Introduction

Stem cells play an important role in tissue engineering and clinical regeneration therapy due to their ability to self-replicate and potential to differentiate into multiple cell lines, called stemness (Keating 2012). The clinical demand for stem cells from various sources far exceeds their supply (Van Zant and Liang 2012). In this regard, an effective solution may be to obtain stem cells through in vitro culture. However, multiple passages of in vitro cultivation gradually decrease the therapeutic properties of stem cells by weakening their proliferation and multipotent differentiation potential, eventually leading to treatment failure (Truong et al. 2019; Zaim et al. 2012). Therefore, develo** more substances that can maintain or even enhance the stemness of stem cells during large-scale in vitro culture is important for their clinical applications.

Extracellular matrix (ECM), a noncellular three-dimensional macromolecular network composed of collagens, fibronectin, elastin, and several other glycoproteins (Ishihara et al. 2014; Isomursu et al. 2019), exerts essential functions on the proliferation and differentiation of stem cells (Du et al. 2011). Extensive research and testing have shown that collagen promotes stem cell migration and adhesion (Sorushanova et al. 2019). Fibronectin promotes cell adhesion similar to collagen, and it also fixes collagen to maintain the ECM stability. In addition, the ECM communicates with cells through integrins located on the surface of the cell membrane, such as αvβ3, αvβ1, and αvβ6, via Arg-Gly-Asp (RGD) sequences (Pierschbacher and Ruoslahti 1984). FN10, a part of the fibronectin type III module, contains RGD sequences that specifically bind αvβ1 to guide the fate of the cells (Bharadwaj et al. 2017; Hocking et al. 1996). Moreover, collagen and fibronectin support the growth of human embryonic stem cells (hESC) without the use of a feeder layer under certain conditions (Lu et al. 2006) and play a role in maintaining the stemness of stem cells (Akhir and Teoh 2020; Thaweekitphathanaphakdee et al. 2019). Notably, as the main strain matrix element, collagen relies on fibronectin, fibronectin-bound, and collagen-bound integrins to complete its corresponding functions (Kadler et al. 2008; Kubow et al. 2015). Meanwhile, natural collagen and fibronectin from animal sources have some problems such as unstable quality and difficult dissolution (Davison-Kotler et al. 2019). In contrast, the recombinant proteins based on collagen and fibronectin and prepared by synthetic biology possess the excellent characteristics of high yield, easy amplification, low cost, and not easily contaminated by mammalian pathogens. These characteristics are slowly being recognized as contributors leading to the future application of recombinant proteins based on collagen and fibronectin in regenerative medicine (Ferrer-Miralles and Villaverde 2013). Therefore, it is necessary to use genetic engineering technology to synthesize a novel functional protein that acts as a biomimetic ECM for stem-cell therapy.

Integrin plays an important role in the long-term ex vivo culture of stem cells. As a member of the integrin family, integrin β3 is a transmembrane receptor that mediates interactions between cells and the ECM and which widely exists on the surface of hematopoietic stem cells and bone marrow mesenchymal stem cells (Pei et al. 2011; Umemoto et al. 2012). The integrin-αvβ3 found on hematopoietic stem cells (HSCs) plays an important role in maintaining stem cell activity by regulating stem cell function and affecting stemness through signal transduction (Umemoto et al. 2012; de Graaf and Metcalf 2011; Ishihara et al. 2014). These integrin functions are dependent on specific ECM ligand–receptor interactions and the specific molecular interactions that lead to cytoskeletal changes that result in different migratory behavior or changes in growth and differentiation (Hynes 2002; Isomursu et al. 2019). It is interesting to design innovative proteins based on the functional characteristics of the principal ECM structural proteins that contain specific ECM ligands, such as RGD, and which can bind with integrin β3 to de-anchor and adsorb cells.

Here, we reported a recombinant protein FCP improved according to the early exploration of our laboratory. FCP structure and functions, including binding to integrin β3, were predicted and recombinant fusion expression assays were conducted. Experiments verified that FCP promoted cell migration and adhesion and maintained cell stemness by binding integrin β3. Its effects could be blocked both by integrin β3 silencing and employing the integrin β3 inhibitor Cilengitide. Therefore, FCP may be used in the in vitro culture of stem cells, such as hematopoietic stem cells and other stem cells with high expression of integrin β3, or even in the field of tissue regeneration engineering.

Materials and methods

Materials

Vectors pPICZαA (Invitrogen, Guangzhou, China) and P. pastoris strain X33 (Invitrogen, Guangzhou, China, Invitrogen™ C18000, ATCC® 28,485™) were used for cloning and heterologous expression. PCR purification kits, gel extraction kits, and micro preparation kits were purchased from Tiangen (Bei**g, China). CC (a collagen-like protein, designed from COLA1, ID: BC036531.2, 2120-2164nt) was artificially synthesized by GL Biochem (Shanghai, China). FN10(ID: U42594.1, 633-770nt) was recombinantly expressed in our Lab (Fig. S2).

ECV304 (Human umbilical vein endothelial cells, ATCC® CRL-1998) and ECV304-eGFP (Human umbilical vein endothelial cells-enhanced green fluorescent protein, ATCC® PCS-100–010) cells were purchased from the Chinese Academy of Sciences (Shanghai, China). They were cultured in Roswell Park Memorial Institute 1640 medium (RPMI 1640, Gibco, Carlsbad, CA, USA) containing 10% fetal bovine serum, penicillin (100 I.U./mL), and streptomycin (100 μg/mL) (Sangon Biotech, Shanghai, China). Rat bone marrow mesenchymal stem cells (rBMSCs) were extracted in our laboratory approved by the Animal Care Committee of **an University (JNU20200826-11) and were performed in accordance with animal ethics guidelines of Agricultural Animals for Research and Teaching at **an University. rBMSCs were cultured in Dulbecco’s modified Eagle’s medium (DMEM, Gibco) supplemented with 10% fetal bovine serum, penicillin (100 I.U./mL), and streptomycin (100 μg/mL).

Construction and identification of FCP

The cDNA encoding FCP was inserted into the pPICZαΑ vector to give a recombinant plasmid named pPICZαΑ-FCP; then, the pPICZαΑ-FCP was transformed into P. pastoris strain X33. Zeocin-resistant clones were selected on YPD plates containing 0.1 mg/mL of Zeocin to obtain different copies of the integrated DNA of FCP. All strains were cultivated overnight at 30 °C in 5 mL of YPD medium, and 1% of the culture was transferred to 25 mL of YPG in a 250 mL baffled flask. The cells were grown at 30 °C in a shaking incubator until the OD600 reached 1.5. The cell pellets were then induced in YPM. At the same time, methanol (0.5, 1.0, 1.5, 2.0%) was added to the culture, and 1-mL samples were collected at 0, 24, 48, and 72 h after induction. The expressions of proteins were monitored by SDS-PAGE. FCP proteins were purified using the Ni Sepharose 6 Fast Flow column combined with gel filtration Sephadex G-25. Polymerase chain reaction (F: 5′-CATGGTGCTCCAGGTGCCCCTGGATC-3′, R: 5′-TGGCTTATCAATTTCAGTTCTGTAG-3′), agarose gel electrophoresis, and western blot (His-Tag Antibody, Affinity, Biosciences, Changzhou, China), etc., were used to verify the authenticity of FCP. Expanded culture of highly expressing P. pastoris strain X33 in 2L YPD media was performed. Additionally, the circular dichroism (CD) spectrum of FCP was measured with a Chirascan plus Circular Dichroism Spectrophotometer (Applied Photophysics Ltd., Leatherhead, Surrey, UK).

Cell migration assay: scratch wound assay

An in vitro scratch wound healing model was used to assess the cell migration induced by FCP. Briefly, ECV304 cells were plated in 12‐well plates and cultured until approximately 95% confluence before the scratch study was undertaken. A scratch was created with a 1 ml sterile pipette tip. After scratching, cells were washed using PBS to remove cell debris caused by the scratch. Another 1 mL treatment solution (250 nmol/L FCP, or FN10, or CC in 1640 medium containing 1% FBS, respectively) was added to cells. Three images of the scratch area were photographed using an inverted microscope (Olympus IX70, Tokyo, Japan) at 0, 18, and 36 h. The ImageJ software was used to determine percentage closure (%).

In another separate experiment, ECV304 cells were treated with or without FCP, integrin β3 siRNA, and NC group (negative control: cells transfected with scrambled siRNA), respectively. All operations are performed as above.

Cell adhesion assay: crystal violet staining

Crystal violet staining was performed to assess the adhesion of FCP. Briefly, culture plates were coated with 250 nmol/L of FCP or Cilengitide overnight at 4 °C. Next, the plates were washed three times with PBS. Subsequently, ECV304 (1.0 × 105 cells/mL) were seeded on tissue culture plates (96-well) and allowed to attach for 4 h at 37 °C. After incubation, non-attached cells were removed by three rinses with PBS. The remaining cells were fixed with 2% paraformaldehyde for 20 min and stained with 1% crystal violet (Solarbio, Bei**g, China) for 20 min.

Immunofluorescence staining assay

After adhesion, cell spread areas were characterized by vinculin and TRITC Phalloidin and detected by immunofluorescence staining assays via a confocal microscope. Briefly, ECV304 (2.0 × 105 cells/mL) were seeded on plates (24-well) and calculated approximately 95% confluence. Then, after being treated with different test substances for 24 h, we discarded culture supernates and washed ECV304 cells using PBS 3 times, following fixed them in 4% paraformaldehyde. In the next moment, washed with PBS and then permeabilized with ice-cold 0.5% Triton X-100. The cells were blocked with PBS containing 2% bovine serum albumin, then incubated with primary antibody of vinculin (Proteintech, Wuhan, China, Cat#26,520–1-AP, 1:100) overnight at 4 °C; after that, Alexa Fluor 488 goat anti-rabbit IgG (Sigma, UK, Ca#SLBX2002, 1:80) were co-incubated at room temperature for 30 min. Next, TRITC Phalloidin (Solarbio, Bei**g, China, Cat#CA1610, 1:200) was incubated at room temperature for 10 min. Finally, the sections were stained with 4′,6-diamidino-2- phenylindole (DAPI) (Cell Biolabs INC, San Diego, CA, Cat#112,002, 1:1000) and imaged with a confocal laser scanning fluorescence microscope (LSM700, Zeiss, Wetzlar, Germany).

There were three separate immunofluorescence staining assays. The test substances of the first immunofluorescence staining assays were FCP, CC, and FN10 (250 nmol/L). ECV304 cells were treated with FCP, integrin β3 siRNA, or scramble siRNA for 24 h separately in the second separated assays. In the last separate experiment, ECV304 cells were treated with FCP, Cilengitide, and FCP combined with Cilengitide (250 nmol/L) for 24 h. All test corresponding control groups were given.

Tube formation assay

The tube formation assay was performed with ECV304-eGFP. Briefly, 96-well plates were coated with 50 μL/well matrigel, then polymerized at 37 °C for 30 min. Then, ECV304-eGFP (3 × 104) were added to each well and incubated at 37 °C and 5% CO2 for 12 h. Cells were respectively treated with or without FCP, FN10, and CC (250 nmol/L) for 4 h. Then, tube formation was observed and captured with an inverted microscope (Olympus IX70, Tokyo, Japan).

Protein modeling and molecular docking

The structural modeling of FCP was performed using bioinformatic webservers such as SWISS-MODEL (https://swissmodel.expasy.org/) (Waterhouse et al. 2018), I-TASSER (https://zhanggroup.org/I-TASSER/) (Yang and Zhang 2015), RoseTTAFold (https://robetta.bakerlab.org/) (Baek et al. 2021) and Alphafold2 (https: // colab.research.google.com / github / deepmind / alphafold / blob / main / notebooks / AlphaFold.ipynb) (Jumper et al. 2021). These webservers cover different modeling methods including homology modeling, threading modeling, deep learning-based ab initio modeling. FCP structure model was evaluated and verified using Ramachandran Plot (Carugo and D**ovic-Carugo 2013), Errat (Colovos and Yeates 1993), Prove (Pontius et al. 1996), Whatcheck (Hooft et al. 1996) and Verify3D (Bowie et al. 1991) evaluation procedures by SAVES6.0 (https://saves.mbi.ucla.edu/).

To explore the molecular interaction, the PatchDock server (http://bioinfo3d.cs.tau.ac.il/PatchDock/php.php) (Schneidman-Duhovny et al. 2005) was used to simulate the interaction between FCP (the highest quality model) and integrin β3 (PDB ID: IJV2).

Utilizing the FireDOCK server (http://bioinfo3d.cs.tau.ac.il/FireDock/) to screen the docking results from PatchDock. Finally, 3D and 2D plot analyses of the protein interaction surface through PyMoL v.2.5.1 and Ligplot+v.2.2 respectively.

Surface plasmon resonance (SPR)

The interactions between FCP and integrin β3 were examined by SPR (Nicoya Life Science, Waterloo, Canada) (** biomaterials is to design biomaterials that contain specific domains which can interact with biological systems to meet various medical and health needs. Protein-functional materials have a distinct advantage in medical and green manufacturing due to their good biocompatibility, biodegradability, multifunctionality, and environmental friendliness.

This study designed a novel recombinant human fusion protein, FCP, with specific protein domains that bind the integrin β3 receptor to communicate with specific stem cells. FCP is composed of FN10 and CC. FN10 is a functional domain of FN that contains the RGD sequence. CC is a recombinant collagen-like protein, developed by our laboratory, that promotes the migration and adhesion of NIH/3T3 cells (Cheng et al. 2020). In a full-thickness animal wound model, CC dressing significantly promotes wound healing and angiogenesis. However, CC exhibits thermal instability, a characteristic that leaves great room for improvement. Thus, in the present study, the CC fragment was combined with FN10 to construct and express FCP protein, improving the performance of the recombinant protein to meet the needs of tissue engineering. Compared with the individual CC and FN10 components, FCP promotes improved cellular proliferation, adhesion, and angiogenesis. The comparative analysis of three proteins shows a satisfying result. All aspects of the performance of FCP improved significantly.

To better understand the structure and function of FCP, the FCP protein structure was predicted using various protein modeling methods. Model quality assessment showed that the FCP model based on ab initio prediction without a homologous template, which was optimized by deep learning, reflected the FCP structure most realistically. Various modeling approaches were tested in optimally constructing a stable FCP structure. The homology and threading methods accurately and efficiently construct the spatial conformations of target proteins with homologous sequences and similar topologies based on the vast amount of known protein structure information obtained from the database. However, these methods are not applicable to unknown proteins that lack template information (Waterhouse et al. 2018; Yang and Zhang 2015). In contrast, for newly designed proteins, such as FCP, a template-independent ab initio prediction modeling approach is best for constructing stable protein models. However, this modeling method is computationally resource intensive and time-consuming. The Alphafold2 server based on a deep learning algorithm was also tested for protein structure modeling; however, the model quality was not satisfactory (Supplement Fig. 3). We speculate that this may be due to the largely irregular curl structure of FCP leading to poor construction.

After obtaining a high-quality structural model of FCP, molecular docking studies of FCP and integrin β3 were performed. Computer simulations showed that the FN10 C-terminus within the FCP protein could form a stable complex with the extracellular segment of integrin β3, consistent with our original design goal. According to these results, modification of the residues on the docking interaction surface of the FCP protein and integrin β3 may potentially enhance the affinity of FCP for integrin β3.

The high affinity of FCP for integrin β3 was verified by SPR assay. Integrin β3 protein, captured on an NTA chip, bound FCP with an affinity constant of 5.37 × 10–8 M, as determined by the SPR assay. Thus, the integrin β3 protein strongly binds the FCP protein (10–8 ~ 10–10 M). The SPR technique is a powerful tool in the study of target molecule interaction and has the advantages of low cost and direct and quantifiable results. However, because SPR is sensitive to interfering factors within the sample, it represents an indirect interaction between proteins.

Collectively, these data suggest that integrin β3, an important receptor on the surface of ECV304 cells, may be a key target for FCP to exert its biological effects. By knocking down integrin β3 expression in ECV304 cells and competitively blocking integrin β3 binding sites with Cilengitide, FCP-induced angiogenesis, migration, and adhesion of ECV304 cells were reduced. rBMSCs were chosen as the model system in the present studies due to ethical constraints and other difficulties in obtaining hHSCs. Considering that rBMSCs are one of the most widely used stem cells in the field of tissue engineering (Chen et al. 2018; Li et al. 2019) and that integrin β3 is highly expressed on the surface of the BMSC cytomembrane (Pei et al. 2011), rBMSCs were utilized in the cloning formation assays. FCP enhanced the colony-forming ability of rBMSCs, and this effect was significantly decreased by Cilengitide pretreatment. After treatment with both FCP and Cilengitide, the related genes were upregulated compared with those in the Cilengitide only group. Likewise, the stemness-related genes NANOG and REX1 expression was increased following FCP administration when compared to control, and this increase remained 14 days after treatment. The gene expression levels of both SOX and ALP, markers of chondrogenic and osteogenic differentiation abilities of rBMSCs, were both higher than that observed in other genes. However, the gene expression of PPAR-γ2, an adipogenic gene, was similar to that of the control (P > 0.05). This finding is difficult to explain, but we speculate that it is related to the function of integrin β3. Further study of these results was outside the scope of the current study due to laboratory constraints. For a sample, we are unable to test karyotypic stability after transfer generation.

Functional recombinant proteins that are used as replacements for native proteins have many unique characteristics. Their high molecular weight, high-frequency amino acids, and special posttranslational modifications serve as bottlenecks in the artificial cell synthesis field and lead to low expression rates and poor adaptation between functional elements and host cells. Functional recombinant protein structures may be unstable or ineffective, limiting efficient production and application of protein-functional materials. Therefore, it is particularly important to ab initio design protein molecules and construct a cell factory to realize functional directional enhancement and industrial production of these functional recombinant proteins. Fibronectin and collagen are principal structural proteins of the ECM. By combining their ECM functions allows us to more comprehensively understand the molecular structure of ECM and its bioactivity in order to design new therapeutic compounds in the future. FCP combines the FN10 (one of fibronectin functional domain-containing RGD sequence) and CC (a kind of human-like collagen) proteins, which have an ideal domain to provide mechanical support for cells. The FCP bionic ECM may provide a microenvironment to support stem cell localization, proliferation, differentiation, and stemness maintenance. In the future, FCP may be used in the in vitro culture of stem cells, such as hematopoietic stem cells and other stem cells with high expression of integrin β3, or even in the field of tissue regeneration engineering.