Article pubs.acs.org/Biomac
Engineering of an Environmentally Responsive Beta Roll Peptide for Use As a Calcium-Dependent Cross-Linking Domain for Peptide Hydrogel Formation Kevin Dooley,† Yang Hee Kim,†,§ Hoang D. Lu,†,∥ Raymond Tu,‡ and Scott Banta*,† †
Department of Chemical Engineering, Columbia University, New York, New York, United States Department of Chemical Engineering, The City College of New York, New York, United States
‡
S Supporting Information *
ABSTRACT: We have created a set of rationally designed peptides that form calcium-dependent hydrogels based on the beta roll peptide domain. In the absence of calcium, the beta roll domain is intrinsically disordered. Upon the addition of calcium, the peptide forms a beta helix secondary structure. We have designed two variations of our beta roll domain. First, we have mutated one face of the beta roll domain to contain leucine residues so that the calcium-dependent structural formation leads to dimerization through hydrophobic interactions. Second, an α-helical leucine zipper domain is appended to the engineered beta roll domain as an additional means of forming intermolecular cross-links. This full peptide construct forms a hydrogel only in calcium-rich environments. The resulting structural and mechanical properties of the supramolecular assemblies are compared with the wild-type domain using several biophysical techniques including circular dichroism, FRET, bis-ANS binding and microrheology. The calcium responsiveness and rheological properties of the leucine beta roll containing construct confirm the potential of this allosterically regulated scaffold to serve as a cross-linking domain for stimulus-responsive biomaterials development.
■
INTRODUCTION The development of smart materials and hydrogels has opened the door to a host of potential applications in such fields as drug delivery, tissue engineering, and microfluidics.1−8 Hydrogels are composed of water-soluble monomers that are physically or covalently cross-linked to form 3D polymer networks. This cross-linking can often be controlled by the incorporation of stimulus-responsive proteins or peptides into the monomeric building block. Stimuli such as pH, temperature, or ionic strength can be used to regulate the assembly of hydrogel networks. Several examples of protein domains that facilitate environmentally responsive gelation have been described, including elastin-like peptides, calmodulin, and αhelical leucine zipper domains.1,9,10 α-Helical leucine zippers are a common structural motif found in DNA binding proteins. The name is derived from the periodic repeat of leucine residues that protrude outward and align along a side of the helix, creating a hydrophobic driving force to form “zipped” coiled coil bundles. These domains have been extensively characterized and have proven useful in developing stimulus-responsive hydrogels as they assemble and dissociate in response to changes in temperature and pH.11,12 Previously, we have appended these domains to enzymes and other globular proteins to create bifunctionalized hydrogel constructs.13−15 In this work, we present a new domain suitable for creating physical cross-links and a hydrogel network. We have used a © 2012 American Chemical Society
beta roll peptide scaffold taken from the block V repeats-intoxin (RTX) domain of adenylate cyclase from Bordetella pertussis. The beta roll motif is a modular repeated sequence that has been shown to be intrinsically disordered in the absence of calcium.16,17 In calcium-rich environments, the peptide forms a beta helix consisting of two short parallel β sheet faces separated by turns. A highly conserved aspartic acid residue in each turn is responsible for calcium binding.18 A Cterminal capping group responsible for entropic stabilization is required for conformational response to calcium. We have previously characterized the calcium responsiveness and capping requirements of the native beta roll domain.19−21 Each beta strand in the folded beta roll domain contains two amino acids that protrude radially from the scaffold and are exposed to the solvent. We hypothesize that mutating the protruding residues on one face to leucine side chains will create a hydrophobic surface suitable for cross-linking in response to calcium binding (Figure 1). Therefore, we have rationally designed an allosterically regulated beta roll motif that should form calcium-dependent cross-links suitable for hydrogel formation. Hydrogel networks modulated by calcium-induced conformational change differ from conventional approaches that rely heavily on pH and Received: February 15, 2012 Revised: April 29, 2012 Published: April 30, 2012 1758
dx.doi.org/10.1021/bm3002446 | Biomacromolecules 2012, 13, 1758−1764
Biomacromolecules
Article
along with the formation of the leucine zipper bundles should provide the physical interactions necessary to form a hydrogel (Figure 2). Gel formation and strength will be modulated using the calcium-dependent allosteric control that is intrinsic to the peptide sequence.
Figure 2. Hydrogel formation. (a) Calcium-induced conformational change of beta roll. In the absence of calcium, the beta roll remains disordered. Upon the addition of calcium, the beta roll undergoes a reversible structural change forming the corkscrew-like beta roll structure. The beta roll is depicted face forward. Calcium ions are shown in red. (b) Hydrogel monomeric building block. The α-helical leucine zipper domain (H) is shown in yellow with the soluble linker domain (S) in blue. The mutant leucine beta roll with the C-terminal capping region is shown in green. (c) Hydrogel transition. Prior to the addition of calcium, the helical domains can form tetrameric bundles, but the beta roll domains remain unstructured. When calcium is added, the folded beta roll domains expose the leucine rich faces, enabling cross-linking and hydrogel network formation. Some folded beta rolls are depicted from a side view, showing how two leucine faces could cross-link.
Figure 1. Beta roll structures. (a) Crystal structure of an extracellular lipase from Pseudomonas sp. MIS38 (PDB 2Z8X) containing a beta roll domain. The folded beta roll domain can be seen in the lower right with the bound calcium ions in red. The structure of this folded beta roll was used as a template to model the block V beta roll domain of adenylate cyclase from B. pertussis. Homology models were generated using SWISS-MODEL, which are shown in panels b and c (Swiss Institute of Bioinformatics). The positions of the bound calcium ions were approximated manually. (b) Model of the WT adenylate cyclase beta roll peptide used in this work along with its primary sequence. The surface-exposed residues in the folded conformation are highlighted in magenta with the residues underlined in the sequence. Calcium ions are shown in red. (c) Model of the mutant leucine beta roll engineered in this work along with its primary sequence. The leucine mutations to the WT beta roll are shown in blue and underlined in the sequence.
Through circular dichroism (CD) spectroscopy, bis-ANS binding, and terbium binding, we show that the leucine mutations have minimal effect on the response of the peptide to calcium. After appending the H and S domains to the wild type (WT) and engineered leucine beta roll (forming HS-WT and HS-leucine respectively), rheological analysis confirms that the hydrogel formation is a result of the leucine mutations as the HS-WT beta roll does not self-assemble in calcium-rich environments.
■
temperature swings to destabilize intermolecular cross-links. Allosteric regulation allows for the precise tuning of gel formation and strength by simply adjusting the calcium concentration, making these gels suitable for applications in systems that do not permit fluctuations in temperature or pH. To assemble a macromolecular hydrogel network, we have fused an α-helical leucine zipper domain (H) to the N-terminus of the leucine beta roll through a soluble linker region (S). Both the H and S domains have been previously characterized.12,22,23 In the absence of calcium, the beta roll should remain disordered, whereas the leucine zippers can form tetrameric coiled coil bundles. The lack of interaction between the disordered beta roll domains should prohibit the formation of a hydrogel. Upon the addition of calcium, the beta roll should undergo a conformational change forming the leucine-rich face. The hypothesized cross-linking between beta roll domains
EXPERIMENTAL SECTION
Materials. The maltose binding protein (MBP) expression kit and all enzymes were purchased from New England Biolabs (Ipswich, MA). Isopropyl β-D-1-thiogalactopyranoside (IPTG) was obtained from Promega (Madison, WI). Halt protease inhibitor cocktail was purchased from Fisher Scientific (Waltham, MA). Amicon centrifugal filters were purchased from Millipore (Billerica, MA). Native PAGE gels, running buffer, protein ladder, and SimplyBlue SafeStain were obtained from Life Technologies (Grand Island, NY). All chemicals and other reagents were purchased from Sigma-Aldrich (St. Louis, MO) unless otherwise specified. Cloning into pMAL and pQE9 Vectors. Both WT and leucine beta roll proteins are expressed using a modified pMAL vector. The intein domain from pET-EI/OPH, a gift from Dr. David Wood (Ohio State University, Columbus, OH), has been cloned into the pMAL vector. (See the Supporting Information for all oligonucleotide sequences.)24 The WT beta roll and the C-terminal capping region 1759
dx.doi.org/10.1021/bm3002446 | Biomacromolecules 2012, 13, 1758−1764
Biomacromolecules
Article
have been amplified out of pDLE9-CyaA, a gift from Dr. Daniel Ladant (Institut Pasteur, Paris, France), using PCR primers with KpnI and HindIII restriction sites for ligation into the pMAL-intein vector.16 The mutant beta roll is constructed by inserting the appropriate leucine codons into two overlapping oligonucleotides encoding for the entire beta roll sequence. The oligonucleotides are annealed and extended to produce the full-length double-stranded leucine beta roll. The C-terminal capping domain is added by overlap extension PCR. KpnI and HindIII sites are added to the capped leucine beta roll before ligation into pMAL-intein. Both HS-WT and HS-leucine beta rolls are expressed using a modified pQE9AC10Acys vector, a gift from Dr. David Tirrell (California Institute of Technology, Pasadena, CA).22 In this work, AC10Acys is termed H−S−H. Both beta roll genes have been amplified by PCR using primers with SphI and SpeI restrictions sites for subsequent cloning into pQE9, which had been previously modified to remove the C-terminal H domain. pMAL vectors are transformed into OmniMAX (Invitrogen) and pQE9 vectors are transformed into SG13009 (QIAGEN) E. coli strains for expression. A schematic of the constructs is provided in Figure 3.
purified using immobilized metal affinity chromatography and a polyhistidine tag. Cultures of Terrific Broth (TB, 1 L) are supplemented with 50 μg/mL kanamycin and 200 μg/mL ampicillin prior to inoculation from an overnight culture of the appropriate vector. Protein expression is induced by the addition of IPTG to a final concentration of 0.5 mM after an OD600 = 0.6 is reached. The expression is carried out for 5 h at 37 °C with shaking. Cells are pelleted and resuspended in 25 mL of HisA buffer (20 mM Tris-HCl, 150 mM NaCl, 40 mM imidazole, pH7.5) supplemented with Halt protease inhibitor cocktail. Cells are harvested and lysed as previously described. Samples are loaded on to a 5 mL nickel-charged HisTrap FF column (GE Healthcare) equilibrated in HisA. Unbound protein is eluted with 10 column volumes of HisA, and the his-tagged construct is eluted with HisB buffer (20 mM Tris-HCl, 150 mM NaCl, 500 mM imidazole, pH 7.5) using a linear gradient to 100% HisB over 20 column volumes. Fractions containing the desired protein are collected and confirmed by SDS-PAGE. Samples are desalted and concentrated by ultrafiltration using 30 kDa MWCO Amicon centrifugal filters. Increased purity can be achieved by size exclusion chromatography. Typical yields ranged from 20 to 30 mg of pure protein per liter culture. Full peptide sequences and SDS-PAGE results are available in the Supporting Information. Circular Dichroism Spectroscopy. CD spectroscopy is performed as previously described.19 In brief, 100 μM samples are loaded into a 0.01 cm path length quartz cuvette and analyzed on a J-815 CD spectrometer (Jasco) equipped with a Peltier junction temperature controller. All measurements are performed in triplicate in 50 mM Tris pH7.5 at 25 °C. Titration data are fit using SigmaPlot (Systat Software) nonlinear regression software. bis-ANS Binding Fluorescence Spectroscopy. Protein samples (250 nM) are loaded in a 1 cm path length cuvette and equilibrated with 0 or 50 mM calcium prior to the addition of 1 μg/mL bis-ANS. Changes in fluorescence emission are measured from 420 to 600 nm using a FMO-427S monochromator (Jasco). Excitation is at 390 nm. All measurements are performed in triplicate in 50 mM Tris pH7.5 at 25 °C. Fluorescence Resonance Energy Transfer. 1 μM protein samples are titrated with terbium chloride. Following excitation of the sample at 282 nm, changes in fluorescence emission from bound terbium ions are monitored at 545 nm. All experiments are performed in 96-well plates (Costar) in 20 mM PIPES pH 6.8, 120 mM NaCl, 10 mM KCl. Terbium is incubated with the protein samples for 30 min at 25 °C prior to reading. All data are fit using SigmaPlot nonlinear regression software. Native Polyacrylamide Gel Electrophoresis (PAGE). Samples (2 μg) of leucine and WT beta roll are run on 4−16% Bis-Tris 1.0 mm gels. The voltage is held constant at 150 V, and the run time is set to 105 min. For runs completed with calcium, 5 mM CaCl2 is added to the running buffer. The gels are stained with SimplyBlue SafeStain according to the manufacturer’s protocol. Hydrogel Preparation. Hydrogel constructs are allowed to selfassemble by reconstituting lyophilized protein with small volumes of water, as previously described.14 HS-WT and HS-leucine beta roll concentrations are determined by UV absorbance at 280 nm using the extinction coefficients ε280 = 24 750 M−1 cm−1 and ε280 = 23 470 M−1 cm−1 respectively. Protein (1.5 mg) is diluted in 250 μL of 5 mM TrisHCl pH7.5 with the appropriate salt concentration. Samples are frozen overnight at −80 °C and lyophilized the following day. The lyophilized protein is rehydrated with 25 μL of Millipore water yielding 6 wt % samples. Mechanical mixing, vortexing, and centrifugation are used to ensure that all of the protein is rehydrated. The samples are centrifuged for 5 min at 13 000 g to remove any air bubbles and allowed to set. Microrheology. Microrheology is used to analyze the mechanical properties of a viscoelastic fluid by following the motion of micrometer-sized spherical particles embedded in the sample. In this study, passive microrheology is used, which relies on the intrinsic Brownian motion of the particles caused by small thermal fluctuations.25,26 The particles’ mean square displacements (MSDs)
Figure 3. Schematic of beta roll constructs. WT and leucine beta rolls were expressed as fusions to maltose binding protein (MBP) and purified by intein cleavage. HS-WT and HS-leucine beta rolls were expressed using the pQE9 vector and purified using polyhistadine tags. H represents an α-helical leucine zipper domain and S represents a randomly coiled linker domain. The blocks are not drawn to scale. Expression and Purification of WT and Leucine Beta Rolls. The WT beta roll and leucine beta roll constructs are expressed identically in LB media containing 2 g/L D-glucose. 1 L cultures (supplemented with 100 μg/mL ampicillin) are inoculated from an overnight culture of the appropriate pMAL-intein vector. The 1 L cultures are incubated at 37 °C with shaking until an OD600 = 0.6 is reached. Protein expression is induced by the addition of IPTG to a final concentration of 0.3 mM. Expression is carried out for 2 h at 37 °C with shaking. Cells are pelleted at 3000g for 15 min, and the supernatant is discarded. The cell pellets are resuspended in 25 mL of MBP column buffer (20 mM Tris-HCl, 200 mM NaCl, 1 mM EDTA, pH7.4) containing Halt protease inhibitor cocktail. Cells are lysed by sonication with a microtip sonicator for 6 min on ice (Misonix Sonicator 3000). The lysate is clarified by centrifugation at 15 000g for 30 min, after which the pellet is discarded. The supernatant is diluted five-fold with MBP column buffer and loaded onto amylose resin columns, according to the manufacturer’s protocol (New England Biolabs). The columns are washed, capped, and filled with 8 mL of intein cleaving buffer (137 mM NaCl, 2.7 mM KCl, 8.1 mM Na2HPO4, 1.76 mM KH2PO4, 40 mM bis-Tris, 2 mM EDTA, pH6.2) for incubation at 37 °C for 12−16 h. Cleaved beta roll is eluted with 50 mL of MBP column buffer, concentrated in 10 kDa MWCO Amicon centrifugal filters, and buffer exchanged with 20 mM bis-Tris, 25 mM NaCl, pH6.0. The samples are run over a 16/10 Q FF ionexchange column (GE Healthcare) using an Ä KTAFPLC (GE Healthcare). Separation between MBP fusions and cleaved beta roll is achieved using an NaCl gradient from 25 to 500 mM over 20 column volumes. Beta roll fractions are collected and desalted prior to SDS-PAGE. Sample concentrations are determined by absorption at 280 nm (Spectramax M2, Molecular Devices) using calculated extinction coefficients (WT, ε280 = 17 780 M−1 cm−1; leucine, ε280 = 16 500 M−1 cm−1). Typical yields ranged from 3 to 7 mg of pure protein per liter of culture. Full peptide sequences and SDS-PAGE results are available in the Supporting Information. Expression and Purification of HS Constructs. Both HS-WT and HS-leucine beta roll constructs are expressed identically and 1760
dx.doi.org/10.1021/bm3002446 | Biomacromolecules 2012, 13, 1758−1764
Biomacromolecules
Article
can be calculated experimentally and are related to the mechanical properties of the fluid through a generalized Stokes−Einstein equation
⟨Δr 2̃ (s)⟩ =
dkBT 3πasG̃(s)
where ⟨Δr̃2(s)⟩ is the time-averaged Laplace transform of the particles’ MSD, d is the dimensionality of the track (2 for this work), kB is the Boltzmann constant, T is the temperature, a is the radius of the tracer particle, s is the Laplace frequency, and G̃ (s) is the frequencydependent Laplace representation of the complex modulus. This is composed of both the elastic (G′) and viscous (G″) moduli.12,27 When reconstituting the lyophilized protein, 1 μm fluorescently labeled tracer particles are added. The samples are thoroughly mixed, loaded onto a glass microscope slide between two strips of Parafilm, and sealed with a glass coverslip. Particle motion is observed using a Nikon Eclipse 50i microscope with a 40× objective. Three hundred frames of video are recorded per run at an exposure time of 33 ms with a Nikon HRD076 camera. Three separate videos are taken per sample to ensure a good statistical average. Readings are made in the middle of each sample so that edge effects can be neglected. Image stacks are created using ImageJ and analyzed using Interactive Data Language (IDL) software. The particle trajectories and rheological properties of each sample are calculated using algorithms created by Crocker et al.28
■
RESULTS AND DISCUSSION Leucine Beta Roll Characterization. A beta roll domain from the adenylate cyclase protein from B. pertussis was rationally mutated to contain a leucine-rich face, and the mutant peptide was expressed, purified, and characterized. Experiments are conducted to compare the mutant construct to the WT to ensure that the mutations did not disrupt the response of the peptide to calcium. CD spectroscopy, bis-ANS dye binding, and terbium binding experiments all suggest that the mutated beta roll exhibits a similar calcium-induced conformational change and calcium binding affinity as compared with the WT peptide. In the absence of calcium, both leucine and WT beta roll CD spectra exhibit a large negative peak at 198 nm indicative of randomly coiled polypeptide. Upon the addition of 50 mM calcium, both constructs show a similar increase in β-sheet secondary structure with a negative peak emerging at 218 nm (Figure 4a,b). These results are consistent with what has been previously reported.19 A calcium titration is performed by monitoring the change in CD signal at 218 nm (Figure 4e). Both data are fit to the Hill equation yielding nearly identical calcium binding affinities and binding cooperativity (Table 1). Bis-ANS binding is used to probe further structural changes in response to calcium. Bis-ANS binds to surface exposed hydrophobic patches resulting in an increase in fluorescence signal in calcium-rich environments.19,29 Both WT and leucine beta roll fluorescence signals are measured in the presence and absence of 50 mM calcium (Figure 4c,d). The slightly elevated signal in Figure 4d is a result of the additional hydrophobic residues in the leucine beta roll. Fluorescence resonance energy transfer (FRET) experiments are performed to confirm the CD and bis-ANS binding data. Terbium, a lanthanide atom, is titrated into beta roll samples. The emission from bound terbium ions in close proximity to excited tyrosine residues is measured spectrophotometrically.30,31 It is important to note that whereas terbium is often used as a calcium analog it does not directly indicate calcium binding. However, when analyzed in coordination with the CD and bis-ANS data, terbium binding does corroborate the claim that both constructs bind calcium in a similar manner. The terbium titrations for the WT and leucine beta roll yield
Figure 4. WT and leucine beta roll calcium responsiveness and characterization. (a) WT and (b) leucine beta roll CD spectra in the presence (···) and absence () of 50 mM calcium showing similar responses. These results are consistent with bis-ANS binding results for WT and leucine beta rolls shown in panels c and d, respectively. The higher bis-ANS signal observed for the leucine construct is due to the increased number of nonpolar residues. The CD calcium titration (e) shows nearly identical curves for both WT (●) and leucine (○) beta roll proteins. The data are fit to the Hill equation, and the parameters are summarized in Table 1. Terbium binding results are shown in panel e for the WT (●) and leucine (○) constructs. Both show very similar responses.
Table 1. Summary of Calcium Binding Affinity and Hill Coefficients for the WT and Leucine Constructs construct
KD (mM)
nH
R2
WT beta roll leu beta roll
0.91 ± 0.02 0.87 ± 0.02
2.91 ± 0.2 2.97 ± 0.2
0.993 0.996
KD values of 0.97 ± 0.04 and 0.91 ± 0.04 mM, respectively (Figure 4f). WT and leucine constructs are analyzed by native PAGE to confirm the oligimerization state of the mutant beta roll in the presence of calcium. Both samples migrate similarly through the native gel in the absence of calcium (Figure S4, Supporting Information). Upon the addition of 5 mM calcium to the running buffer, there is a clear difference in migration between the WT and leucine beta roll. The leucine beta roll appears to run larger, suggesting the formation of an oligomer, most likely caused by the cross-linking of the leucine-rich faces. Whereas these gels are not entirely quantitative, they do suggest an apparent difference in size, only in the presence of calcium. 1761
dx.doi.org/10.1021/bm3002446 | Biomacromolecules 2012, 13, 1758−1764
Biomacromolecules
Article
HS-Leucine Beta Roll Characterization. An α-helical leucine zipper domain (H) along with a randomly coiled soluble domain (S) is fused to the N-terminus of the leucine and WT beta rolls. Similar characterization experiments are performed to determine if these domains have any effect on calcium-induced structural change. The CD spectra do show changes in response to calcium, but the signal is dominated by the largely helical content of the H domain (Figure S3, Supporting Information). Bis-ANS binding experiments show no discernible difference between the HS-WT and HS-leucine proteins (Figure 5a,b). It is important to note that the
(Data not shown). At weight percentages of 6%, we observe a calcium-dependent phase transition, and behavior at this concentration is further explored. Samples (6 wt %) of HS-WT and HS-leucine beta roll are prepared as described above. After tracer particles are imbedded in the samples, video microscopy is used to record the motion of the particles. The trajectories and mechanical properties are calculated using IDL software. The viscous (G″) and elastic (G′) moduli of HS-WT and HS-leucine as a function of frequency are shown in Figure 6. Both WT and leucine
Figure 5. HS-WT and HS-leucine beta roll calcium responsiveness and characterization. (a) HS-WT and (b) HS-leucine beta roll bis-ANS binding spectra in the presence (···) and absence () of 50 mM calcium. These data are consistent with the bis-ANS data presented in Figure 4, with the increases in baseline fluorescence due to binding to H and S domains. Terbium binding titrations are shown in panel c for the HS-WT (●) and HS-leucine (○) constructs. CD spectra are included in the Supporting Information.
fluorescence signals are much higher in this case due to bisANS binding to the H and S domains. However, the relative change in peak intensity upon the addition of calcium is conserved among all four constructs (∼2.5× higher). The terbium titrations for HS-WT and HS-leucine (Figure 5c) yield KD values of 0.93 ± 0.03 and 0.86 ± 0.02 mM, respectively. HS-Leucine Beta Roll Microrheology. Mechanical properties of HS-WT and HS-leucine beta roll samples are characterized using microrheology. The Brownian motion of small particles imbedded into the sample is recorded using video microscopy, and the average MSD of the particles is calculated as a function of time. The MSD of the tracer particles allows one to quantify the mechanical properties of the fluid where they are embedded.12,27 Once the MSD is obtained, the frequency-dependent viscous and elastic moduli of a sample can be calculated using the modified Stokes−Einstein equation given in the Experimental Section. Both constructs demonstrate concentration-dependent gelation. Initial experiments show that at protein concentrations below 5 wt % the samples remain viscous with and without calcium. Conversely, at weight percentages above 10%, the samples are completely elastic
Figure 6. HS-WT and HS-leucine beta roll microrheology. Viscous (○) and elastic (●) moduli have been calculated for 6 wt % HS-WT and HS-leucine beta roll samples. Both constructs appear mostly viscous in buffer (a,b) and in 50 mM magnesium (c,d). The HS-WT beta roll remains viscous in 50 mM calcium as well (e). The HSleucine sample shows a shift in mechanical properties after the addition of calcium, gaining elasticity as compared with the HS-WT control. MSD versus τ plots are included in the Supporting Information for all microrheology experiments.
constructs appear to be viscous liquids in buffer (Figure 6a,b) and in the presence of 50 mM magnesium (Figure 6c,d). However, whereas the WT construct remains viscous in 50 mM calcium (Figure 6e), the leucine construct forms a hydrogel (Figure 6f). To supplement this data, we performed a calcium titration with 6 wt % HS-leucine beta roll samples. The calcium concentration is varied from 0 to 10 mM, and the resultant rheological plots are given in Figure 7. Large shifts in the elastic and viscous moduli are observed at 500 μM calcium with a crossover frequency of ∼3 s−1. As the calcium concentration is 1762
dx.doi.org/10.1021/bm3002446 | Biomacromolecules 2012, 13, 1758−1764
Biomacromolecules
Article
not play an important role in protein folding.19 Although the CD spectra of the constructs containing the H and S domains are dominated by the highly helical H domain, there appear to be conformational changes following the addition of calcium. The microrheology data present substantial differences in viscoelastic properties between the HS-WT and HS-leucine beta roll peptides in the presence of calcium. At 6 wt %, both constructs exhibit viscous character in buffer and in buffer supplemented with 50 mM magnesium. The magnesium control shows that ionic effects did not influence the changes in mechanical properties of both samples. When calcium is added to the HS-WT protein, the sample remains viscous. Here we suspect that the WT beta roll is fully folded, as indicated by the CD data. However, this calcium-induced structural response does not promote the formation of a hydrogel network because the WT beta roll domains do not interact. Upon the addition of calcium to the HS-leucine beta roll, we observe a significant change in rheological properties. The sample appears to be elastic, showing frequency-independent viscous and elastic moduli. At 50 mM calcium, we expect the leucine beta roll to be completely folded, and the engineered hydrophobic leucine face should be exposed to the solvent. This hydrophobic face promotes cross-linking of the beta roll domains. The calciumdependent physical cross-linking coupled to the coiled coil bundles formed by the leucine zipper domains provides enough interaction to alter the mechanical properties of the sample and create a hydrogel network. It may also be possible for the leucine zipper domains to interact with the leucine beta roll domains, and this would introduce a different mode of crosslinking within the hydrogels. The transition from viscous liquid to hydrogel shown in Figure 6 is consistent with the leucine beta roll CD titration in Figure 4, where we observe the transition from disordered to structured peptide between 0.5 and 3.0 mM calcium. At concentrations higher than 3 mM, the beta roll is completely folded. A strong parallel can be drawn with the rheology data in Figure 6. The sample was found to transition from a viscous liquid to a hydrogel between 0.5 and 5 mM calcium. By 10 mM, the hydrogel is completely formed as the beta roll domains should be completely folded, maximizing the cross-linking. Most stimulus-responsive hydrogels described in the literature use cross-link forming scaffolds; then, a trigger is found to destabilize the binding interaction. For example, the leucine zipper-based hydrogels are destabilized by changes in pH because this interferes with the α-helix formation, and the elastin-like peptide based hydrogels take advantage of the unique inverse temperature transition of these peptides to destabilize the hydrogel. We have taken a different approach in this work, where we have chosen a scaffold that undergoes a specific and unique conformational transition from an intrinsically disordered structure to the folded beta roll domain in response to calcium. The beta roll domain is not normally involved in biomolecular recognition or self-assembly, so this feature was engineered into the scaffold to control self-assembly by calcium addition. Eliminating the reliance on temperature and pH swings to modulate self-assembly may allow for the use of these hydrogels in more biologically relevant environments, where changes in temperature or pH are not tolerated. Although the gels produced in this proof-of-concept work have relatively low elastic moduli, the leucine beta roll can be optimized in future work to create stronger hydrogels for different applications. Because the beta roll is a modular repeat protein, the number of repeats and composition of the
Figure 7. HS-leucine beta roll calcium titration. Viscous (○) and elastic (●) moduli have been calculated for 6 wt % HS-leucine beta roll samples at various calcium concentrations showing the transition from viscous liquid to hydrogel. MSD versus τ plots are included in the Supporting Information for all microrheology experiments.
increased, the crossover frequency continues to decrease. At 10 mM calcium, the sample is elastic in the frequency range we explored. Discussion. Several biophysical techniques are used in this work to probe the calcium binding, structural confirmation, and mechanical properties of the WT and leucine beta roll peptide constructs. We show that the leucine mutations made to the WT beta roll result in no change in calcium responsiveness or binding affinity; similar conformational changes are observed in the mutant beta roll as shown by CD and bis-ANS binding. This is expected as the residues selected for mutation most likely do not participate in calcium binding.18 Assuming the beta roll adopts a structure similar to those derived crystallographically in other RTX containing proteins, the amino acid side chains that are mutated project radially outward, away from the hydrophobic core minimizing any potential steric effects. Furthermore, native PAGE data indicates a calcium-dependent difference in migration between the mutant and WT proteins, likely caused by leucine beta roll cross-linking. This premise was elucidated through the rheological experiments after cloning both constructs into the pQE9 vector. Appending the H and S domains to the N-termini of the WT and leucine beta roll also results in minimal effects on response to calcium, as shown by the bis-ANS and terbium binding data. This is consistent with our previous results showing that the native N-terminal capping group of the beta roll domain does 1763
dx.doi.org/10.1021/bm3002446 | Biomacromolecules 2012, 13, 1758−1764
Biomacromolecules
Article
(7) Qiu, Y.; Park, K. Adv. Drug Delivery Rev. 2001, 53, 321. (8) Chockalingam, K.; Blenner, M.; Banta, S. Protein Eng., Des. Sel. 2007, 20, 155. (9) Banta, S.; Wheeldon, I. R.; Blenner, M. Annu. Rev. Biomed. Eng. 2010, 12, 167. (10) Topp, S.; Prasad, V.; Cianci, G. C.; Weeks, E. R.; Gallivan, J. P. J. Am. Chem. Soc. 2006, 128, 13994. (11) Kennedy, S. B.; Littrell, K.; Thiyagarajan, P.; Tirrell, D. A.; Russell, T. P. Macromolecules 2005, 38, 7470. (12) Xu, C.; Breedveld, V.; Kopeček, J. Biomacromolecules 2005, 6, 1739. (13) Lu, H. D.; Wheeldon, I. R.; Banta, S. Protein Eng., Des. Sel. 2010, 23, 559. (14) Wheeldon, I. R.; Barton, S. C.; Banta, S. Biomacromolecules 2007, 8, 2990. (15) Wheeldon, I. R.; Gallaway, J. W.; Barton, S. C.; Banta, S. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 15275. (16) Bauche, C. J. Biol. Chem. 2006, 281, 16914. (17) Chenal, A.; Guijarro, J. I.; Raynal, B.; Delepierre, M.; Ladant, D. J. Biol. Chem. 2009, 284, 1781. (18) Angkawidjaja, C.; Paul, A.; Koga, Y.; Takano, K.; Kanaya, S. FEBS Lett. 2005, 579, 4707. (19) Blenner, M. A.; Shur, O.; Szilvay, G. R.; Cropek, D. M.; Banta, S. J. Mol. Biol. 2010, 400, 244. (20) Szilvay, G. R.; Blenner, M. A.; Shur, O.; Cropek, D. M.; Banta, S. Biochemistry 2009, 48, 11273. (21) Shur, O. Protein Sci. 2011, 20, 925. (22) Shen, W.; Lammertink, R. G. H.; Sakata, J. K.; Kornfield, J. A.; Tirrell, D. A. Macromolecules 2005, 38, 3909. (23) Shen, W.; Zhang, K.; Kornfield, J. A.; Tirrell, D. A. Nat. Mater. 2006, 5, 153. (24) Fong, B. A.; Wu, W. Y.; Wood, D. W. Trends Biotechnol. 2010, 28, 272. (25) Cicuta, P.; Donald, A. M. Soft Matter 2007, 3, 1449. (26) Mason, T.; Ganesan, K.; Van Zanten, J.; Wirtz, D.; Kuo, S. Phys. Rev. Lett. 1997, 79, 3282. (27) Mason, T. G.; Weitz, D. Phys. Rev. Lett. 1995, 74, 1250. (28) Crocker, J. C.; Valentine, M. T.; Weeks, E. R.; Gisler, T.; Kaplan, P. D.; Yodh, A. G.; Weitz, D. A. Phys. Rev. Lett. 2000, 85, 888. (29) Takashi, R.; Tonomura, Y.; Morales, M. F. Proc. Natl. Acad. Sci. U.S.A. 1977, 74, 2334. (30) Huang, Y.; Zhou, Y.; Castiblanco, A.; Yang, W.; Brown, E. M.; Yang, J. J. Biochemistry 2008, 48, 388. (31) Yang, W.; Jones, L. M.; Isley, L.; Ye, Y.; Lee, H. W.; Wilkins, A.; Liu, Z.; Hellinga, H. W.; Malchow, R.; Ghazi, M. J. Am. Chem. Soc. 2003, 125, 6165.
repeating unit can be easily altered to extend the size and makeup of the hydrophobic domain. In addition, alternative cross-linking strategies could be incorporated such as the inclusion of specific ionic interactions as has been explored for the leucine zipper domains. Also, the beta roll has a second face amenable to mutation, which could be used to create leucinerich surfaces on both sides of the beta roll domain.
■
CONCLUSIONS In this work, we present a rationally engineered peptide that can be used to create allosterically controlled hydrogel networks. Leucine mutations are inserted into the beta roll peptide scaffold to create a hydrophobic surface suitable for cross-linking, which is exposed only after calcium binding. An α-helical leucine zipper domain with a randomly coiled linker are attached to the N-terminus of the beta roll to provide additional physical cross-linking. Hydrogels form only in calcium-rich environments, where the folded leucine beta roll domains provide the necessary hydrophobic interface. The WT beta roll remains a viscous liquid regardless of the calcium concentration.
■
ASSOCIATED CONTENT
S Supporting Information *
SDS-PAGE results for WT, leucine, HS-WT, and HS-leucine beta roll domains. CD results of HS-WT and HS-Leu beta roll domains at various calcium concentrations. Native PAGE results for WT and leucine beta roll domains with and without calcium. MSD versus τ plots for all microrheology experiments. Primary sequences and molecular weights for all constructs used in this work. Oligonucleotide sequences for all cloning experiments. This material is available free of charge via the Internet at http://pubs.acs.org
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Phone: (212) 854-7531. Present Addresses §
Food Ingredients Center, CJ CheilJedang, Seoul, South Korea. Department of Bioengineering, University of Pennsylvania, Philadelphia, Pennsylvania, United States. ∥
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We gratefully acknowledge funding from a MURI award from the Air Force Office of Scientific Research (AFOSR) [FA955006-1-0264 to S.B.], the Columbia Technology Ventures office, and the U.S. Department of Education Ronald E. McNair Postbaccalaureate Achievement Program [P217A070257 to H.D.L].
■
REFERENCES
(1) Ehrick, J. D.; Deo, S. K.; Browning, T. W.; Bachas, L. G.; Madou, M. J.; Daunert, S. Nat. Mater. 2005, 4, 298. (2) Endo, T.; Ikeda, R.; Yanagida, Y.; Hatsuzawa, T. Anal. Chim. Acta 2008, 611, 205. (3) He, C.; Kim, S. W.; Lee, D. S. J. Controlled Release 2008, 127, 189. (4) Kopecek, J. Biomaterials 2007, 28, 5185. (5) Lee, K. Y.; Mooney, D. J. Chem. Rev. 2001, 101, 1869. (6) Nettles, D. L.; Kitaoka, K.; Hanson, N. A.; Flahiff, C. M.; Mata, B. A.; Hsu, E. W.; Chilkoti, A.; Setton, L. A. Tissue Eng., Part A 2008, 14, 1133. 1764
dx.doi.org/10.1021/bm3002446 | Biomacromolecules 2012, 13, 1758−1764
