Design of bFGF-tethered self-assembling extracellular matrix proteins via coiled-coil triple-helix formation

The pBluescript SKII(-) plasmid was purchased from Toyobo (Japan). The pUC18, pET28b, pET32c plasmids and E. coli BL21 (DE3) were obtained from Novagen. Escherichia coli KRX was purchased from Promega. Synthesized oligo DNA were obtained from FASMAC (Japan). Restriction enzymes and ligase were purchased from Takara Bio (Japan). All other chemicals were analytical grade.

Three plasmids were constructed to conduct following experiments described in this paper: pET28b-hA-ERE-hA-CHis, pET28b-hC-ERE-hC-CHis, and pET32c-NHis-hB-(GGGS)₈-bFGF. The pET32c-NHis-ERE, pET28b-hA-CHis, pET28b-hB-CHis, pET28b-hC-CHis, pBS-(GGGS)₈, and pET32c-NHis-bFGF plasmids were previously constructed in our lab [14, 36].

To obtain the ERE fragment, pET32c-NHis- ERE was digested with BamHI and BglII. The resultant fragment was ligated into a pUC18 vector, which included the adapter sequence BamHI-BglII-BspHI-EcoRI, digested with BglII. pUC18-ERE was digested with BamHI and EcoRI and ligated to pET28b-hA-CHis or pET28b-hC-CHis using the same enzymes. The resultant pET28b-hA-ERE-CHis or pET28b-hC-ERE-CHis sequences were digested with XbaI and BspHI and inserted into the XbaI and NcoI digested sites of the pET28b-hA-CHis or pET28b-hC-CHis plasmids, respectively. The resulting plasmids were named pET28b-hA-ERE-hA-CHis and pET28b-hC-ERE-hC-CHis, respectively.

To obtain the (GGGS)₈ linker fragment, pBS-(GGGS)₈ was digested with BglII and EcoRI. The resulting fragment was ligated to pET28b- hB-CHis digested with BamHI and EcoRI. The resultant pET28b-hB-(GGGS)₈-CHis sequence was digested with NcoI and EcoRI and ligated to the pET32c-NHis-bFGF plasmid, which included the same restriction enzymes sites. The resulting plasmid was named pET32c-NHis-hB-(GGGS)₈-bFGF.

For expression of hA-ERE-hA (ARA) and hC-ERE-hC (CRC), E. coli KRX competent cells were transformed with pET28b-hA-ERE-hA-CHis and pET28b-hC-ERE-hC-CHis plasmids, respectively. Transformed E. coli cells were cultured at 37 °C in a Luria−Bertani (LB) medium with 20 μg ml⁻¹ kanamycin at an optical density of 0.6 at 660 nm. Protein expression was induced by addition of 1 mM isopropyl-β-D-thiogalactopyranoside (IPTG) and 0.1% rhamnose. Cells were cultured for 5 h at 30 °C and harvested by centrifugation. Collected cells were suspended in Bug Buster Protein Extraction Reagent (Novagen) with Benzonase Nuclease (Novagen), followed by rotation at room temperature for 30 min after pelleting by centrifugation (8000 g, 4 °C) for 15 min, supernatants were collected. The objective fusion proteins were purified by His-select Nickel Affinity Gel (Sigma-Aldrich) using a Poly-Prep column (Bio-Rad) from the soluble fractions. After 2 h incubation at 4 °C, the column was washed five times with the wash buffer (16 mM Na₂HPO₄, 4 mM NaH₂PO₄, 500 mM NaCl, pH 8.0) of three times volume of resin. Then, the column was washed three times with the same buffer containing 5 and 10 mM imidazole each (six times in total) of three times volume of resin. The fusion proteins were eluted three times with the elution buffer (16 mM Na₂HPO₄, 4 mM NaH₂PO₄, 500 mM NaCl, 500 mM imidazole, pH 7.4) of an equivalent volume of resin. The fusion proteins were then dialyzed in a phosphate-buffered saline (PBS) and concentrated using an Amicon Ultra filter (Merck Millipore).

For expression of hB-bFGF, E. coli BL21 (DE3) competent cells were transformed with pET32c-NHis-hB-(GGGS)₈-bFGF plasmid. Transformed E. coli cells were cultured at 37 °C in LB medium with 50 μg ml⁻¹ ampicillin at an optical density of 0.6 at 660 nm. Protein expression was induced by addition of 1 mM IPTG. Cells were cultured overnight at 16 °C. The protein was purified by the procedure described above for ARA and CRC.

The purity of all fusion proteins were evaluated by 15% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The concentrations of the purified fusion proteins were examined using the bicinchoninic acid (BCA) Protein Assay Kit (Pierce).

ARA and CRC proteins were dissolved in 3-morpholinopropanesulfonic acid (MOPS), pH 7.0, at a final concentration of 50 μM. ARA was mixed with an equivalent volume of CRC, and the mixture was incubated at 4 °C for 1 h to promote self-assembly. The sample was adsorbed onto a hydrophilically treated specimen grid covered with a carbon film (Nisshin EM) for 1 min. After washing with ultrapure water, the grid was stained with 0.25% phosphotungstic acid (PTA) for 1 min. After washing with ultrapure water again, the stained sample was dried on filter paper at room temperature for 2 h or more. TEM images were taken with a transmission electron microscope H-7500 (Hitachi) at 120 kV.

Human umbilical vein endothelial cells (HUVECs) were obtained from Kurabo (Japan). These cells were maintained at 37 °C under 5% CO₂ condition in HuMedia-EG2 supplemented with 2% fetal bovine serum, 10 ng ml⁻¹ human EGF, 5 ng ml⁻¹ human bFGF, 1.34 μg ml⁻¹ hydrocortisone, 10 μg ml⁻¹ heparin, 50 μg ml⁻¹ gentamicin, and 50 ng ml⁻¹ amphotericin B.

The helix-fused ERE proteins (ARA and CRC) were prepared in PBS at a final concentration of 100 nM. These protein solutions were added to a 96-well suspension culture plate (Sumilon MS-8096R) to coat the surface. A fibronectin-coated surface was prepared for as a positive control. Uncoated plates and plates coated with the helix-combined linker proteins hA-(GGGS)₈-hA (ALA) and hC-(GGGS)₈- hC (CLC) were prepared for as negative controls. After 1 h incubation at 37 °C, the protein-coated wells were washed with PBS. HUVECs were seeded in each plate at a density of 10⁴ cells/well. After 3 h incubation at 37 °C, plate surfaces were washed with PBS. Attached cells on the plates were analyzed by Cell Counting Kit (CCK-8) (Dojindo) using water-soluble tetrazolium salt, WST-8. The activities were measured at a wavelength of 450 nm as described in the CCK-8. The cells were concomitantly stained with Calcein-AM (Dojindo) and observed using a fluorescence microscopy.

HUVECs were utilized to evaluate the cell proliferation activity of hB-bFGF. HUVECs were seeded in a 96-well tissue culture plate (Costar 3595) at a density of 2000 cells/well. Then hB-bFGF was supplied with the HuMedia-EG2 without bFGF. The cells were cultured at 37 °C under 5% CO₂ condition. The culture media were changed on the first and fourth day. The day on which hB-FGF was added was defined as day 0. The activity was evaluated at an absorbance of 450 nm as described in the CCK-8 on days 1 and 4.

The surfaces of a non-treated 96-well plate (Sumilon MS-8096R) were modified with 100 nM of ARA–CRC by hydrophobic interaction. After 1 h incubation at 37 °C, the surfaces were washed five times with PBS-T (PBS including 0.05% Tween 20) followed by blocking with Blocking One (Nacalai Tesque) incubated overnight at 4 °C. After washing five times with PBS-T, various concentrations (1, 10, 100, and 1000 nM) of hB-bFGF were added to the ARA–CRC modified plates. After incubation for 1 h at 37 °C to allow triple helix formation, the surfaces of the plates were washed five times with PBS-T followed by reaction with 1/5000 diluted rabbit anti-bFGF antibody (Sigma) incubated for 40 min at room temperature. After washing with PBS-T, anti-rabbit IgG antibody labeled peroxidase (Jackson ImmunoResearch) was added and incubated for 40 min at room temperature. After washing with PBS-T, 3, 3', 5, 5'-tetramethylbenzidine (TMB) horseradish peroxidase (HRP) substrate (Kirkegaard and Perry Laboratories) was reacted with protein-modified surfaces of the plates. Atter 2 min incubation, 1 N HCl was added to stop the reaction, and the absorbances at 450 nm were measured.

The surfaces of a 96-well hydrophobic plate (Sumilon MS-8096R) were coated with 100 nM of ARA–CRC and incubated for 1 h at 37 °C. After washing three times with PBS, 10 nM of hB-bFGF was added and incubated for 1 h at 37 °C to form coiled-coil triple helix structure. After washing three times with PBS, HUVECs suspended in HuMedia-EG2 without bFGF were seeded onto the wells at a density of 2000 cells/well. The culture media were changed on the second, fifth, and seventh day. The numbers of cells on days 2 and 7 were counted by WST-8 assay using the CCK-8.

The constructed helix-fused proteins are illustrated in figure 2(a). ARA and CRC were created by fusing helix peptides hA or hC to both ends of the matrix protein ERE, which is composed of two (APGVGV)₁₂ peptides along with the RGD tripeptide motif. The RGD sequence is a major cell attachment site found in various ECMs that interact with integrin receptors [37]. On the other hand, bFGF-tethered hB peptides (hB-bFGF) were constructed by fusing hB with the linker peptide (GGGS)₈ to the N-terminus of bFGF. Note that in our previous work, C-terminus modified bFGF-fusion proteins were obtained from only insoluble fraction, and showed no proliferation activity, even after refolding. The (GGGS)₈ linker was also inserted between hB and bFGF to enhance the flexibility of bFGF and efficiently transmit a signal to the cell without losing activity upon immobilization of bFGF to the scaffold via coiled-coil interactions. These fusion proteins were genetically produced by E. coli KRX or BL21 (DE3) as a Histidine-tag (His-tag)-fusion protein and purified from the soluble fractions using a coordinate bond with His-select nickel affinity resins. The purities of the fusion proteins were evaluated by SDS-PAGE (figures 2(b) and (c)). The bands appeared at the expected molecular weights (ARA: 24.4 kDa; CRC: 24.4 kDa; hB-bFGF: 30.0 kDa). The purities of fusion proteins ARA, CRC, and hB-bFGF were 80.2%, 84.9%, and 60.3%, respectively, estimated by ImageJ software. The fusion proteins were considered pure enough for further experiments.

Zoom In Zoom Out Reset image size

TEM imaging was used to examine the self-assembling capacity of a mixture of ARA and CRC. Fibril formation was expected upon specific coiled-coil interactions between hA of ARA and hC of CRC. The helix peptides hA, hB, and hC form not only heterotrimer coiled-coils but also heterodimer coiled-coils comprised of each combination of two helix peptides, each of which exhibits different associative strengths (hA-hC > hB-hC > hA-hB) [34]. Representative TEM images of the ARA–CRC complex are shown in figure 3. The sample formed a fibrillar structure in which two fibers of 500 nm in width were twisted upon other. Mature fibrils appeared to be formed by lateral aggregation arising from the elastin-derived hydrophobic APGVGV repeats following the formation of nanofibers by supramolecular assembly between ARA and CRC. This result suggests that the introduced helix peptides promote self-assembly of the matrix proteins. Although some studies have addressed nanofiber formation by self-assembly of short-chain peptides, nanofibers composed of ECM proteins can be constructed by fusing helix peptides that form specific coiled-coils to both ends of the matrix proteins. In this study, we used the repeated APGVGV sequence which is one of the motifs called elastin-like polypeptides (ELPs) as a structural frame of the ECM proteins. It has been proposed that the APGVGV polypeptide forms a stable β-spiral structure [38] that provides flexibility and stability in the ECM proteins. Meanwhile, the APGVGV motif has a propensity to aggregate in high concentration due to increasing hydrophobic interaction. To inhibit aggregation and control fiber width, incorporation of the hydrophilic group could be effective. Indeed, previous experiments by our group have shown that inducing polyaspartic acid chains regulated aggregation of ELPs by the negatively charged repulsion. As a result, the polyaspartic acid fused ELPs formed size-controlled nanoparticles [39–41]. Therefore, aggregation of high-hydrophobic ECM proteins including ELPs could be controlled by introducing electrically charged amino acid chain such as polyaspartic acid.

Zoom In Zoom Out Reset image size

HUVECs were seeded onto a surface coated with ARA and CRC to evaluate the cell adhesion activity of these ECM proteins. Non-coated, hA-(GGGS)₈-hA (ALA)-coated and hC-(GGGS)₈-hC (CLC)-coated plates were used as negative controls, while ERE-coated and fibronectin-coated plates were used as positive controls. The repeating APGVGV sequence contained in ERE, ARA, and CRC is highly hydrophobic, allowing for the ECM proteins to adsorb hydrophobic plate surface. After 3 h of incubation, the numbers of cells on the plates were examined (figure 4(a)). HUVECs seeded onto the surfaces coated with ALA or CLC were barely attached to the plate, similar to the non-treated plate. Meanwhile, adequate cell adhesion and extension were observed on the plates coated with ARA, CRC, and their mixture (ARA–CRC), as well as ERE-coated and fibronectin-coated plates (figure 4(b)). These results suggest that cell adhesion was promoted by the RGD sequence of ARA and CRC. Thus, the introduced helix peptides do not disturb the ERE functions, and the designed ECM proteins exhibit cell adhesion activities comparable to that of fibronectin.

Zoom In Zoom Out Reset image size

Cell proliferation activity of hB-bFGF was investigated. The hB-bFGF was added to the HUVEC culture medium at the final concentration of 10 or 100 nM. The bFGF is an essential factor for the proliferation of HUVECs in low serum condition. The proliferation of HUVECs induced by adding hB-bFGF is shown in figure 5. HUVECs significantly proliferated in the presence of bFGF or hB-bFGF, while the same cells showed no growth without bFGF. The helix-fused protein induced cell growth at levels comparable to wild-type bFGF at each concentration. The results indicate that hB-bFGF preserved the activity as bFGF even after fusing hB-(GGGS)₈.

Zoom In Zoom Out Reset image size

ELISA analyses were carried out to examine the binding activity of hB-bFGF to ARA–CRC via heterotrimer coiled-coil interaction. Various concentrations of hB-bFGF were immobilized to the ARA– CRC modified surfaces. Anti-bFGF antibody was used to determine the amounts of hB-bFGF on the surface. As shown in figure 6(a), the signals detected with hB-bFGF increased in a concentration-dependent manner up to 1000 nM. However, there is a possibility of nonspecific binding of hB-bFGF. To elucidate that the interaction between hB-bFGF and ARA–CRC was due to the triple-helix formation between hA, hB, and hC presented each protein, competitive inhibition of the binding of hB-bFGF with hB peptide was performed. Various concentrations of hB inhibitor peptide were mixed with 100 nM of hB-bFGF solution, and the mixtures are reacted with the ARA–CRC modified surfaces. The intensities of the signal based on the amounts of hB-bFGF were evaluated as described above and shown in figure 6(b). The hB peptide inhibited the hB-bFGF binding to the surfaces modified with ARA–CRC dependent on the hB concentration. It is evident that the interaction between hB-bFGF and ARA–CRC was due to the triple-helix formation between hA, hB, and hC. Thus, the helix peptides functioned as not only self-assembling domains but also as growth factor-tethering domains for ECM proteins.

Zoom In Zoom Out Reset image size

The proliferation of HUVECs cultured on the surface of bFGF-tethered self-assembling ECM proteins was evaluated. HUVECs were grown on the hB-bFGF-presenting ARA–CRC coated plate surface, and their proliferation activities were examined (figure 7). As control experiments, HUVECs were seeded onto the ARA–CRC-coated plate with or without bFGF. In these cases, cell growth was scarcely observed or less active than when utilizing hB-bFGF. It is presumed that bFGF was washed out at the time of medium change on the second, fifth, and seventh day. Meanwhile, hB-bFGF tethered self-assembling ECM protein-absorbed surface induced significant cell proliferation. These results indicated that hB-bFGF exhibits intact bFGF activity without impairment and sustained its activity for a longer period by immobilizing to the ECM proteins via triple-helix formation.

Zoom In Zoom Out Reset image size