissociated states of the helical peptide. To do so, stable conformations of the helical peptide and its assemblies were chosen from configurations in the last 20 ns of the equilibration simulations described above, and multiple FEP simulations with restrained peptide backbones were conducted under the same conditions as in the equilibration simulations. These were used, along with the computed free energy cost of the backbone restraints, to compare the free energies of association on a per chain basis between associated states with different pockets (e.g. dimers with HP vs ALA pockets) and different states of association (dimers, trimers, and tetramers with HP pockets).

3. Results & Discussion Protein Design. The overall design motif for the e-gel protein is the well-known telechelic architecture from the field of associating polymers28. Synthetic telechelic polymers are polymers that have solvent incompatible terminal groups; while the main chain is soluble in the solvent of choice, the terminal groups are not, and thus they tend to aggregate into clusters. This aggregation results in a physically crosslinked gel state, where the crosslinks are dynamic entities with finite mean lifetimes. A classic example for aqueous solvents is the case of poly(ethylene oxide) main chains with short alkyl end groups. Designer protein analogues of these telechelics were pioneered by Tirrell and co-workers29, where the central domain was a disordered sequence that is soluble at moderate to high values of pH and where the associating end domains were amphiphilic alpha-helices (coiled-coil variants of the so-called leucine zipper motif 30-35) that have an innate preference to form bundles in aqueous solution. Over the last few decades, synthetic proteins have been successfully engineered with coiled-coil domains to accomplish a variety of functions, including self-assembly of fibrils, hydrogels and biofunctional surfaces, liquid crystallinity, piezoelectricity, modular molecular recognition, and controlled hydrogel erosion.29,36-45 We also use such a coiled-coil based telechelic protein approach for the design of the minimalist e-gel system herein. The central and end domains of our system have specific requirements to facilitate the formation of an efficient, switchable e-gel system. In particular, the central domain should have both a pH-



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

dependent solubility and electrophoretic mobility, to facilitate transport of the protein to the gelling electrode of the electrochemical cell. The associating coiled-coil ends should preferably undergo a sharp transition between the disordered and functional alpha-helical conformations at a convenient value of solution pH. Figure 1 shows a schematic representation of the expected change of state of a triblock CcECc protein with a change of pH. We now discuss each of these domains in more detail. Electrophoretic central domain. A spider silk glue sequence (EEPETPSPET)10 was selected as the central electrophoretic E domain motif. This sequence is based on one of the aggregate spider silk glue protein sequences cloned from the golden orb weaver spider (Nephila clavipes).46 The native spider silk glue is one of the strongest and most effective biological glues. It contains numerous repeats of the EEPETPSPET sequence and is also heavily glycosylated.47,48 It has been suggested that the strong adhesive properties of this protein mainly come from the glycans.49 However, the amino acid sequence serves here as an excellent template for an electrophoretically direct-able domain because of its high fraction of Glu residues, which make the chimeric protein carry a strong net negative charge above the pKa of the Glu. In a region of relative high pH, as is found near the cathode of an electrochemical cell, such an E domain would have a high electrophoretic mobility that would drive it away from the cathode and toward the anode. Moreover, once near the anode, where the pH approaches the pKa of the Glu, the domain chain conformation is expected to become more compact, since its solubility at reduced pH is only supported by weak residual charge of the Glu residues and by the less numerous polar Ser and Thr residues. Thus, a somewhat denser polypeptide environment is achieved that facilitates the assembly of the coiled-coil crosslinks. Inducible coiled-coil end domains. At the amino acid sequence level, classical coiled-coils are characteristically comprised of consecutive heptad motifs, denoted as a-b-c-d-e-f-g. The amino acid residues at the a and d positions are predominantly hydrophobic, and pack side-by-side into a hydrophobic core in a “knobs-into-holes” manner.30-33 The proximal e and g positions, when occupied by charged residues, mediate the intra- and inter-helical electrostatic interactions, and thus modulate the stability of each oligomeric state.30-33,50 The residues at the surface exposed b, c, and f positions play a role in the inherent stability of the helical state and also act collectively to organize solvent structure in the vicinity of the coiled-coil assemblies.33-35,51 Conformational switching between multi-stranded helix bundles and single-stranded random coils in response to changes in pH is a central consideration in the design of our pH-reversible e-gel system. For electrochemical control of conformational switching, it is highly desirable to control the nominal value of the pH at the trigger point and to minimize the pH span of the transition. Past research efforts have focused on widening the range of the trigger pH points in these systems. Histidine substitutions, for example, were introduced at the conventionally hydrophobic core d positions of alternate heptads in a coiled-coil peptide, to increase the inter-helical repulsion at the low end of the pH scale and, consequently, triggered formation of α-helical fibers at pH ≈ 6.0, near the pKa of the free imidazole side chain of histidine.45,52 The pKa values of the amino acid side chain groups have also been modulated by selected conformational attributes, such as peptide chain folding, inter-chain association and the establishment of a local hydrophobic pocket. For instance, when glutamic acid residues (with a free side chain carboxyl group pKa of 4.3), were placed in the peripheral e and g positions, the conformational switching in the reported coiled-coil peptides shifted to a pH between 5.5 and 6.0.50 Even at the surface b and c positions, charged residues, e.g., lysine and glutamic acid, were found to affect the pH-dependent stability of coiled-coil motifs.51 The sequence design of our terminal e-gel association Cc motif is comprised of modified sequences of previously reported dimerization motifs utilized in de novo fibers and hydrogels.33,36 Table 1 gives the sequence of Cc. A cursory inspection of the sequence reveals many of the standard features of classical coiled-coils discussed above. Notable differences include (i) a glutamic acid substitution in the hydrophobic core d position 25 that was chosen to destabilize coiled-coil assemblies at elevated pH, (ii)


Page 8 of 28

Page 9 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


glutamic acid substitutions two at the surface-exposed b and c positions in the two middle repeats (positions 9 and 17) that were included in order to stabilize tetrameric assemblies, and (iii) alanine residues placed at all surface-exposed f positions, in order to enhance coiled-coil stability at low pH.34,35 We show below that this design is capable of completing the coil-helix switching within a narrow pH range between 5.2 and 6.0. This is in contrast to other reported pH switchable systems, for which the transition ranges generally span 2 to 3 pH units. Moreover, in the low pH regime, this Cc motif favors the formation of tetrameric coiled-coils. Coiled-coil domain behavior. Detailed experimental and computational studies were carried out to the characterize the secondary structure, thermodynamic stability, and association behavior of the Cc peptides prior to their use as assembly motifs in the CcECc system. Circular dichroism (CD) was used to characterize the secondary structure of the Cc peptide and its dependence on pH and temperature. Figure 2A shows CD scans from 190 to 250 nm for Cc peptide solutions (30 µM in aqueous buffers, described in the methods section below, at 20°C) over a pH range from 2.0 to 11.0. The α-helical structure is characterized by two negative dichroic minima at 208 and 222 nm, and higher-order coiled-coil oligomerization is reflected by a minimum at 222 nm that is more negative than the one at 208 nm.53 The pH-dependence of the ratio of ellipticities at 222 and 208 nm, ρ=[θ]222nm/[θ]208nm, is summarized in Figure 2B. These CD results reveal a sharp helix-coil transition occurring in the pH range between pH 5.2 and 6.0, demonstrating high pH sensitivity and selectivity for this system. Notably, the ratio ρ dropped significantly (from ρ≈1.1 to 0.24) across the trigger point (pHTrigger ≈ 5.3), suggesting that a sudden loss of oligomerization accompanies the loss of secondary structure. The thermal stability of the Cc peptides at low pH was also characterized using differential scanning calorimetry (DSC) for 5 mM peptide solutions in 10 mM citric buffer at pH 3.5. Figure 2C shows a temperature-induced denaturation transition occurring over a temperature range between 45 and 80°C, as indicated by a melting peak centered around 74°C. The asymmetric shape of the Cc peptide thermal transition suggests the existence of several distinct oligomeric states of this peptide in solution. The nature of these oligomers was next studied by analytical size exclusion chromatography (SEC) (conducted at pH 4 on 150 µM peptide solutions in 10 mM acetate buffer). Figure 2D presents SEC data showing four clear oligomeric states. Specifically, deconvolution of the SEC data indicated the existence of a dominant tetramer (39%) in coexistence with monomeric peptide (33%) and with dimeric (14%) and trimeric aggregates (15%) in these acidic conditions where helical secondary structure is stable. To further understand the coiled-coil oligomerization at the molecular level, all atom molecular dynamics (MD) simulations, conducted in 10 mM NaCl solutions under NPT conditions at 300 K and 1 atm., were performed using the NAMD simulation package23. Dimeric, trimeric, and tetrameric configurations of the Cc peptide were constructed from its helical monomer conformation. In particular, aggregates with hydrophobic cores of the classical coiled-coil type comprised of the a and d hydrophobic residue faces (denoted HP), hydrophobic cores comprised of alanine residues from the f positions (denoted ALA), and mixed cores comprised of a-d hydrophobic residue faces in contact with alanines from f positions (denoted ALA-HP) were initially investigated. Each oligomeric state was equilibrated for at least 100 ns. Dimeric, trimeric, and tetrameric oligomers of the Cc peptide with the classical HPcores were found to form stable assemblies in low pH conditions (as were, to a lesser degree, dimers with mixed ALA-HP cores). The changes in free energies associated with the formation of oligomers with HP-cores were then obtained from multiple trials using the free energy perturbation (FEP) method26,27. The calculated free energy of association on a per peptide basis was lowest for HP-based tetramers (-13.2 ± 2.2 kCal/mol), followed by trimers (-10.3 ± 2.3 kCal/mol), and dimers (-4.6 ± 2.2 kCal/mol), in qualitative agreement with the SEC data. Several competing physical effects appear to play roles in the preference for tetramers of the Cc peptide at low pH (Figure 3A). The Glu in the core d position at residue 25, introduced to suppress oligomerization above the pH trigger point, also seems to play a destabilizing role at low pH in dimeric and trimeric oligomers, due to steric hindrance of the Glu side chain in the hydrophobic pocket. However, this Glu side chain appears to be well accommodated in the tetrameric state of association (Figure 3C). On the other hand, the Glu residues at the b and c positions interact with Glu and Lys residues at e and g



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 28

positions to stabilize the tetrameric state at low pH (Figure 3B). However, above the trigger pH, the negatively-charged Glu at residue 25 can no longer be accommodated in the pocket, leading to loss of oligomer stability and the concomitant suppression of the alpha-helical conformation of individual peptides by intra-molecular electrostatic repulsion between the many Glu residues in the solvent exposed faces. Recombinant protein design and production. A recombinant fusion protein CcECc (Mw=29.64kDa) with the silk repeating motif flanked by coiled-coil sequence was next created to be tested in terms of its e-gel forming ability and reversibility (Table 2). Constructs with only the E domain (Mw=24.46kDa) were also produced and used as a control. Expression vectors encoding the E and CcECc fusion proteins, each including a terminal hexa-histidine tag (Figure S1), were transformed into BL21 E.coli and grown in standard shake flask culture to an OD600nm=0.8. Protein expression was then induced at 22oC, and the fusion protein was purified through his-tag affinity chromatography. The yields of purified proteins from the soluble fraction of E. coli lysate were approximately 40-50 mg per liter of culture. SDS-PAGE gels indicated that the final eluted E and CcECc fusion proteins were approximately 90% pure (Figure S2). MADLI-TOF mass spectrometry also confirmed the identity of E and CcECc fusion proteins at the expected molecular weights (Figures S3C & S3B). For comparison, MADLI-TOF mass spectrometry data for the Cc peptide is also provided (Figure S3C). Amino acid analysis indicated that composition of the purified proteins were in good agreement with the designed sequence (Figures S4A & S4B). Characterization of CcECc protein in solution. We next performed studies to characterize the molecular properties of CcECc in solution. First, potentiometric titration was used to determine the pKa of the CcECc telechelic triblock construct. Titration curves for CcECc in deionized (DI) water and a control titration curve for DI water along are shown in Figure 4A. Due in part to the large number of Glu residues in the E block (containing 40% glutamic acid residues), a relative low value of pKa=3.7 was found for the CcECc solution, while DI water control exhibited a slightly acidic value of pKa=6.0, presumably due to dissolved CO2 from the atmosphere. Next, CD studies of both CcECc and E solutions (10 µM and 12 µM, respectively, in 10 mM buffers (described in the Methods section below) were performed as a function of pH and temperature. Figure 4B shows CD spectra of CcECc as a function of wavelength at 20°C over a pH range from 3.0 to 8.0, revealing a pH-dependent conformational transition between random coils and partially α-helical structures between pH 5 and 6, as was found previously for the Cc domains on their own. For comparison, no significant differences were observed in the CD spectra of the E protein (middle block only), shown at pH 3.0 and 8.0 (Figure 4B, inset), suggesting that the measured α-helical content is due to the associating Cc end-domains. Accordingly, the ellipticity ratio ρ for CcECc (black diamonds) shows a sharp increase upon reduction in pH from ~7 to ~5, while ρ for the reference E protein (blue triangles) is almost unaffected by pH (Figure 4C). The increase in ρ for CcECc at low pH mirrors the behavior observed for the solutions of Cc peptides previously discussed and thus is an indication of intra- and inter-molecular association of the amphiphilic Cc end-domains of the triblocks into bundles. In the case of CcECc, at low concentrations one expects the formation of compact micelle-like aggregates of the triblock proteins, as shown schematically in Figure 5A. In synthetic telechelic polymers, such aggregates are known as “flower-like” micelles. As with small molecule amphiphiles, there is a thermodynamic transition with increasing concentration from individual solvated chains in very dilute conditions, where entropic considerations dominate, to compact micellar aggregates above a critical micelle concentration (CMC) of CcECc protein at sufficiently low pH. A combination of optical and electrophoretic methods was used to characterize the micelle formation process. Figure 5B shows spectrophotometry data of the optical density at 400 nm for a series of aqueous CcECc solutions at pH 3.5 and 10 as a function of concentration. For the acidic buffer (pH 3.5, red circles), the optical density rises first rapidly with concentration of CcECc until a cusp is reached, beyond which there is a very gradual rise in optical density. The concentration corresponding to this cusp in optical density is the CMC, φcmc ~0.6 mg/mL. Above this CMC, the sample


Page 11 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


becomes progressively more turbid, as is forms increasingly large superstructure (see inset). Association only occurs at sufficiently low pH for the end domains of the protein to exist in an amphiphilic state. Thus, the solutions prepared at pH 10, for which the proteins exist in a non-associating disordered coil state, are essentially transparent (see figure inset) over a wide range of concentrations and therefore have an optical density that only rises slowly with concentration, shown in the blue squares in Figure 5B, and therefore no measureable CMC. Dynamic light scattering measurements were used to obtain estimates of the hydrodynamic radius and zeta potential of aggregates in various solution conditions. Figure 5C shows data for apparent hydrodynamic radius as a function of pH for samples with concentration ~0.7 mg/mL (slightly above the CMC). For low pH conditions (pH < 5), where the associating helical end-domains are functional (as determined by the CD studies above) micelles of mean size 100 nm are observed. At higher pH values (pH > 6), objects of mean size ~300nm are observed. These are likely not well-defined micelles, but rather very loosely associated denatured protein chains, as these pH conditions correspond to fully charged Glu side-chains and unstructured end-domains. Indeed, with increasing pH we expect progressively more charged and swollen coils of the CcECc construct, as is clearly shown by the zeta potential data in Figure 5D. Characterization of CcECc protein e-gel phases. Beyond the CMC, there is a second, continuous transition in solutions of telechelic polymers from the assembly of distinct micelles to the formation of a continuous network, due to the eventual preference for inter-molecular vs intra-molecular association. The same transition should occur for our telechelic protein hydrogels in appropriate pH conditions. Hydrogel forming conditions can be achieved, in principle, by preparing bulk samples at sufficiently high concentration and low pH. Alternatively, such conditions can occur by accumulation of protein at an appropriate surface, for instance at the anode of an electrochemical cell. We developed two such electrochemical cells for characterization of the e-gel process in CcECc. The first e-gel cell was composed of a milli-fluidic rectangular flow channel with two parallel gold-coated interior side walls that is viewed from above in an inverted microscope (Figure S4). When a sufficiently large voltage is applied between the two parallel gold-coated walls, the growth of an e-gel phase on one electode surface results. For instance, a +12V applied voltage is seen in Figure 6A to generate e-gel formation upward from the bottom anode surface. The anode releases hydrogen ions (H+), which migrate toward the cathode. Thus, the vicinity of the anode is a low-pH environment (as shown by the orange color of pH indicator added to the buffer), where the associating Cc domains of the telechelic proteins are active. Likewise, the vicinity of the cathode is enriched in OH- ions that turn the pH indicator purple. Moreover, since the CcECc has a net negative charge everywhere except in the vicinity of the anode, it will be driven electrophoretically from the cathode to the anode, enriching the concentration of e-gel protein at the anode surface. The e-gel growth and dissolution kinetics was monitored optically near the anode using solutions (~1.0 mg/mL in water, without added pH indicator for clarity) in response to a series of applied voltages across the cell. Figure 6B shows the e-gel response to the following time series of applied voltages: +12V for 5 s, 0V for 25 s, -12V for 5 s, and 0V for the last 5 s. Immediately upon switching on the positive voltage, the opaque e-gel layer begins to grow rapidly. During the first ~10 s of the off period, the front edge of the growing e-gel kept advancing, until the thickness reached a transient plateau (the e-gel state would eventually dissolve on its own in the absence of an applied potential). Upon the subsequent reversal of voltage at t=30 s, the e-gelled CcECc dissolved rapidly to zero thickness in ~5 s. This process is reproduced with subsequent applications of this applied voltage time series. The second e-gel cell was composed of a 9 mm diameter reservior in a PDMS spacer sheet 3mm in thickness that was capped by a lower electrode surface (Al foil) and open at the top, where an Al electrode contacts the free surface. Figure 7 shows a series of images that follow the e-gel process for a ~1.0 mg/mL CcECc solution (Figure 7A) subjected to a potential difference of 12V between lower and upper Al foil electrodes, corresponding to an electric field of 4 V/mm across the gap. The e-gel phase formed rapidly on the lower anode after application of the 12V potential difference (Figure 7B) and subsequently dissolved after turning off the potential difference (Figure 7C & D), in a reversible manner.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 28

The lifetime of the e-gel state in the absence of applied potential is shorter in this geometry (3-5 s) than in the channel geometry shown in Figure S5 (~30 s, see Figure 6B), presumably due to differences in the electrochemical conditions. To investigate the properties of the e-gel phase, we performed electron microscopy and particletracking microrheology measurements. Figure S6B shows cryo-TEM images of solutions before and after in situ electrogelation on a carbon TEM grid (see Methods section for details). While the solution phase of CcECc showed no characteristic microstructural features (Figure S6A), the e-geled phase exhibits a textured morphology with a characteristic length scale in the range 50-100 nm (Figure S6B). To probe the viscoelastic features of the CcECc system, we performed particle-tracking measurements, using standard video microscopy methods (see Methods section for details), of the thermal fluctuations of a dilute suspension of 5 µm polystyrene latex particles added to CcECc protein solutions residing in the 9 mm circular electrochemical cell with transparent glass upper and lower electrodes under several electrochemical conditions. Image analysis of the video was used to extract trajectories of the individual particles. Control experiments were also performed for tracer particles in pure water. Figure 8A shows typical trajectories for tracer particles in pure water (black), in pH 7.4 CcECc solution (green), in pH 4 CcECc solution (blue), and in an e-gel layer near the anode surface immediately after switching off a 12V electrochemical potential (red). In pure water, the tracer particles exhibited unrestricted diffusion in a purely viscous solvent, while in the buffered solutions the particle motion was increasingly inhibited with decreasing pH, due to increasing steric hindrance by the network of CcECc proteins that envelop the particles. At pH 7.4, the CcECc proteins formed a weakly entangled and more viscous solution that only modestly slows tracer motion, while at pH 4, transient viscoelastic network structures formed by the association of the telechelic proteins acted to more strongly restrict tracer motion. Finally, in the e-gel phase, the particles were seen to be well confined within a concentrated elastic gel phase, as they only executed very small amplitude fluctuations around their mean positions (the motion of a tracer particle in the e-gel phase is so restricted that the particle seems stationary over a one minute time window compared with other samples shown in Figure 8A). Analysis of the trajectory data allowed for the calculation of the average mean squared displacement, , of the tracer particles as a function of time, as shown in Figure 8B. For tracer particles in water (black), ∝ t , confirming the pure diffusive motion of the control. For tracer particles in buffered CcECc solutions, the motion was somewhat suppressed and x weakly sub-diffusive, ∝ t with x

Electrochemically Directed Assembly of Designer Coiled-Coil

Recommend Documents