www.advhealthmat.de
FULL PAPER
www.MaterialsViews.com
Rational Design of Multifunctional Hetero-Hexameric Proteins for Hydrogel Formation and Controlled Delivery of Bioactive Molecules Xiaoli Zhang, Hao Zhou, Ying Xie, Chunhua Ren, Dan Ding,* Jiafu Long,* and Zhimou Yang* biomaterials because they mimick extracellular matrix of tissues and organs. The hydrogels containing proteins in the 3D networks are known as protein-based hydrogels. Due to the unique property of proteins (e.g., perfect polydispersity, bioactivity, exact control over monomer sequence, and ability to fine-tune molecular-level biochemical interactions), protein-based hydrogels have attracted extensive research interests in recent two decades,[4] and they have exhibited great potential for tissue engineering,[5] bioactive molecule sensing,[6] and regenerative medicine.[7] Protein-based hydrogels are usually formed by either thermo-inducement or mixing two components that are capable of forming covalent bonds or having specific non-covalent interactions. For example, coiled-coil peptides or biohybrids consisted of coiled-coil peptides can form oligomeric structures at elevated temperatures that serve as cross-linkers for hydrogel formation.[2a,8] The covalent strategy for protein-based hydrogel formation often utilizes recombinant proteins to crosslink polymers by Michael addition reactions, site selective conjugation, and enzyme-catalyzed ligations.[9] Furthermore, specific non-covalent interactions including protein–ligand,[7,10] protein–protein,[11] protein–peptide,[12,13] and protein–polysacharide interactions[14] have also been applied to form protein-based hydrogels. Noteworthy is that these hydrogels formed by specific non-covalent interactions are physical hydrogels that possess excellent injectable property.[4d] Besides, the simple mixing procedure for preparing physical hydrogels also renders its promising application in such as cell delivery[11,12,15] and controlled drug delivery.[6a,7] In order to obtain protein-based hydrogels by non-covalent interactions, the design and production of proteins with multiple binding sites are the key. Recombinant proteins with multiple binding sites can be prepared through conjugation with multi-armed polymers, by rational design to produce those with tandem repeating binding domains, and by generation of those that can form oligomer structures. For example, Kiick and co-workers have reported on conjugates of multi-armed poly(ethylene glycol) (PEG)–LMWH (a low-molecular-weight heparin-modified star polymer) and a dimeric, heparin-binding
A hetero-hexameric protein system is developed in this study, which not only functions as cross-linkers for hydrogel formation but also offers docking sites for controlled delivery of bioactive molecules. First, a hexameric protein with two, four, and six tax-interacting protein-1 (TIP-1), respectively (named as 2T, 4T, and 6T), is designed and obtained. As the hexapeptide ligand (WRESAI) can specifically bind to TIP-1 with high affinity, the hexameric proteins of 2T, 4T, and 6T can be used to crosslink the self-assembling nanofibers of NapGFFYGGGWRESAI, leading to formation of injectable biohybrid hydrogels with tunable mechanical properties. Furthermore, a hetero-hexameric protein containing four TIP-1 and two C-terminal moiety of the pneumococcal cellwall amidase LytA (C-LytA) proteins is designed and engineered (named as 4T2C). The 4T2C proteins can not only serve as cross-linkers for hydrogel formation but also provide docking sites for loading and controlled release of model drug Rhoda-GGK′. This study opens up new opportunities for further development of multifunctional hetero- recombinant protein-based hydrogels for biological applications.
1. Introduction We reported on the first example of hetero-hexameric protein system that not only functioned as cross-linkers for hydrogelations but also provided docking sites for controlled delivery of bioactive molecules. Hydrogels possess water-soluble 3D networks of polymers,[1] proteins,[2] or small molecules[3] formed by covalent or non-covalent bonds, and they are promising
X. L. Zhang, Dr. H. Zhou, Y. Xie, Dr. D. Ding, Prof. J. F. Long, Prof. Z. M. Yang State Key Laboratory of Medicinal Chemical Biology and College of Life Sciences Nankai University Tianjin 300071, P. R. China E-mail: [email protected]; [email protected]; [email protected] X. L. Zhang, C. H. Ren, Prof. Z. M. Yang Key Laboratory of Bioactive Materials Ministry of Education Collaborative Innovation Center of Chemical Science and Engineering (Tianjin) Nankai University Tianjin 300071, P. R. China
DOI: 10.1002/adhm.201300660
1804
wileyonlinelibrary.com
© 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Healthcare Mater. 2014, 3, 1804–1811
www.advhealthmat.de www.MaterialsViews.com
2. Results and Discussion 2.1. Preparation and Characterization of 2Tgel, 4Tgel, and 6Tgel Recently, we reported on the crystal structure of a hetero-hexameric structure formed by proteins of Elp-4, Elp-5, and Elp-6. Proteins of Elp-5 and Elp-6 were needed to be expressed in Escheria coli together and Elp-4 could be expressed by itself. Once mixing the three proteins at equal molar ratio, they would spontaneously form a very tight hetero-hexameric protein structure.[17] Based on this information, we rationally designed hexameric proteins with two, four, and six binding sites by engineering
Adv. Healthcare Mater. 2014, 3, 1804–1811
FULL PAPER
vascular endothelial growth factor (VEGF) for hydrogel formation. The resulting hydrogels can be applied for targeted delivery of VEGF.[16] Heilshorn and co-workers have rationally designed recombinant proteins with tandem repeating WW (single-letter amino acid symbol: W) and proline-rich peptide domains. The specific protein–protein interaction between WW domains and proline-rich peptides leads to the formation of physical hydrogels via a simple mixing process, which can be applied for cell encapsulation and cell delivery.[11] Ito et al. have also reported a trimeric CutA protein that can cross-link fourarmed PEG to form hydrogels for cell culture.[12] Stimulated by these successful examples of strategies used to form physical protein-based hydrogels, our group have designed a recombinant protein of ulbiquitin-like domain and tax-interacting protein-1 (ULD-TIP-1) that can form a tight tetramer structure. The tetrameric ULD-TIP-1 protein can then serve as a cross-linker to enhance the inter-fiber interaction between self-assembled nanofibers, leading to the formation of biohybrid hydrogels.[13] To the best of our knowledge, there are only two examples of using oligomeric recombinant proteins as cross-linkers to form hydrogels until now.[12,13] What is more, the two reported oligomeric recombinant proteins are both homo oligomers. Hetero-oligomeric recombinant proteins hold advantages because more than one kind of protein can be engineered to them. The resulting recombinant proteins can not only serve as crosslinkers for hydrogelations but also provide docking sites for bioactive molecules. Therefore, protein-based hydrogels formed by hetero-oligomeric proteins are multifunctional and can be applied for many applications such as controlled delivery of bioactive molecules. Here, we report for the first time on such an example of recombinant protein system that can form hetero-hexameric structures for hydrogelations. The hetero-hexameric protein (4T2C) with two kinds of different binding sites was engineered by mixing of proteins of C-LytA-Elp-4, TIP-1-Elp-5, and TIP1-Elp-6. Proteins of Elp-4, Elp-5, and Elp-6 could guarantee a hetero-hexameric structure formation. TIP-1 proteins on the hetero-hexameric protein allowed for crosslinking self-assembling nanofibers containing hexapeptide ligand WRESAI. On the other hand, the C-terminal moiety of the pneumococcal cell-wall amidase LytA (C-LytA) on the hetero-hexameric protein provided docking sites for model drug loading and controlled delivery. This study would open up new opportunities for further development of multifunctional hetero-recombinant protein-based hydrogels for biological applications.
Figure 1. Proposed structures of recombinant proteins based on Elp-4, Elp-5, and Elp-6 hetero-hexameric proteins. A) protein of TIP-1-Elp-4 yElp-5 yElp-6 (2T), B) protein of yElp-4 TIP-1-Elp-5 TIP-1-Elp-6 (4T), C) protein of TIP-1-Elp-4 TIP-1-Elp-5 TIP-1-Elp-6 (6T), and D) protein of C-lytA-Elp-4 TIP-1-Elp-5 TIP-1-Elp-6 (4T2C). The protein of yElp4, yElp5, yElp6, TIP-1, and C-LytA is in green, blue, red, yellow, and purple, respectively.
the hexameric protein with two, four, and six TIP-1 proteins, respectively (Figure 1). Since the TIP-1 protein could bind to the peptide of WRESAI very tightly with a dissociate constant (Kd) of 6.5 × 10−9 M,[18] the hexameric proteins with different numbers of binding sites could therefore be applied to crosslink self-assembling nanofibers to form biohybrid hydrogels (Figure 2A). The resulting hydrogels might possess different properties when different proteins were used. For example, the mechanical properties of the gels might be different upon proteins with various numbers of binding sites were used due to the different cross-linking densities in the gels. We first constructed two plasmids; one could be used to express TIP-1-Elp-4 and the other one to express TIP1-Elp-5 and TIP-1-Elp-6 simultaneously. Each recombinant protein could be expressed and obtained in moderate yields (>10 mg L−1) and in very good qualities (Figure S-1, Supporting Information). Mixing Elp-5, Elp-6, and TIP-1-Elp-4 at equal molar ratio could lead to the formation of a hexameric protein with two TIP-1 proteins (2T, Figure 1A). The analytical gel filtration result (Figure S-1A, Supporting Information) and ultracentrifugation result (Figure S-2A, Supporting Information) demonstrated the formation of a hexameric structure, indicating that the introduction of two TIP-1 proteins would not disturb the formation of the tight hexameric structure. Hexameric proteins with four and six TIP-1 proteins could also be obtained by the similar strategy (mixing Elp-4, TIP-1-Elp-5, and TIP-1-Elp-6 to produce 4T (Figure 1B) and mixing TIP-1-Elp-4, TIP-1-Elp-5, and TIP-1-Elp-6 to produce 6T (Figure 1C)). After obtaining these hetero-hexameric proteins with different numbers of binding sites, we tested their properties as crosslinkers
© 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
wileyonlinelibrary.com
1805
www.advhealthmat.de
FULL PAPER
www.MaterialsViews.com
Figure 2. A) A schematic illustration for using 6T proteins to cross-link self-assembling nanofibers of Nap-GFFYGGGWRESAI for hydrogel formation; B) Optical images of 1) solution of Nap-GFFYGGGWRESAI (0.5 wt%) in phosphate buffer saline (PBS, pH 7.4), 2) a gel formed by adding 0.25% of 2T to solution in 1), 3) a gel formed by adding 0.25% of 4T to solution in 1), and 4) a gel formed by adding 0.25% of 6T to solution in 1); C) Dynamic time sweep to indicate hydrogel formations by adding 2T, 4T, or 6T protein (0.25%) to solution of Nap-GFFYGGGWRESAI (0.5 wt%) (squares in black: 2T, up triangles in blue: 4T, down triangles in olive: 6T); D) Dynamic time sweep to indicate hydrogel formations by adding different concentrations of 6T protein to solution of Nap-GFFYGGGWRESAI (0.5 wt%) (squares in black: 0.05%, up triangles in blue: 0.10%, down triangles in olive: 0.25% and diamonds in magenta: 0.40% equiv. to Nap-GFFYGGGWRESAI, closed symbols: G’ and open symbols: G”).
for hydrogel formation. The Nap-GFFYGGGWRESAI (Scheme S-1, Supporting Information) peptide could self-assembled into nanofibers in aqueous solution but were not able to form a hydrogel because of the weak interactions between the selfassembled nanofibers. We subsequently investigated whether our hetero-hexameric proteins could serve as cross-linkers to enhance the inter-nanofiber interactions in order to promote hydrogelation. As shown in Figure 2B, the addition of 0.25 wt% (relative to Nap-GFFYGGGWRESAI) of 2T, 4T, or 6T protein to the phosphate buffer saline (PBS, pH 7.4) solution of NapGFFYGGGWRESAI led to the formation of biohybrid hydrogels (2Tgel, 4Tgel, and 6Tgel, respectively) within 5 min. The hydrogel formation was supported by the upside-down optical images (Figure 2B). The rheological measurements with the mode of dynamic time sweep also indicated the formation of biohybrid hydrogels upon mixing the protein and the peptide solutions (Figure 2C). Noteworthy is that the mechanical properties were different for the resulting 2Tgel, 4Tgel, and 6Tgel. As shown in Figure 2C, the order of elasticity (G′) value of the gel was followed the trend of 6Tgel > 4Tgel > 2Tgel. The results revealed that hydrogels formed by proteins with more binding sites (more crosslinking densities) would have better mechanical properties. The mechanical property of each kind of hydrogel could also be controlled by the amounts of the corresponding proteins as crosslinkers, and higher concentrations of proteins would lead to hydrogels with larger G′ values (Figure 2D and Figures S-10–12, Supporting Information). Moreover, upon fixation of the protein concentration, the G′ value of the gel increased with the increase of peptide 1806
wileyonlinelibrary.com
concentration (Figure S-13, Supporting Information). These observations revealed that the mechanical property of the protein-based hydrogels could be tuned by several factors, which included using hexameric proteins with different numbers of binding sites, using different concentrations of proteins, and using different concentrations of peptides. Hydrogels with adjustable mechanical properties would have great potential in cell culture, controllable cell differentiation, and controlled drug release. Moreover, the formed hydrogels had a fast recovery property. For instance, the gels containing various concentrations of NapGFFYGGGWRESAI (ranging from 0.3 to 1.0 wt%) and 0.1 wt% (relative to Nap-GFFYGGGWRESAI) of 6T showed shearthining properties. As shown in Figure 3, when subjected to a large external stress (50% stain), the gels experienced a gel– sol-phase transition. After removal of the external stress, both G′ and G′′ of the gels elevated rapidly and the gel mechanical strengths were significantly recovered within 600 s. These data suggested that the hydrogels formed by our method possessed thixotropic as well as rapid recovery properties and were thus injectable. We then used transmission electron microscopy (TEM) to investigate the nanostructures in solution of NapGFFYGGGWRESAI and the resulting gels. As shown in Figure S-15 (Supporting Information), we observed nanofibers with a diameter of around 35 nm in a PBS solution containing 0.5 wt% of Nap-GFFYGGGWRESAI. After adding 6T protein, these nearly parallel arranged-nanofibers (Figure S-15, Supporting Information) entangled with each other to form 3D
© 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Healthcare Mater. 2014, 3, 1804–1811
www.advhealthmat.de www.MaterialsViews.com
FULL PAPER Figure 3. The recovery property of the gels. Rheological measurements in dynamic time sweep mode for hydrogel containing 0.10% of 6T equiv. to Nap-GFFYGGGWRESAI with A) 0.3 wt% (squares in black), B) 0.5 wt% (up triangles in blue), C) 0.8 wt% (down triangles in olive), and D) 1.0 wt% (diamonds in magenta) of Nap-GFFYGGGWRESAI in PBS buffer solution. (closed symbols: G′ and open symbols: G′′).
networks, leading to the hydrogel formation. We observed the 3D networks of nanofibers in all hydrogels with different concentrations of 6T (Figure 4). However, the diameter and
morphology of the nanofibers in various gels were different. As shown in Figure 4, the diameter of the nanofibers in gels was 38.5 ± 3.9, 48.0 ± 3.0, 51.7 ± 5.1, and 60.3 ± 4.6 nm when the concentration of 6T was 0.05, 0.10, 0.25, and 0.40 wt% to NapGFFYGGGWRESAI, respectively. The difference in nanofiber diameter between each two gel groups is statistically significant (P < 0.05; 50 nanofibers were measured in each gel group by TEM observation). Similar trends were also observed in other gels with 2T or 4T protein (Figure S-16 and S-17, Supporting Information). These observations in TEM images correlated well with the mechanical properties of the gels and indicated that gels with bigger G′ value would possess larger-sized nanofibers. 2.2. Preparation and Characterization of 4T2Cgel
Figure 4. TEM images of gels containing 0.5 wt% of NapGFFYGGGWRESAI and different concentrations of 6T: A) 0.05%, B) 0.10%, C) 0.25%, and D) 0.40% equiv. relative to Nap-GFFYGGGWRESAI.
Adv. Healthcare Mater. 2014, 3, 1804–1811
Since the hexameric protein was formed by Elp-4, Elp-5, and Elp-6 that are capable of being expressed separately, two kinds of proteins with different binding sites could be rationally engineered to the hexameric protein together. The resulting hexameric protein could therefore not only serve as the nanofiber cross-linkers but also provide docking sites for bioactive molecules. In order to achieve this goal, we engineered a plasmid that could express a recombinant protein of C-LytA-Elp-4. The Lytic amidase was anchored at bacterial surface through the interaction between its C-terminal (C-LytA) and the bacterial surface choline.[19] The mixing of C-LytA-Elp-4, TIP-1-Elp-5, and TIP-1-Elp-6 might lead to the formation of a hexameric protein (protein 4T2C in Figure 1D) with two kinds of binding sites;
© 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
wileyonlinelibrary.com
1807
www.advhealthmat.de
FULL PAPER
www.MaterialsViews.com
Figure 5. A) Analytical gel filtration profile and its SDS-PAGE gel of protein 4T2C, B) Rheological measurements in dynamic time sweep mode for gels containing 0.5 wt% of Nap-GFFYGGGWRESAI at the frequency of 1 rad s−1 and strain of 0.5% with different concentrations of 4T2C in PBS buffer solution (down triangles in olive: 0.25% and diamonds in magenta: 0.40% equiv. to Nap-GFFYGGGWRESAI, closed symbols: G′ and open symbols: G′′), C) A schematic illustration for using Rhoda-GGK′-bound 4T2C protein to cross-link self-assembling nanofibers of Nap-GFFYGGGWRESAI to form a hydrogel, and D) drug release profile of Rhoda-peptides from hydrogels (higher binding affinity of peptides to TIP-1 protein results in lower release speed, and vice versa).
one from TIP-1 for nanofiber crosslinking and the other one from C-LytA for bioactive molecule loading. The protein 4T2C could indeed be obtained in a moderate yield (about 5 mg L−1) and good quality (Figure S-1D, Supporting Information). The results by analytical gel filtration (Figure 5A) and ultracentrifugation (Figure S-2D, Supporting Information) indicated the formation of a hetero-hexameric protein structure. The property of the obtained hexameric protein 4T2C for hydrogel formation was subsequently tested. Adding 0.25 wt% (relative to Nap-GFFYGGGWRESAI) of 4T2C protein to the PBS solution of Nap-GFFYGGGWRESAI (0.5 wt%) led to the formation of biohybrid hydrogel (4T2Cgel) within 5 min. The 4T2Cgel was then characterized by rheological and TEM measurements. As shown in Figure 5B, the G′ value of the resulting gels significantly increased with the increase of amounts of 4T2C protein for crosslinking. This result revealed that the hydrogels formed by more 4T2C proteins would have better mechanical properties and that the mechanical property of the hydrogels could be adjusted by two factors including using different concentrations of 4T2C proteins and using different concentrations of peptides. Figure S-18 (Supporting Information) displayed the TEM images of gels containing 0.5 wt% of NapGFFYGGGWRESAI and different amounts of 4T2C proteins. 3D networks of nanofibers were observed in all of the hydrogels with various 4T2C amounts. Moreover, when higher concentration of 4T2C protein was used, the diameter of nanofibers in the resulting 4T2Cgel was larger in the TEM observations,
1808
wileyonlinelibrary.com
which agreed well with the corresponding mechanical property trend of the gels. The rheological and TEM observations revealed that 4T2C proteins are efficient cross-linkers to enhance the internanofiber interactions, thus strengthening the mechanical property of 4T2Cgel. As our hydrogels had the multi-condition controlled property, we predicted that it would be beneficial to their future applications in the fields such as stem cell controlled differentiation and controlled drug release. 2.3. Application of 4T2Cgel in Controlled Delivery of Model Drug In order to demonstrate the potential application of 4T2Cgel in controlled drug release, we synthesized a dye-peptide conjugate (Rhoda-GGK′; Scheme S-2, Supporting Information; Figures S-6 and S-7, Supporting Information for its characterization) as a model drug. It is noted that Rhoda-GGK′ was tested to possess a Kd value to C-LytA binding of 4.5 ± 1.1 × 10−6 M measured by isothermal titration calorimetry (ITC) (Figure S-3, Supporting Information). Besides C-LytA that could bind to Rhoda-GGK′, the four TIP-1 in the 4T2C allowed the heteroproteins to crosslink nanofibers of Nap-GFFYGGGWRESAI, afforded Rhoda-GGK′-loaded 4T2Cgels (Figure 5C). The in vitro releasing profile of Rhoda-GGK′ from the 2C4Tgels was then studied. After the formation of Rhoda-GGK′-loaded 4T2Cgels with concentrations of the 4T2C protein of 0.10, 0.20, 0.25,
© 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Healthcare Mater. 2014, 3, 1804–1811
www.advhealthmat.de www.MaterialsViews.com
3. Conclusions We have designed and obtained 2T, 4T, and 6T proteins with two, four, and six binding sites, respectively, which could be utilized as crosslinkers to form injectable biohybrid hydrogels with controllable mechanical properties. More importantly, we have also engineered a hetero-hexameric protein 4T2C with two kinds of different binding sites. The 4T2C proteins have been demonstrated to not only serve as cross-linkers for hydrogel formation but also provide docking sites for model drug RhodaGGK′ loading. The resulting 4T2Cgels could therefore be a very promising system for controlled delivery of bioactive molecules. This successful example of hetero-oligomeric proteinbased hydrogel system will inspire more exciting research in this emerging field.
4. Experimental Section Chemicals: Fmoc-amino acids were obtained from GL Biochem (Shanghai, China). Rhodamine B was received from J&K. All the other starting materials were obtained from Alfa. Chemical reagents and solvents were used as received from commercial sources. General Methods: 1H NMR spectra were obtained on Bruker ARX 400 using DMSO-d6 as the solvent; HR-MS were received from VG ZAB-HS system (England). HPLC was conducted at LUMTECH HPLC (Germany) system using a C18 RP column with MeOH (0.05% of TFA) and water (0.05% of TFA) as the eluents; TEM images were done on a Tecnai G2 F20 system, operating at 200 kV, TEM samples were prepared as following: a carbon-coated copper grid (from Zhongjingkeyi Technology Co. Ltd., Beijing, P. R. China) was vertically dipped into the hydrogel for 5 s, washed by water two times, and then placed in a dessicator overnight before the TEM measurement. LC-MS was conducted at the LCMS-20AD (Shimadzu) system, and rheology was performed on an AR 1500ex (TA instrument) system using a parallel plate (40 mm) at the gap of 300 µm. Protein Expression and Purification: Expression and purification of proteins were performed by standard recombinant protein technology as previously described.[17] Briefly, DNA fragments corresponding to the ChBD of the major lytic amidase (C-lytA) of S. pneumoniae and human TIP-1 were amplified by polymerase chain reaction (PCR). Gene fragments corresponding to yeast Elp4 (residues 67–372 from full-length
Adv. Healthcare Mater. 2014, 3, 1804–1811
residues 1–456) (yElp4), Elp5 (residues 1–238 from full-length residues 1–309) (yElp5), and full-length Elp6 (residues 1–273) (yElp6) were amplified by PCR from the S. cerevisiae genome. To make four singlechain fusion proteins of the TIP1-Elp4, TIP1-Elp5, TIP1-Elp6, C-lytA-Elp4, the DNA fragment of TIP-Elp4 and C-lytA-Elp4 were cloned into the modified pET32a vector (Novagen). The S-tag and thrombin recognition sites were replaced with a sequence encoding a 3C protease-cleavable segment (Leu-Glu-Val-Leu-Phe-Gln-Gly-Pro). The DNA fragments of TIP1-Elp5 and TIP1-Elp6 were cloned into the pETDuet-1 (Novagen) vector to co-express these two proteins. All resulting proteins contained Trx-His6 tags at their N termini. BL21(DE3) CodonPlus E. coli cells harboring the expression plasmid were grown in LB medium at 37 °C until the OD600 reached 0.6 and then induced with 0.3 × 10−3 M isopropyl-β-D-thiogalactoside at 16 °C for about 16–18 h. After being spun at 5000 rpm for 15 min, E. coli cells were resuspended in T50N500I5 buffer (50 × 10−3 M Tris–HCl pH 7.9, 500 × 10−3 M NaCl, and 5 × 10−3 M imidazole) supplemented with 1 × 10−3 M phenylmethylsulfonyl fluoride, 1 µg mL−1 leupeptin and 1 µg mL−1 antipain. The cells were then lysed by sonication. After the lysates had been centrifuged at 18 000 rpm for 30 min, the supernatant was loaded onto a Ni-NTA agarose column (Qiagen) that was equilibrated with T50N500I5 buffer. The Ni-NTA column was washed with 5 column volumes of T50N500I5 buffer. The Trx-his6-tagged protein was eluted with T50N500I5 buffer containing 500 × 10−3 M imidazole. The eluted proteins loaded on a HighLoad 26/60 Superdex-200 size-exclusion column (GE Healthcare) and eluted with T50N300D1 buffer (50 × 10−3 M Tris–HCl pH 7.4, 300 × 10−3 M NaCl, and 1 × 10−3 M DTT) at a flow rate of 2.5 mL min−1. Each fraction of the column elute was 5 mL. The protein peak was identified by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gel. After digestion with PreScission Protease to cleave the N-terminal Trx-His6-tag, the target protein was purified on a HighLoad 26/60 Superdex-200 sizeexclusion column in 50 × 10−3 M Tris–HCl pH 7.4, 300 × 10−3 M NaCl, and 1 × 10−3 M DTT. However, it is worth mentioning that TIP1-Elp5 and TIP1-Elp6 plasmid should transform into Pubs520 E. Coli cells. The purification method remains the same. Analytical Gel Filtration: Size-exclusion chromatography was performed on an AKTA purifier system using a Superdex 200 10/300 column (GE Healthcare) for TIP1-Elp4 yElp5 yElp6, yElp4 TIP1-Elp5 TIP1-Elp6, TIP1 Elp4 TIP1-Elp5 TIP1-Elp6, C-lytA-Elp4 TIP1-Elp5 TIP1-Elp6. Protein samples were dissolved in buffer containing 50 × 10−3 M Tris–HCl pH 7.4, 300 × 10−3 M NaCl, and 1 × 10−3 M DTT. The column was calibrated with a gel filtration standard from Bio-Rad. Analytical Ultracentrifugation: Sedimentation velocity (SV) was performed in a Beckman/Coulter XL-I analytical ultracentrifuge using double-sector. An additional protein purification step on a Superose 12 10/300 column (GE Healthcare) in 50 × 10−3 M Tris–HCl pH 7.4, 300 × 10−3 M NaCl, and 1 × 10−3 M DTT was performed before the experiments. The experiments were conducted at 30 000 rpm, and 4 °C using interference detection and double-sector cells loaded with 1 mg mL−1 of the protein. The buffer composition (density and viscosity) and protein partial specific volume (V-bar) were obtained using the program SEDNTERP.[20] The data were analyzed using the programs SEDFIT and SEDPHAT. Isothermal Titration Calorimetry: ITC measurements were carried out on a MicroCal Isothermal Titration Calorimeter iTC200 (GE Healthcare) in PBS buffer pH 7.4 at 16 °C. The peptide was dissolved in the same buffer mentioned above and adjusted to pH 7.4. Both peptide and protein solutions were degassed by being spun at 13 000 rpm for 15 min. To measure the binding constant of C-lytA with the peptide (Rhoda-GGK′), an initial injection (0.4 µL) followed by 19 injections (2 µL) peptide (0.750 × 10−3 M) into the calorimeter cell, which was completely filled with protein solution (0.015 × 10−3 M), were collected at 120 s intervals while being stirred at 1000 rpm. The titration data and binding plot were analyzed using MicroCal Origin software with one-site binding model. Peptide Systhesis: The peptide derivative was prepared by solidphase peptide synthesis (SPPS) using 2-chlorotrityl chloride resin and the corresponding N-Fmoc protected amino acids with side chains
© 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
wileyonlinelibrary.com
FULL PAPER
and 0.30 wt% equiv. to Nap-GFFYGGGWRESAI, respectively (Rhoda-GGK′ was fixed at 0.1 wt% for all 4T2Cgels), 0.2 mL of PBS buffer was added on the top of each Rhoda-GGK′-loaded 4T2Cgel. At designated time intervals, the upper PBS buffer was totally taken out for measurement of accumulating release amount of Rhoda-GGK′ from each 4T2Cgel and 0.2 mL of fresh PBS buffer was subsequently added. Figure 5D showed the release profiles of Rhoda-GGK′ from the Rhoda-GGK′-loaded 4T2Cgels. It is obvious that Rhoda-GGK′ could be released from each 4T2Cgel in a sustained manner. Additionally, the amounts of 4T2C proteins in the 4T2Cgels could significantly influence the release profile of Rhoda-GGK′ and higher 4T2C protein concentrations led to slower release rates of RhodaGGK′, which would be mainly due to the more binding sites for Rhoda-GGK′ provided by the proteins. The results demonstrated that our hetero-hexameric protein 4T2C can not only serve as crosslinkers for hydrogel formation, but also endow the resulting protein-based hydrogel with plenty of docking sites for loading and controlled release of model drug Rhoda-GGK′.
1809
www.advhealthmat.de
FULL PAPER
www.MaterialsViews.com
1810
properly protected. The first amino acid was loaded on the resin at the C-terminal with the loading efficiency about 0.5 mmol g−1. 20% piperidine in anhydrous N,N′-dimethylformamide (DMF) was used during deprotection of Fmoc group. Then the next Fmoc-protected amino acid was coupled to the free amino group using O-(Benzotriazol1-yl)-N,N,N′,N′-tetramethyluroniumhexafluorophosphate (HBTU) as the coupling reagent. The growth of the peptide chain was according to the established Fmoc SPPS protocol. At the final step, Naphthalene acetic acid or Rhodamine B was used to attach on the peptide. After the last coupling step, excessive reagents were removed by a single DMF wash for 5 min (5 mL per gram of resin), followed by five steps of washing using DCM for 1 min (5 mL g−1 of resin). The peptide derivative was cleaved using 95% of trifluoroacetic acid with 2.5% of TMS and 2.5% of H2O for 45 min. 20 mL per gram of resin of ice-cold diethylether was then added to cleavage reagent. The resulting precipitate was centrifuged for 10 min at 4 °C at 10 000 rpm. Afterward the supernatant was decanted and the resulting solid was dissolved in DMSO for HPLC separation. Nap-GFFYGGGWRESAI: (This compound was reported in our previous study.)[13]1H NMR (300MHz, DMSO-d6) δ 10.78 (s, 1H), 9.20 (s, 1H), 8.25–8.31 (m, 2H), 8.13–8.20 (m, 3H), 8.02–8.11 (m, 5H), 7.80–7.89 (m, 3H), 7.74 (S, 1H), 7.59–7.64 (d, J = 7.61Hz, 1H), 7.45– 7.56 (m, 2H), 7.39–7.44 (t, 1H)7.28–7.34 (d, J = 7.31Hz, 2H), 7.11–7.23 (m, 12H), 7.01–7.07 (t, 3H),6.93–6.99 (t, 1H), 6.61–6.68 (d, J = 6.64Hz, 2H), 4.53–4.63 (m, 2H), 4.43–4.53 (m, 3H), 4.28–4.39 (m, 4H), 4.09– 4.15 (m,1H), 3.68–3.80 (m, 6H), 3.53–3.64 (m, 6H), 3.46–3.51 (m, 1H), 3.15–3.19 (d, J = 3.17Hz, 2H), 3.06–3.14 (m, 3H), 2.88–3.00 (m, 4H), 2.61–2.79 (m, 4H), 2.26–2.34 (m, 2H), 1.89–1.96 (m, 1H), 1.69–1.80 (m, 3H), 1.33–1.58 (m, 5H), 1.14–1.28 (m, 6H), 0.79–0.89 (t, 6H). MS: calc. M+ = 1614.7, obsvd. (M+1)+ = 1615.7. Rhoda-GGK′: 1H NMR (400MHz, DMSO-d6) δ 8.07–8.12 (d, 1H), 7.84–7.90 (m, 2H), 7.54–7.60 (m, 2H), 7.42 (s, 1H), 7.29 (s, 1H), 7.16 (s, 1H), 7.07–7.13 (m, 1H), 4.16–4.25 (m, 2H), 4.37–4.46 (m, 9H), 3.23– 3.30 (m, 2H), 3.04 (s, 9H), 1.72–1.81 (m, 1H), 1.60–1.71 (m, 3H), 1.22– 1.33 (m, 3H), 1.01–1.12 (t, 12H). MS: calc. M+ = 728.4, obsvd. (M+1)+ = 729.4. Preparation of Solution ofNap-GFFYGGGWRESAI: 5.0 mg of NapGFFYGGGWRESAI (3.10 µmol) was dissolved in 0.50 mL of PBS buffer solution containing 0.66 mg (2 equiv. to Nap-GFFYGGGWRESAI) of Na2CO3 (2 equiv. of Na2CO3 were used to neutralize the carboxylic acids on Nap-GFFYGGGWRESAI to make the final pH value to 7.4). The mixture was heated to form a transparent solution, which was followed by cooling back to room temperature for use. Preparation of Solution of Rhoda-GGK′: 5.0 mg of Rhoda-GGK′ (6.86 µmol) was dissolved in 1.25 mL of PBS buffer solution containing 1.46 mg (2 equiv. to Rhoda-GGK’) of Na2CO3 (2 equiv. of Na2CO3 were used to neutralize the carboxylic acids on Rhoda-GGK′ to make the final pH value to 7.4). The mixture was heated to form a transparent solution and then cooling back to room temperature for use. Preparation of Hydrogels: Take 0.10 mL the solution of NapGFFYGGGWRESAI into a small bottle and then the resulting solution was heated to form a transparent solution, which was followed by cooling back to room temperature, affording a nanofiber solution of Nap-GFFYGGGWRESAI. 0.10 mL of PBS buffer solution containing 65.81 µg, 131.62 µg, 329.05 µg, or 526.48 µg of 2T protein (0.31 nmol, 0.62 nmol, 1.55 nmol, or 2.48 nmol, 0.05%, 0.10%, 0.25%, or 0.40% equiv. to Nap-GFFYGGGWRESAI) was added to the nanofiber solution. Gels would form after the mixture was rapidly stirred within seconds and then kept at room temperature (22–25 °C) for about 3–30 min. Other hexamer proteins of 4T, 6T, or 4T2C with different concentrations were also used to form gels by using similar procedures. Rheology: Rheology test was done on an AR 1500ex (TA instrument) system, 40 mm parallel plates were used during the experiment at the gap of 300 µm. The dynamic time sweep was conducted at the frequency of 1 rad s−1 and the strain of 0.5%. Dynamic strain sweep was performed and the strain values within the linear range were chosen for the following dynamic frequency sweep. The gels were also characterized by the mode of dynamic frequency sweep in the region of 0.1–100 rad s−1
wileyonlinelibrary.com
at the strain of 0.5%. The recovery property of the gels was probed as following: the large amplitude of stain (50%) at the frequency of 0.5 rad s−1 was first applied to the gels for 600 s. And then the recovery property of the gels was measured right after the removal of the external large amplitude of force (50% of strain) at the strain of 0.3% and the frequency of 0.5 rad s−1 for 3600 s. Release Profile: 0.2 mL of 4T2Cgel containing Rhodamine–peptide conjugates was formed as described above (the final concentrations of the protein respectively were 0.10%, 0.20%, 0.25%, and 0.30% equiv. to Nap-GFFYGGGWRESAI, Rhodamine–peptide was 0.1 wt%). After 10 h, each of the gels was treated with 0.2 mL of fresh PBS buffer solutions (pH 7.4). 0.2 mL of the upper buffer solution was taken out and used to tested by UV–vis Spectrophotometer at the wavelength of 560 nm at designated time intervals. A fresh 0.2 mL of PBS was added back to the gel. Standard curve of the model drug was determined before the test. The experiment was conducted in three parallel experiments. In addition, 0.2 mL of 4T2Cgel containing Rhodamine–peptide conjugates was formed in syringes, then injected in the small bottle and formed hydrogels (the final concentrations of the protein respectively were 0.20%, 0.25%, and 0.30% equiv. to Nap-GFFYGGGWRESAI, Rhodamine– peptide was 0.1 wt%). Then measured them by using similar procedures as above. Statistical Analysis: Quantitative data were expressed as mean ± standard deviation. Statistical comparisons were made by ANOVA analysis and Student's t-test. A P value
FULL PAPER
www.MaterialsViews.com
Rational Design of Multifunctional Hetero-Hexameric Proteins for Hydrogel Formation and Controlled Delivery of Bioactive Molecules Xiaoli Zhang, Hao Zhou, Ying Xie, Chunhua Ren, Dan Ding,* Jiafu Long,* and Zhimou Yang* biomaterials because they mimick extracellular matrix of tissues and organs. The hydrogels containing proteins in the 3D networks are known as protein-based hydrogels. Due to the unique property of proteins (e.g., perfect polydispersity, bioactivity, exact control over monomer sequence, and ability to fine-tune molecular-level biochemical interactions), protein-based hydrogels have attracted extensive research interests in recent two decades,[4] and they have exhibited great potential for tissue engineering,[5] bioactive molecule sensing,[6] and regenerative medicine.[7] Protein-based hydrogels are usually formed by either thermo-inducement or mixing two components that are capable of forming covalent bonds or having specific non-covalent interactions. For example, coiled-coil peptides or biohybrids consisted of coiled-coil peptides can form oligomeric structures at elevated temperatures that serve as cross-linkers for hydrogel formation.[2a,8] The covalent strategy for protein-based hydrogel formation often utilizes recombinant proteins to crosslink polymers by Michael addition reactions, site selective conjugation, and enzyme-catalyzed ligations.[9] Furthermore, specific non-covalent interactions including protein–ligand,[7,10] protein–protein,[11] protein–peptide,[12,13] and protein–polysacharide interactions[14] have also been applied to form protein-based hydrogels. Noteworthy is that these hydrogels formed by specific non-covalent interactions are physical hydrogels that possess excellent injectable property.[4d] Besides, the simple mixing procedure for preparing physical hydrogels also renders its promising application in such as cell delivery[11,12,15] and controlled drug delivery.[6a,7] In order to obtain protein-based hydrogels by non-covalent interactions, the design and production of proteins with multiple binding sites are the key. Recombinant proteins with multiple binding sites can be prepared through conjugation with multi-armed polymers, by rational design to produce those with tandem repeating binding domains, and by generation of those that can form oligomer structures. For example, Kiick and co-workers have reported on conjugates of multi-armed poly(ethylene glycol) (PEG)–LMWH (a low-molecular-weight heparin-modified star polymer) and a dimeric, heparin-binding
A hetero-hexameric protein system is developed in this study, which not only functions as cross-linkers for hydrogel formation but also offers docking sites for controlled delivery of bioactive molecules. First, a hexameric protein with two, four, and six tax-interacting protein-1 (TIP-1), respectively (named as 2T, 4T, and 6T), is designed and obtained. As the hexapeptide ligand (WRESAI) can specifically bind to TIP-1 with high affinity, the hexameric proteins of 2T, 4T, and 6T can be used to crosslink the self-assembling nanofibers of NapGFFYGGGWRESAI, leading to formation of injectable biohybrid hydrogels with tunable mechanical properties. Furthermore, a hetero-hexameric protein containing four TIP-1 and two C-terminal moiety of the pneumococcal cellwall amidase LytA (C-LytA) proteins is designed and engineered (named as 4T2C). The 4T2C proteins can not only serve as cross-linkers for hydrogel formation but also provide docking sites for loading and controlled release of model drug Rhoda-GGK′. This study opens up new opportunities for further development of multifunctional hetero- recombinant protein-based hydrogels for biological applications.
1. Introduction We reported on the first example of hetero-hexameric protein system that not only functioned as cross-linkers for hydrogelations but also provided docking sites for controlled delivery of bioactive molecules. Hydrogels possess water-soluble 3D networks of polymers,[1] proteins,[2] or small molecules[3] formed by covalent or non-covalent bonds, and they are promising
X. L. Zhang, Dr. H. Zhou, Y. Xie, Dr. D. Ding, Prof. J. F. Long, Prof. Z. M. Yang State Key Laboratory of Medicinal Chemical Biology and College of Life Sciences Nankai University Tianjin 300071, P. R. China E-mail: [email protected]; [email protected]; [email protected] X. L. Zhang, C. H. Ren, Prof. Z. M. Yang Key Laboratory of Bioactive Materials Ministry of Education Collaborative Innovation Center of Chemical Science and Engineering (Tianjin) Nankai University Tianjin 300071, P. R. China
DOI: 10.1002/adhm.201300660
1804
wileyonlinelibrary.com
© 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Healthcare Mater. 2014, 3, 1804–1811
www.advhealthmat.de www.MaterialsViews.com
2. Results and Discussion 2.1. Preparation and Characterization of 2Tgel, 4Tgel, and 6Tgel Recently, we reported on the crystal structure of a hetero-hexameric structure formed by proteins of Elp-4, Elp-5, and Elp-6. Proteins of Elp-5 and Elp-6 were needed to be expressed in Escheria coli together and Elp-4 could be expressed by itself. Once mixing the three proteins at equal molar ratio, they would spontaneously form a very tight hetero-hexameric protein structure.[17] Based on this information, we rationally designed hexameric proteins with two, four, and six binding sites by engineering
Adv. Healthcare Mater. 2014, 3, 1804–1811
FULL PAPER
vascular endothelial growth factor (VEGF) for hydrogel formation. The resulting hydrogels can be applied for targeted delivery of VEGF.[16] Heilshorn and co-workers have rationally designed recombinant proteins with tandem repeating WW (single-letter amino acid symbol: W) and proline-rich peptide domains. The specific protein–protein interaction between WW domains and proline-rich peptides leads to the formation of physical hydrogels via a simple mixing process, which can be applied for cell encapsulation and cell delivery.[11] Ito et al. have also reported a trimeric CutA protein that can cross-link fourarmed PEG to form hydrogels for cell culture.[12] Stimulated by these successful examples of strategies used to form physical protein-based hydrogels, our group have designed a recombinant protein of ulbiquitin-like domain and tax-interacting protein-1 (ULD-TIP-1) that can form a tight tetramer structure. The tetrameric ULD-TIP-1 protein can then serve as a cross-linker to enhance the inter-fiber interaction between self-assembled nanofibers, leading to the formation of biohybrid hydrogels.[13] To the best of our knowledge, there are only two examples of using oligomeric recombinant proteins as cross-linkers to form hydrogels until now.[12,13] What is more, the two reported oligomeric recombinant proteins are both homo oligomers. Hetero-oligomeric recombinant proteins hold advantages because more than one kind of protein can be engineered to them. The resulting recombinant proteins can not only serve as crosslinkers for hydrogelations but also provide docking sites for bioactive molecules. Therefore, protein-based hydrogels formed by hetero-oligomeric proteins are multifunctional and can be applied for many applications such as controlled delivery of bioactive molecules. Here, we report for the first time on such an example of recombinant protein system that can form hetero-hexameric structures for hydrogelations. The hetero-hexameric protein (4T2C) with two kinds of different binding sites was engineered by mixing of proteins of C-LytA-Elp-4, TIP-1-Elp-5, and TIP1-Elp-6. Proteins of Elp-4, Elp-5, and Elp-6 could guarantee a hetero-hexameric structure formation. TIP-1 proteins on the hetero-hexameric protein allowed for crosslinking self-assembling nanofibers containing hexapeptide ligand WRESAI. On the other hand, the C-terminal moiety of the pneumococcal cell-wall amidase LytA (C-LytA) on the hetero-hexameric protein provided docking sites for model drug loading and controlled delivery. This study would open up new opportunities for further development of multifunctional hetero-recombinant protein-based hydrogels for biological applications.
Figure 1. Proposed structures of recombinant proteins based on Elp-4, Elp-5, and Elp-6 hetero-hexameric proteins. A) protein of TIP-1-Elp-4 yElp-5 yElp-6 (2T), B) protein of yElp-4 TIP-1-Elp-5 TIP-1-Elp-6 (4T), C) protein of TIP-1-Elp-4 TIP-1-Elp-5 TIP-1-Elp-6 (6T), and D) protein of C-lytA-Elp-4 TIP-1-Elp-5 TIP-1-Elp-6 (4T2C). The protein of yElp4, yElp5, yElp6, TIP-1, and C-LytA is in green, blue, red, yellow, and purple, respectively.
the hexameric protein with two, four, and six TIP-1 proteins, respectively (Figure 1). Since the TIP-1 protein could bind to the peptide of WRESAI very tightly with a dissociate constant (Kd) of 6.5 × 10−9 M,[18] the hexameric proteins with different numbers of binding sites could therefore be applied to crosslink self-assembling nanofibers to form biohybrid hydrogels (Figure 2A). The resulting hydrogels might possess different properties when different proteins were used. For example, the mechanical properties of the gels might be different upon proteins with various numbers of binding sites were used due to the different cross-linking densities in the gels. We first constructed two plasmids; one could be used to express TIP-1-Elp-4 and the other one to express TIP1-Elp-5 and TIP-1-Elp-6 simultaneously. Each recombinant protein could be expressed and obtained in moderate yields (>10 mg L−1) and in very good qualities (Figure S-1, Supporting Information). Mixing Elp-5, Elp-6, and TIP-1-Elp-4 at equal molar ratio could lead to the formation of a hexameric protein with two TIP-1 proteins (2T, Figure 1A). The analytical gel filtration result (Figure S-1A, Supporting Information) and ultracentrifugation result (Figure S-2A, Supporting Information) demonstrated the formation of a hexameric structure, indicating that the introduction of two TIP-1 proteins would not disturb the formation of the tight hexameric structure. Hexameric proteins with four and six TIP-1 proteins could also be obtained by the similar strategy (mixing Elp-4, TIP-1-Elp-5, and TIP-1-Elp-6 to produce 4T (Figure 1B) and mixing TIP-1-Elp-4, TIP-1-Elp-5, and TIP-1-Elp-6 to produce 6T (Figure 1C)). After obtaining these hetero-hexameric proteins with different numbers of binding sites, we tested their properties as crosslinkers
© 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
wileyonlinelibrary.com
1805
www.advhealthmat.de
FULL PAPER
www.MaterialsViews.com
Figure 2. A) A schematic illustration for using 6T proteins to cross-link self-assembling nanofibers of Nap-GFFYGGGWRESAI for hydrogel formation; B) Optical images of 1) solution of Nap-GFFYGGGWRESAI (0.5 wt%) in phosphate buffer saline (PBS, pH 7.4), 2) a gel formed by adding 0.25% of 2T to solution in 1), 3) a gel formed by adding 0.25% of 4T to solution in 1), and 4) a gel formed by adding 0.25% of 6T to solution in 1); C) Dynamic time sweep to indicate hydrogel formations by adding 2T, 4T, or 6T protein (0.25%) to solution of Nap-GFFYGGGWRESAI (0.5 wt%) (squares in black: 2T, up triangles in blue: 4T, down triangles in olive: 6T); D) Dynamic time sweep to indicate hydrogel formations by adding different concentrations of 6T protein to solution of Nap-GFFYGGGWRESAI (0.5 wt%) (squares in black: 0.05%, up triangles in blue: 0.10%, down triangles in olive: 0.25% and diamonds in magenta: 0.40% equiv. to Nap-GFFYGGGWRESAI, closed symbols: G’ and open symbols: G”).
for hydrogel formation. The Nap-GFFYGGGWRESAI (Scheme S-1, Supporting Information) peptide could self-assembled into nanofibers in aqueous solution but were not able to form a hydrogel because of the weak interactions between the selfassembled nanofibers. We subsequently investigated whether our hetero-hexameric proteins could serve as cross-linkers to enhance the inter-nanofiber interactions in order to promote hydrogelation. As shown in Figure 2B, the addition of 0.25 wt% (relative to Nap-GFFYGGGWRESAI) of 2T, 4T, or 6T protein to the phosphate buffer saline (PBS, pH 7.4) solution of NapGFFYGGGWRESAI led to the formation of biohybrid hydrogels (2Tgel, 4Tgel, and 6Tgel, respectively) within 5 min. The hydrogel formation was supported by the upside-down optical images (Figure 2B). The rheological measurements with the mode of dynamic time sweep also indicated the formation of biohybrid hydrogels upon mixing the protein and the peptide solutions (Figure 2C). Noteworthy is that the mechanical properties were different for the resulting 2Tgel, 4Tgel, and 6Tgel. As shown in Figure 2C, the order of elasticity (G′) value of the gel was followed the trend of 6Tgel > 4Tgel > 2Tgel. The results revealed that hydrogels formed by proteins with more binding sites (more crosslinking densities) would have better mechanical properties. The mechanical property of each kind of hydrogel could also be controlled by the amounts of the corresponding proteins as crosslinkers, and higher concentrations of proteins would lead to hydrogels with larger G′ values (Figure 2D and Figures S-10–12, Supporting Information). Moreover, upon fixation of the protein concentration, the G′ value of the gel increased with the increase of peptide 1806
wileyonlinelibrary.com
concentration (Figure S-13, Supporting Information). These observations revealed that the mechanical property of the protein-based hydrogels could be tuned by several factors, which included using hexameric proteins with different numbers of binding sites, using different concentrations of proteins, and using different concentrations of peptides. Hydrogels with adjustable mechanical properties would have great potential in cell culture, controllable cell differentiation, and controlled drug release. Moreover, the formed hydrogels had a fast recovery property. For instance, the gels containing various concentrations of NapGFFYGGGWRESAI (ranging from 0.3 to 1.0 wt%) and 0.1 wt% (relative to Nap-GFFYGGGWRESAI) of 6T showed shearthining properties. As shown in Figure 3, when subjected to a large external stress (50% stain), the gels experienced a gel– sol-phase transition. After removal of the external stress, both G′ and G′′ of the gels elevated rapidly and the gel mechanical strengths were significantly recovered within 600 s. These data suggested that the hydrogels formed by our method possessed thixotropic as well as rapid recovery properties and were thus injectable. We then used transmission electron microscopy (TEM) to investigate the nanostructures in solution of NapGFFYGGGWRESAI and the resulting gels. As shown in Figure S-15 (Supporting Information), we observed nanofibers with a diameter of around 35 nm in a PBS solution containing 0.5 wt% of Nap-GFFYGGGWRESAI. After adding 6T protein, these nearly parallel arranged-nanofibers (Figure S-15, Supporting Information) entangled with each other to form 3D
© 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Healthcare Mater. 2014, 3, 1804–1811
www.advhealthmat.de www.MaterialsViews.com
FULL PAPER Figure 3. The recovery property of the gels. Rheological measurements in dynamic time sweep mode for hydrogel containing 0.10% of 6T equiv. to Nap-GFFYGGGWRESAI with A) 0.3 wt% (squares in black), B) 0.5 wt% (up triangles in blue), C) 0.8 wt% (down triangles in olive), and D) 1.0 wt% (diamonds in magenta) of Nap-GFFYGGGWRESAI in PBS buffer solution. (closed symbols: G′ and open symbols: G′′).
networks, leading to the hydrogel formation. We observed the 3D networks of nanofibers in all hydrogels with different concentrations of 6T (Figure 4). However, the diameter and
morphology of the nanofibers in various gels were different. As shown in Figure 4, the diameter of the nanofibers in gels was 38.5 ± 3.9, 48.0 ± 3.0, 51.7 ± 5.1, and 60.3 ± 4.6 nm when the concentration of 6T was 0.05, 0.10, 0.25, and 0.40 wt% to NapGFFYGGGWRESAI, respectively. The difference in nanofiber diameter between each two gel groups is statistically significant (P < 0.05; 50 nanofibers were measured in each gel group by TEM observation). Similar trends were also observed in other gels with 2T or 4T protein (Figure S-16 and S-17, Supporting Information). These observations in TEM images correlated well with the mechanical properties of the gels and indicated that gels with bigger G′ value would possess larger-sized nanofibers. 2.2. Preparation and Characterization of 4T2Cgel
Figure 4. TEM images of gels containing 0.5 wt% of NapGFFYGGGWRESAI and different concentrations of 6T: A) 0.05%, B) 0.10%, C) 0.25%, and D) 0.40% equiv. relative to Nap-GFFYGGGWRESAI.
Adv. Healthcare Mater. 2014, 3, 1804–1811
Since the hexameric protein was formed by Elp-4, Elp-5, and Elp-6 that are capable of being expressed separately, two kinds of proteins with different binding sites could be rationally engineered to the hexameric protein together. The resulting hexameric protein could therefore not only serve as the nanofiber cross-linkers but also provide docking sites for bioactive molecules. In order to achieve this goal, we engineered a plasmid that could express a recombinant protein of C-LytA-Elp-4. The Lytic amidase was anchored at bacterial surface through the interaction between its C-terminal (C-LytA) and the bacterial surface choline.[19] The mixing of C-LytA-Elp-4, TIP-1-Elp-5, and TIP-1-Elp-6 might lead to the formation of a hexameric protein (protein 4T2C in Figure 1D) with two kinds of binding sites;
© 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
wileyonlinelibrary.com
1807
www.advhealthmat.de
FULL PAPER
www.MaterialsViews.com
Figure 5. A) Analytical gel filtration profile and its SDS-PAGE gel of protein 4T2C, B) Rheological measurements in dynamic time sweep mode for gels containing 0.5 wt% of Nap-GFFYGGGWRESAI at the frequency of 1 rad s−1 and strain of 0.5% with different concentrations of 4T2C in PBS buffer solution (down triangles in olive: 0.25% and diamonds in magenta: 0.40% equiv. to Nap-GFFYGGGWRESAI, closed symbols: G′ and open symbols: G′′), C) A schematic illustration for using Rhoda-GGK′-bound 4T2C protein to cross-link self-assembling nanofibers of Nap-GFFYGGGWRESAI to form a hydrogel, and D) drug release profile of Rhoda-peptides from hydrogels (higher binding affinity of peptides to TIP-1 protein results in lower release speed, and vice versa).
one from TIP-1 for nanofiber crosslinking and the other one from C-LytA for bioactive molecule loading. The protein 4T2C could indeed be obtained in a moderate yield (about 5 mg L−1) and good quality (Figure S-1D, Supporting Information). The results by analytical gel filtration (Figure 5A) and ultracentrifugation (Figure S-2D, Supporting Information) indicated the formation of a hetero-hexameric protein structure. The property of the obtained hexameric protein 4T2C for hydrogel formation was subsequently tested. Adding 0.25 wt% (relative to Nap-GFFYGGGWRESAI) of 4T2C protein to the PBS solution of Nap-GFFYGGGWRESAI (0.5 wt%) led to the formation of biohybrid hydrogel (4T2Cgel) within 5 min. The 4T2Cgel was then characterized by rheological and TEM measurements. As shown in Figure 5B, the G′ value of the resulting gels significantly increased with the increase of amounts of 4T2C protein for crosslinking. This result revealed that the hydrogels formed by more 4T2C proteins would have better mechanical properties and that the mechanical property of the hydrogels could be adjusted by two factors including using different concentrations of 4T2C proteins and using different concentrations of peptides. Figure S-18 (Supporting Information) displayed the TEM images of gels containing 0.5 wt% of NapGFFYGGGWRESAI and different amounts of 4T2C proteins. 3D networks of nanofibers were observed in all of the hydrogels with various 4T2C amounts. Moreover, when higher concentration of 4T2C protein was used, the diameter of nanofibers in the resulting 4T2Cgel was larger in the TEM observations,
1808
wileyonlinelibrary.com
which agreed well with the corresponding mechanical property trend of the gels. The rheological and TEM observations revealed that 4T2C proteins are efficient cross-linkers to enhance the internanofiber interactions, thus strengthening the mechanical property of 4T2Cgel. As our hydrogels had the multi-condition controlled property, we predicted that it would be beneficial to their future applications in the fields such as stem cell controlled differentiation and controlled drug release. 2.3. Application of 4T2Cgel in Controlled Delivery of Model Drug In order to demonstrate the potential application of 4T2Cgel in controlled drug release, we synthesized a dye-peptide conjugate (Rhoda-GGK′; Scheme S-2, Supporting Information; Figures S-6 and S-7, Supporting Information for its characterization) as a model drug. It is noted that Rhoda-GGK′ was tested to possess a Kd value to C-LytA binding of 4.5 ± 1.1 × 10−6 M measured by isothermal titration calorimetry (ITC) (Figure S-3, Supporting Information). Besides C-LytA that could bind to Rhoda-GGK′, the four TIP-1 in the 4T2C allowed the heteroproteins to crosslink nanofibers of Nap-GFFYGGGWRESAI, afforded Rhoda-GGK′-loaded 4T2Cgels (Figure 5C). The in vitro releasing profile of Rhoda-GGK′ from the 2C4Tgels was then studied. After the formation of Rhoda-GGK′-loaded 4T2Cgels with concentrations of the 4T2C protein of 0.10, 0.20, 0.25,
© 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Healthcare Mater. 2014, 3, 1804–1811
www.advhealthmat.de www.MaterialsViews.com
3. Conclusions We have designed and obtained 2T, 4T, and 6T proteins with two, four, and six binding sites, respectively, which could be utilized as crosslinkers to form injectable biohybrid hydrogels with controllable mechanical properties. More importantly, we have also engineered a hetero-hexameric protein 4T2C with two kinds of different binding sites. The 4T2C proteins have been demonstrated to not only serve as cross-linkers for hydrogel formation but also provide docking sites for model drug RhodaGGK′ loading. The resulting 4T2Cgels could therefore be a very promising system for controlled delivery of bioactive molecules. This successful example of hetero-oligomeric proteinbased hydrogel system will inspire more exciting research in this emerging field.
4. Experimental Section Chemicals: Fmoc-amino acids were obtained from GL Biochem (Shanghai, China). Rhodamine B was received from J&K. All the other starting materials were obtained from Alfa. Chemical reagents and solvents were used as received from commercial sources. General Methods: 1H NMR spectra were obtained on Bruker ARX 400 using DMSO-d6 as the solvent; HR-MS were received from VG ZAB-HS system (England). HPLC was conducted at LUMTECH HPLC (Germany) system using a C18 RP column with MeOH (0.05% of TFA) and water (0.05% of TFA) as the eluents; TEM images were done on a Tecnai G2 F20 system, operating at 200 kV, TEM samples were prepared as following: a carbon-coated copper grid (from Zhongjingkeyi Technology Co. Ltd., Beijing, P. R. China) was vertically dipped into the hydrogel for 5 s, washed by water two times, and then placed in a dessicator overnight before the TEM measurement. LC-MS was conducted at the LCMS-20AD (Shimadzu) system, and rheology was performed on an AR 1500ex (TA instrument) system using a parallel plate (40 mm) at the gap of 300 µm. Protein Expression and Purification: Expression and purification of proteins were performed by standard recombinant protein technology as previously described.[17] Briefly, DNA fragments corresponding to the ChBD of the major lytic amidase (C-lytA) of S. pneumoniae and human TIP-1 were amplified by polymerase chain reaction (PCR). Gene fragments corresponding to yeast Elp4 (residues 67–372 from full-length
Adv. Healthcare Mater. 2014, 3, 1804–1811
residues 1–456) (yElp4), Elp5 (residues 1–238 from full-length residues 1–309) (yElp5), and full-length Elp6 (residues 1–273) (yElp6) were amplified by PCR from the S. cerevisiae genome. To make four singlechain fusion proteins of the TIP1-Elp4, TIP1-Elp5, TIP1-Elp6, C-lytA-Elp4, the DNA fragment of TIP-Elp4 and C-lytA-Elp4 were cloned into the modified pET32a vector (Novagen). The S-tag and thrombin recognition sites were replaced with a sequence encoding a 3C protease-cleavable segment (Leu-Glu-Val-Leu-Phe-Gln-Gly-Pro). The DNA fragments of TIP1-Elp5 and TIP1-Elp6 were cloned into the pETDuet-1 (Novagen) vector to co-express these two proteins. All resulting proteins contained Trx-His6 tags at their N termini. BL21(DE3) CodonPlus E. coli cells harboring the expression plasmid were grown in LB medium at 37 °C until the OD600 reached 0.6 and then induced with 0.3 × 10−3 M isopropyl-β-D-thiogalactoside at 16 °C for about 16–18 h. After being spun at 5000 rpm for 15 min, E. coli cells were resuspended in T50N500I5 buffer (50 × 10−3 M Tris–HCl pH 7.9, 500 × 10−3 M NaCl, and 5 × 10−3 M imidazole) supplemented with 1 × 10−3 M phenylmethylsulfonyl fluoride, 1 µg mL−1 leupeptin and 1 µg mL−1 antipain. The cells were then lysed by sonication. After the lysates had been centrifuged at 18 000 rpm for 30 min, the supernatant was loaded onto a Ni-NTA agarose column (Qiagen) that was equilibrated with T50N500I5 buffer. The Ni-NTA column was washed with 5 column volumes of T50N500I5 buffer. The Trx-his6-tagged protein was eluted with T50N500I5 buffer containing 500 × 10−3 M imidazole. The eluted proteins loaded on a HighLoad 26/60 Superdex-200 size-exclusion column (GE Healthcare) and eluted with T50N300D1 buffer (50 × 10−3 M Tris–HCl pH 7.4, 300 × 10−3 M NaCl, and 1 × 10−3 M DTT) at a flow rate of 2.5 mL min−1. Each fraction of the column elute was 5 mL. The protein peak was identified by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gel. After digestion with PreScission Protease to cleave the N-terminal Trx-His6-tag, the target protein was purified on a HighLoad 26/60 Superdex-200 sizeexclusion column in 50 × 10−3 M Tris–HCl pH 7.4, 300 × 10−3 M NaCl, and 1 × 10−3 M DTT. However, it is worth mentioning that TIP1-Elp5 and TIP1-Elp6 plasmid should transform into Pubs520 E. Coli cells. The purification method remains the same. Analytical Gel Filtration: Size-exclusion chromatography was performed on an AKTA purifier system using a Superdex 200 10/300 column (GE Healthcare) for TIP1-Elp4 yElp5 yElp6, yElp4 TIP1-Elp5 TIP1-Elp6, TIP1 Elp4 TIP1-Elp5 TIP1-Elp6, C-lytA-Elp4 TIP1-Elp5 TIP1-Elp6. Protein samples were dissolved in buffer containing 50 × 10−3 M Tris–HCl pH 7.4, 300 × 10−3 M NaCl, and 1 × 10−3 M DTT. The column was calibrated with a gel filtration standard from Bio-Rad. Analytical Ultracentrifugation: Sedimentation velocity (SV) was performed in a Beckman/Coulter XL-I analytical ultracentrifuge using double-sector. An additional protein purification step on a Superose 12 10/300 column (GE Healthcare) in 50 × 10−3 M Tris–HCl pH 7.4, 300 × 10−3 M NaCl, and 1 × 10−3 M DTT was performed before the experiments. The experiments were conducted at 30 000 rpm, and 4 °C using interference detection and double-sector cells loaded with 1 mg mL−1 of the protein. The buffer composition (density and viscosity) and protein partial specific volume (V-bar) were obtained using the program SEDNTERP.[20] The data were analyzed using the programs SEDFIT and SEDPHAT. Isothermal Titration Calorimetry: ITC measurements were carried out on a MicroCal Isothermal Titration Calorimeter iTC200 (GE Healthcare) in PBS buffer pH 7.4 at 16 °C. The peptide was dissolved in the same buffer mentioned above and adjusted to pH 7.4. Both peptide and protein solutions were degassed by being spun at 13 000 rpm for 15 min. To measure the binding constant of C-lytA with the peptide (Rhoda-GGK′), an initial injection (0.4 µL) followed by 19 injections (2 µL) peptide (0.750 × 10−3 M) into the calorimeter cell, which was completely filled with protein solution (0.015 × 10−3 M), were collected at 120 s intervals while being stirred at 1000 rpm. The titration data and binding plot were analyzed using MicroCal Origin software with one-site binding model. Peptide Systhesis: The peptide derivative was prepared by solidphase peptide synthesis (SPPS) using 2-chlorotrityl chloride resin and the corresponding N-Fmoc protected amino acids with side chains
© 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
wileyonlinelibrary.com
FULL PAPER
and 0.30 wt% equiv. to Nap-GFFYGGGWRESAI, respectively (Rhoda-GGK′ was fixed at 0.1 wt% for all 4T2Cgels), 0.2 mL of PBS buffer was added on the top of each Rhoda-GGK′-loaded 4T2Cgel. At designated time intervals, the upper PBS buffer was totally taken out for measurement of accumulating release amount of Rhoda-GGK′ from each 4T2Cgel and 0.2 mL of fresh PBS buffer was subsequently added. Figure 5D showed the release profiles of Rhoda-GGK′ from the Rhoda-GGK′-loaded 4T2Cgels. It is obvious that Rhoda-GGK′ could be released from each 4T2Cgel in a sustained manner. Additionally, the amounts of 4T2C proteins in the 4T2Cgels could significantly influence the release profile of Rhoda-GGK′ and higher 4T2C protein concentrations led to slower release rates of RhodaGGK′, which would be mainly due to the more binding sites for Rhoda-GGK′ provided by the proteins. The results demonstrated that our hetero-hexameric protein 4T2C can not only serve as crosslinkers for hydrogel formation, but also endow the resulting protein-based hydrogel with plenty of docking sites for loading and controlled release of model drug Rhoda-GGK′.
1809
www.advhealthmat.de
FULL PAPER
www.MaterialsViews.com
1810
properly protected. The first amino acid was loaded on the resin at the C-terminal with the loading efficiency about 0.5 mmol g−1. 20% piperidine in anhydrous N,N′-dimethylformamide (DMF) was used during deprotection of Fmoc group. Then the next Fmoc-protected amino acid was coupled to the free amino group using O-(Benzotriazol1-yl)-N,N,N′,N′-tetramethyluroniumhexafluorophosphate (HBTU) as the coupling reagent. The growth of the peptide chain was according to the established Fmoc SPPS protocol. At the final step, Naphthalene acetic acid or Rhodamine B was used to attach on the peptide. After the last coupling step, excessive reagents were removed by a single DMF wash for 5 min (5 mL per gram of resin), followed by five steps of washing using DCM for 1 min (5 mL g−1 of resin). The peptide derivative was cleaved using 95% of trifluoroacetic acid with 2.5% of TMS and 2.5% of H2O for 45 min. 20 mL per gram of resin of ice-cold diethylether was then added to cleavage reagent. The resulting precipitate was centrifuged for 10 min at 4 °C at 10 000 rpm. Afterward the supernatant was decanted and the resulting solid was dissolved in DMSO for HPLC separation. Nap-GFFYGGGWRESAI: (This compound was reported in our previous study.)[13]1H NMR (300MHz, DMSO-d6) δ 10.78 (s, 1H), 9.20 (s, 1H), 8.25–8.31 (m, 2H), 8.13–8.20 (m, 3H), 8.02–8.11 (m, 5H), 7.80–7.89 (m, 3H), 7.74 (S, 1H), 7.59–7.64 (d, J = 7.61Hz, 1H), 7.45– 7.56 (m, 2H), 7.39–7.44 (t, 1H)7.28–7.34 (d, J = 7.31Hz, 2H), 7.11–7.23 (m, 12H), 7.01–7.07 (t, 3H),6.93–6.99 (t, 1H), 6.61–6.68 (d, J = 6.64Hz, 2H), 4.53–4.63 (m, 2H), 4.43–4.53 (m, 3H), 4.28–4.39 (m, 4H), 4.09– 4.15 (m,1H), 3.68–3.80 (m, 6H), 3.53–3.64 (m, 6H), 3.46–3.51 (m, 1H), 3.15–3.19 (d, J = 3.17Hz, 2H), 3.06–3.14 (m, 3H), 2.88–3.00 (m, 4H), 2.61–2.79 (m, 4H), 2.26–2.34 (m, 2H), 1.89–1.96 (m, 1H), 1.69–1.80 (m, 3H), 1.33–1.58 (m, 5H), 1.14–1.28 (m, 6H), 0.79–0.89 (t, 6H). MS: calc. M+ = 1614.7, obsvd. (M+1)+ = 1615.7. Rhoda-GGK′: 1H NMR (400MHz, DMSO-d6) δ 8.07–8.12 (d, 1H), 7.84–7.90 (m, 2H), 7.54–7.60 (m, 2H), 7.42 (s, 1H), 7.29 (s, 1H), 7.16 (s, 1H), 7.07–7.13 (m, 1H), 4.16–4.25 (m, 2H), 4.37–4.46 (m, 9H), 3.23– 3.30 (m, 2H), 3.04 (s, 9H), 1.72–1.81 (m, 1H), 1.60–1.71 (m, 3H), 1.22– 1.33 (m, 3H), 1.01–1.12 (t, 12H). MS: calc. M+ = 728.4, obsvd. (M+1)+ = 729.4. Preparation of Solution ofNap-GFFYGGGWRESAI: 5.0 mg of NapGFFYGGGWRESAI (3.10 µmol) was dissolved in 0.50 mL of PBS buffer solution containing 0.66 mg (2 equiv. to Nap-GFFYGGGWRESAI) of Na2CO3 (2 equiv. of Na2CO3 were used to neutralize the carboxylic acids on Nap-GFFYGGGWRESAI to make the final pH value to 7.4). The mixture was heated to form a transparent solution, which was followed by cooling back to room temperature for use. Preparation of Solution of Rhoda-GGK′: 5.0 mg of Rhoda-GGK′ (6.86 µmol) was dissolved in 1.25 mL of PBS buffer solution containing 1.46 mg (2 equiv. to Rhoda-GGK’) of Na2CO3 (2 equiv. of Na2CO3 were used to neutralize the carboxylic acids on Rhoda-GGK′ to make the final pH value to 7.4). The mixture was heated to form a transparent solution and then cooling back to room temperature for use. Preparation of Hydrogels: Take 0.10 mL the solution of NapGFFYGGGWRESAI into a small bottle and then the resulting solution was heated to form a transparent solution, which was followed by cooling back to room temperature, affording a nanofiber solution of Nap-GFFYGGGWRESAI. 0.10 mL of PBS buffer solution containing 65.81 µg, 131.62 µg, 329.05 µg, or 526.48 µg of 2T protein (0.31 nmol, 0.62 nmol, 1.55 nmol, or 2.48 nmol, 0.05%, 0.10%, 0.25%, or 0.40% equiv. to Nap-GFFYGGGWRESAI) was added to the nanofiber solution. Gels would form after the mixture was rapidly stirred within seconds and then kept at room temperature (22–25 °C) for about 3–30 min. Other hexamer proteins of 4T, 6T, or 4T2C with different concentrations were also used to form gels by using similar procedures. Rheology: Rheology test was done on an AR 1500ex (TA instrument) system, 40 mm parallel plates were used during the experiment at the gap of 300 µm. The dynamic time sweep was conducted at the frequency of 1 rad s−1 and the strain of 0.5%. Dynamic strain sweep was performed and the strain values within the linear range were chosen for the following dynamic frequency sweep. The gels were also characterized by the mode of dynamic frequency sweep in the region of 0.1–100 rad s−1
wileyonlinelibrary.com
at the strain of 0.5%. The recovery property of the gels was probed as following: the large amplitude of stain (50%) at the frequency of 0.5 rad s−1 was first applied to the gels for 600 s. And then the recovery property of the gels was measured right after the removal of the external large amplitude of force (50% of strain) at the strain of 0.3% and the frequency of 0.5 rad s−1 for 3600 s. Release Profile: 0.2 mL of 4T2Cgel containing Rhodamine–peptide conjugates was formed as described above (the final concentrations of the protein respectively were 0.10%, 0.20%, 0.25%, and 0.30% equiv. to Nap-GFFYGGGWRESAI, Rhodamine–peptide was 0.1 wt%). After 10 h, each of the gels was treated with 0.2 mL of fresh PBS buffer solutions (pH 7.4). 0.2 mL of the upper buffer solution was taken out and used to tested by UV–vis Spectrophotometer at the wavelength of 560 nm at designated time intervals. A fresh 0.2 mL of PBS was added back to the gel. Standard curve of the model drug was determined before the test. The experiment was conducted in three parallel experiments. In addition, 0.2 mL of 4T2Cgel containing Rhodamine–peptide conjugates was formed in syringes, then injected in the small bottle and formed hydrogels (the final concentrations of the protein respectively were 0.20%, 0.25%, and 0.30% equiv. to Nap-GFFYGGGWRESAI, Rhodamine– peptide was 0.1 wt%). Then measured them by using similar procedures as above. Statistical Analysis: Quantitative data were expressed as mean ± standard deviation. Statistical comparisons were made by ANOVA analysis and Student's t-test. A P value
