Accepted Manuscript Title: Graded functionalization of biomaterial surfaces using mussel-inspired adhesive coating of polydopamine Authors: Sajeesh Kumar Madhurakkat Perikamana, Young Min Shin, Jin Kyu Lee, Yu Bin Lee, Yunhoe Heo, Taufiq Ahmad, So Yeon Park, Jisoo Shin, Kyung Min Park, Hyun Suk Jung, Seung-Woo Cho, Heungsoo Shin PII: DOI: Reference:

S0927-7765(17)30537-4 http://dx.doi.org/10.1016/j.colsurfb.2017.08.022 COLSUB 8778

To appear in:

Colloids and Surfaces B: Biointerfaces

Received date: Revised date: Accepted date:

9-5-2017 11-8-2017 14-8-2017

Please cite this article as: Sajeesh Kumar Madhurakkat Perikamana, Young Min Shin, Jin Kyu Lee, Yu Bin Lee, Yunhoe Heo, Taufiq Ahmad, So Yeon Park, Jisoo Shin, Kyung Min Park, Hyun Suk Jung, Seung-Woo Cho, Heungsoo Shin, Graded functionalization of biomaterial surfaces using mussel-inspired adhesive coating of polydopamine, Colloids and Surfaces B: Biointerfaceshttp://dx.doi.org/10.1016/j.colsurfb.2017.08.022 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Graded functionalization of biomaterial surfaces using

mussel-inspired

adhesive

coating

of

polydopamine Sajeesh Kumar Madhurakkat Perikamana a,b, ‡, Young Min Shin a, ‡, Jin Kyu Lee a,b, Yu Bin Lee a,b

, Yunhoe Heo a,b, Taufiq Ahmad a,b, So Yeon Park c, Jisoo Shin d, Kyung Min Park e, Hyun Suk

Jung c, Seung-Woo Cho d and Heungsoo Shin a,b* a

Department of Bioengineering, Hanyang University, 222 Wangsimni-ro, Seongdong-gu, Seoul,

04763, Republic of Korea b

BK21 Plus Future Biopharmaceutical Human Resources Training and Research Team, 222

Wangsimni-ro, Seongdong-gu, Seoul, 04763, Republic of Korea c

School of Advanced Materials Science & Engineering, Sungkyunkwan University, Suwon

16419, Republic of Korea. d

Department of Biotechnology, Yonsei University, 50 Yonsei-ro, Seodaemun-gu, Seoul 120–749,

Republic of Korea e

Division of Bioengineering, College of Life Sciences and Bioengineering, Incheon National

University, 119 Academy-ro, Yeonsu-gu, Incheon 22012, Republic of Korea

1

*Corresponding author Email: [email protected] Graphical Abstract

Highlights   

A method for generating polydopamine (PD) gradient on biomaterial has been reported. PD gradient substrates showed gradient in surface chemistry and hydrophilicity. PD gradient substrates were able to modulate the spatial distribution of biomolecules.

ABSTRACT

Biomaterials with graded functionality have various applications in cell and tissue engineering. In this study, by controlling oxidative polymerization of dopamine, we demonstrated universal techniques for generating chemical gradients on various materials with adaptability for secondary molecule immobilization. Diffusion-controlled oxygen supply was successfully exploited for coating of polydopamine (PD) in a gradient manner on different materials, regardless of their surface chemistry, which resulted in gradient in hydrophilicity and surface roughness. The PD gradient controlled graded adhesion and spreading of human mesenchymal stem cells (hMSCs) and endothelial cells. Furthermore, the PD gradient on these surfaces served as a template to allow for graded immobilization of different secondary biomolecules such as cell adhesive arginineglycine-aspartate (RGD) peptides and siRNA lipidoid nanoparticles (sLNP) complex, for site-

2

specific adhesion of human mesenchymal stem cells, and silencing of green fluorescent protein (GFP) expression on GFP-HeLa cells, respectively. In addition, the same approach was adapted for generation of nanofibers with surface in graded biomineralization under simulated body fluid (SBF). Collectively, oxygen-dependent generation of PD gradient on biomaterial substrates can serve as a simple and versatile platform that can be used for various applications realizing in vivo tissue regeneration and in vitro high-throughput screening of biomaterials. KeyWords: gradient systems; tissue engineering; chemical modification; polydopamine; biomolecule gradient. 1.

Introduction

There is currently a great interest in graded presentation of biomolecules on a material surface for regulating cell functions both in vitro and in vivo or to screen and optimize cell-biomolecule interactions.1-3 Several recent reports have implemented biomolecule gradient platform to study biological process such as cell migration, stem cell differentiation, angiogenesis, or spatially defined cell adhesion and spreading.4-7 For instance, a linear cyclic arginine-glycine-aspartate (RGDfK) peptide gradient was formed to analyze the role of ECM concentration in cell polarization and migration.8 Different aspects of stem cell responses were screened on an amine functional gradient platform and optimal surface chemistry for retention of pluripotency were identified.9 A nanofiber scaffold with hydroxyapatite gradient or three-dimensional scaffolds with a gradient distribution of osteogenic gene were also reported for engineering bone-soft tissue interfacial site.10, 11 Presently, most technologies for the graded biomolecule immobilization are based on the surface modification of substrates with certain functional groups that have specific interactions with the biomolecules.12-14 Some representative systems involve the use of self-assembled monolayers

3

(SAMs) or plasma-based approaches and, thereby, demonstrating the fabrication of tunable surfaces chemistries and the secondary attachment of various biomolecules. For example, gradients in the densities of amine or carboxyl groups have been prepared via controlled plasma treatment using a covered mask, or by moving the sample stage to control the differentiation of stem cells or the growth of rat neuronal progenitor cells (PC-12).15-17 Similarly, gradient surfaces varying from high hydroxyl group density to high aldehyde group density were prepared through diffusion-controlled plasma polymerization for the gradient immobilization of nerve growth factor (NGF).18 Different fabrication techniques such as nanolithography, programmed inkjet printing, or controlled UV exposure, have also been combined with SAMs to generate substrates with gradients of laminin or bone morphogenetic protein-2 (BMP-2).19-22 Although the aforementioned technologies have been successful in generating specific gradients for each objective, these processes are mostly time-consuming and rely on relatively complex setups. Most importantly, the gradient formation is often limited to certain substrate chemistries and types of target molecules, which require individualized designs depending on their applications. Until now, there has been no report on a facile strategy through which a gradient of desirable signals for the modulation of cell or tissue function could be adaptable for different materials. Previously, inspired by the strong adhesiveness of marine mussels, it was demonstrated that dopamine could undergo oxidative self-polymerization to synthesize polydopamine (PD) under alkaline conditions; this PD could then be coated onto various materials independent of their surface properties, and subsequently, could facilitate the secondary immobilization of biomolecules, genes, and minerals.23-26 This polymerization is controlled by various parameters such as pH, dopamine concentration, temperature, and oxygen concentration.25, 27-29 In this study, we hypothesized that (1) a controlled supply of oxygen could control dopamine polymerization in

4

a spatially controlled manner and (2) the resulting graded PD coating on substrates could then be amenable for generation of a gradient of various biomolecules. Given that, we developed a system in which a gradient in oxygen concentration was created across the depth of the reaction solution and investigated its effect on PD gradient formation on different substrates including parafilm, polydimethylsiloxane (PDMS), and poly (L-lactic acid) (PLLA) nanofibers. We then investigated the effect of PD gradient platform on the adhesion of stem cells and endothelial cells. In addition, the effect of gradient in PD coating on graded functionalization of various biomaterials surfaces with different biomolecules was also investigated. Specifically, a cell-instructive peptide, and siRNAs were immobilized in a gradient manner, which were confirmed to control cell functions in a gradient manner. Finally, we also investigated the potential for fabricating a triple-sandwich gradient on a nanofiber surface to regulate spatially controlled biomineralization. 2. Experimental Section 2.1 Materials: Dulbecco’s modified Eagle’s medium (DMEM) was purchased from Gibco (Grand Island, NY, USA), 3,4-dihydroxyphenethylamine from Sigma (St. Louis, MO, USA), fetal bovine serum (FBS) from United Search Partners (Austin, TX, USA), and Tris–HCl from Shelton Scientific, Inc. (Peosta, IA, USA). The fluorescent probes, rhodamine–phalloidin, Hoechst 33258, and Alexa Fluor 488 rabbit anti-mouse IgG, were obtained from Molecular Probes (Eugene, OR, USA). A Milli-Q Plus System (Millipore, Billerica, MA, USA) was used to produce ultrapure water. PLLA was purchased from Boehringer Ingelheim GmbH (Resomer L214S, Essen, Germany; 5.7–8.5 dl/g viscosity), and 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) was from Wako (Osaka, Japan). All other chemicals and solvents were of analytical grade and were used without further purification.

5

2.2 Oxygen tension measurement in dopamine reaction solution: We monitored dissolved oxygen (DO) levels throughout the dopamine solutions using commercially available oxygen sensors (Microx-4 Portable Oxygen Meters, Presens, Regensburg, Germany), as previously described.30, 31

We prepared 4 ml of dopamine solution (2 mg/mL in pH 8.5 Tris-buffer) in 10-mL glass vials

for DO measurements. To monitor DO levels throughout the solutions, we measured DO levels at different depths using implantable microsensors. We chose three different depth points: the airwater interface (top), middle (5 mm), and bottom, which was 10 mm deep to the interface. 2.3 PD gradient generation on different substrates: To generate PD gradients, samples were fixed to a plastic square dish using magnets, as shown in the provided schematic. Square dishes were then placed over a PDMS triangular block fabricated in-house, which had a slope of 10°. Dopamine solutions (in Tris-HCL, 2 mg/ml, and pH 8.5) were then carefully poured into each square dish, and the dishes were incubated for 4 h at room temperature. After the 4-h incubation, samples were washed with distilled water for 2 h with gentle shaking. The water-air interface point was designated as the 0-mm point, and from there, we cut 2 mm × 4 mm samples every 5 mm over a length of 20 mm for different analyses. For the secondary immobilization studies, we extended the gradient length to 30 mm to allow for convenient analysis, and the analyzed points were spaced every 7.5 mm. 2.4 PD gradient characterization: Quantification of PD amount coated on the films was performed using a modified micro BCA assay. In brief, samples from each point were cut into 2 mm × 4 mm sections and incubated with 200 µl of micro BCA working solution (Pierce, IL, USA) for 2 h at 37 °C. The amount of working solution that reacted with the PD-coated substrate was detected at 562 nm using a SpectraMax M2 microplate reader (Molecular Devices, CA, USA). Substrate hydrophobicity at each point was analyzed by imaging the water contact angle on the substrate

6

(Phoenix 300, Surface Electro Optics Co. Ltd., Suwon, Korea) and analyzed using Image J (National Institutes of Health, Bethesda, MD, USA). Optical images of water droplets on the parafilm were photographed 5 min after dropping 5 µl of colored dye onto each film. Surface roughness was characterized by an atomic force microscope (XE-100, Park Systems, Suwon, Korea). Surface atomic composition of PD-parafilm gradient was analyzed with X-ray photoelectron spectroscopy (XPS) (Theta Probe base system; Thermo Fisher Scientific). 2.5 Cell cultures on PD gradient substrates: hMSCs and HUVECs were purchased from Lonza Group Ltd. (Basel, Switzerland). hMSCs were maintained as a monolayer in low-glucose DMEM with 10% FBS and 1% penicillin-streptomycin at 37 °C, 95% air, and 5% CO2 in a humidified environment. For the cell adhesion analysis, hMSCs (passage number 5) were enzymatically removed from culture flasks using 0.125% trypsin-ethylenediaminetetraacetic acid. HUVECs were maintained in endothelial growth medium (EGM-2 BulletKit, Lonza Group Ltd.) and they were trypsinized at passage number 7 for cell adhesion studies on PD gradient substrates. For cell culturing, PD gradient samples were fixed to plastic Petri dishes using sterile magnets. All samples were sterilized with 70% ethanol and by UV exposure. hMSCs and endothelial cells were seeded at a concentration of 20,000 cells/cm2. After 12 h, samples were stained with calcein reagent at a ratio of 1:10,000 for 15 min, and images were taken using a fluorescent microscope. The number of cells and cell spreading at each position were analyzed using ImageJ (five images from each position) and are presented as averages. For paxillin staining, samples were fixed using 4% paraformaldehyde and immunostained using anti-paxillin (1:200, BD Biosciences, Franklin Lakes, NJ, USA) for 1 h at 37 °C. Samples were subsequently stained with a 1:100 dilution of Alexa Fluor 488 rabbit anti-mouse IgG, 1:200 rhodamine–phalloidin, and 1:5000 Hoechst 33258 for 1 h at 37

7

°C. Images were taken at 40 × magnifications using a confocal microscope (TE2000 and Eclipse C1, Nikon, Tokyo, Japan). 2.6 RGD gradient generation on PD gradient substrates: Initially, a thin PDMS film was fabricated by mixing silicone elastomer base and curing agent in a 10:1 ratio. For RGD gradient fabrication, a PD gradient was initially generated on the PDMS film as previously mentioned and was then incubated with RGD peptide at a concentration of 1 mg/ml in Tris-HCl solution for 12 h at 37 °C. For RGD quantification at each point, we cut PD-PDMS gradient samples from each point (2 × 4 mm) and incubated each with 200 μl of RGD solution for 12 h. After the immobilization process, the supernatant was collected and reacted with fluorescamine solution (100 μg/ml in acetone) in a 96-well plate (RGD solution = 75 μl, fluorescamine solution = 25 μl). The reaction was performed for 1 min at room temperature. The fluorescent intensity of each sample was measured (480 nm of excitation, 520 nm of emission) using a plate reader (SpectraMax M2, Molecular Devices). The amount of RGD in 200 μl of the original solution was also measured, and the immobilized amount of RGD on the PD gradient was calculated as the difference between the amount of RGD in the original solution and that in the supernatants. 2.7 PLLA nanofiber fabrication: PLLA nanofibers were fabricated using an electrospinning technique. Polymer solution was prepared by dissolving PLLA in HFIP solution (3%). Polymer solution was then ejected through a 23G blunt-end needle at a speed of 2 ml/h. The polymer stream was directed towards an aluminum foil-covered rotating metal collector (200 rpm) that was supplied with a 13 kV power supply. The fabricated fibers were dried overnight at room temperature and used for further experiments.

8

2.8 sLNP complex gradient immobilization: First, lipidoid (ND98) was synthesized as previously described.32 To prepare sLNP complexes, ND98 stock solution was diluted to 40 ng/µl with 25 mM sodium acetate buffer (pH 5.2, Sigma-Aldrich). Then, ND98 solution was added to a diluted siRNA solution (4 ng/µl; Dy647-siRNA, GE Dharmacon, Lafayette, CO, USA; siGFP, ST Pharm, Seoul, Korea) at equal volume ratios (a siRNA:lipidoid weight ratio of 1:10). sLNP complexes were spontaneously formed via ionic interaction between the negatively charged siRNA and positively charged lipidoid after incubation for 20 minutes at room temperature.33 The sLNP complexes thus formed were loaded onto PD-gradient PLLA nanofibers and incubated for 2 h at room temperature. The nanofibers were then washed with 25 mM sodium acetate buffer (pH 5.2) three times. The sLNP complexes immobilized on the nanofibers in a PD-gradient manner were detected by visualization with fluorescent dye (Dy647)-labeled sLNP complexes using a confocal microscope (LSM 880, Carl Zeiss, Oberkochen, Germany). 2.9 GFP-HeLa cell culture and gradient silencing of GFP expression: GFP-HeLa cells were kindly provided by Prof. Hyukjin Lee, College of Pharmacy, Ewha Womans University, and were cultured in high-glucose DMEM (Gibco BRL, Gaithersburg, MD, USA) supplemented with 10% (v/v) FBS (Gibco BRL) and 1% (v/v) penicillin-streptomycin (Gibco BRL). For reverse transfection by siGFP-LNP complexes loaded on PD-gradient PLLA nanofibers, GFP-HeLa cells were seeded on the nanofibers at a density of 2.5 × 104 cells/cm2. After two days, cells were fixed with 4% paraformaldehyde (Sigma-Aldrich) and then nuclei were stained with 4',6-diamidino-2phenylindole (Sigma-Aldrich). Down regulated GFP expression in the cells was observed using a confocal microscope (LSM 880), and the proportion of GFP-positive cells in the total cell population was calculated as the percentage ratio of the number of GFP-positive cells to the total cell number, based on the images obtained from fluorescence observation.

9

2.10 Mineral gradient fabrication on the PD-PLLA gradient: For mineral gradient fabrication, we initially generated a PD gradient on PLLA nanofibers for 1 h and then incubated them with gelatin solution (1 mg/ml in Tris-HCl) for 30 min. To remove physically attached gelatin from the nanofibers, we washed the gelatin nanofibers in warm distilled water for 2 h with continuous stirring at 100 rpm. Samples were then air dried and incubated with 10× simulated body fluid (SBF) solution for 2 h. Mineralized nanofibers were then washed again in distilled water for 1 h and dried at 37 °C. The mineralized gelatin-PD-PLLA nanofiber morphology was observed using a field emission scanning electron microscope (FE-SEM; JSM-7600F, JEOL) as a function of the position on the nanofibers. Mineral content was calculated using an EDX spectrometer attached to a FE-SEM. Sample structural characterization was investigated using an X-ray diffractometer (New D8-Advanced, Bruker, Billerica, MA, USA). 2.11 Statistics All the regression analysis was performed by Sigma plot regression wizard (Systat Software Inc, Erkrath, Germany) using linear regression curve. The error bars are denoted as +/- SD in the manuscript. 3. Results and discussion 3.1. Analysis of oxygen concentrations during dopamine polymerization Figure 1a depicts the experimental setup in which the substrate of interest was modified with a PD surface gradient. We hypothesized that (1) dopamine polymerization would rapidly deplete oxygen in the solution, and further polymerization would be restricted by oxygen diffusion, and therefore, (2) simple immersion of the substrate at a defined tilting angle into the dopamine

10

solution could modulate dopamine polymerization in a gradient dependent manner. As shown in Figure 1b, in the absence of dopamine, there were no differences in dissolved oxygen concentration in the top and bottom positions. However, in the presence of dopamine, the partial pressure of oxygen levels in a buffer solution at different depths (0, 5, and 10 mm from the airwater interface) showed a sudden reduction within 30 min, indicating the depletion of dissolved oxygen via initial dopamine polymerization. After 30 min, the three depth points exhibited different oxygen levels; in particular, the top position, nearest to the air-water interface, maintained higher oxygen levels (three or four times higher) than the other positions (middle and bottom) (Figure 1c). These results indicate that oxygen concentrations in the dopamine reaction solution were spatially regulated by initial oxygen consumption during dopamine polymerization and further by oxygen diffusion from ambient air. Among the different parameters that affect dopamine polymerization, oxygen concentration has been found to be very critical, since without the presence of oxygen, no evidence of polymerization was observed under various pH levels and incubation times.34 In addition, oxygen removal by purging with argon significantly altered dopamine reaction kinetics.23 The proposed dopamine polymerization reaction involves dopamine oxidation to dopamine-quinone with several intermediate structural changes leading to a final PD layer. All of these intermediate reactions are mediated by oxidation, intra-molecular cyclization, and self-rearrangement into the final product.34, 35 Therefore, we anticipated that a graded oxygen supply would result in controlled dopamine polymerization corresponding to solution depth. 3.2. PD gradient characterization We then investigated whether the above oxygen gradient in a dopamine solution could generate PD gradient substrates. We initially generated a PD gradient on parafilm as a model substrate, since it is a very hydrophobic and easily accessible material in all laboratories. Recently,

11

a PD coating was successfully employed to modify parafilm, which showed improved hydrophilicity and protein immobilization efficiency.36 After a 4-h incubation in the solution at a 10° tilt angle, the brown color intensity along the sample became faded out in proportion to oxygen concentration. In addition, the spreading of water droplets on the film between the air-water interface (denoted as the 0-mm position) and the bottom (20-mm position) changed significantly (Figure 2a). Consistent with this observation, a modified micro bicinchoninic acid (BCA) assay revealed that different PD amounts were deposited at different positions (relative value: top [100%] → middle [57.6 ± 6.3%], bottom [46.8 ± 6.5%]) (Figure 2b). Furthermore, the water contact angle showed a steady decrease in hydrophilicity from the bottom position (95.3 ± 3.2°) to the top position (65.2 ± 2.2°) (Figure 2c). Characterizations of surface roughness revealed increased surface roughness at the 0 mm position as compared to the 10 mm and 20 mm positions in the film with the linear gradient, suggesting that increased polymerization occurred at positions close to the interface (0 mm) compared to the lower (10 and 20 mm) positions (Figure S1a, and b). The surface chemistry of PD gradient was analyzed using X-ray photoelectron spectroscopy (XPS). Figure S1c shows the wide spectrum of samples where a peak corresponds to the nitrogen peak, observed at 399 eV for both 0 and 20 mm position. High resolution nitrogen peaks showed an increase in peak intensity at 0 mm position compared to 20 mm position (Figure S1d). Furthermore, N/C molar ratio was higher at 0 mm position (0.077) as compared to that at the 20 mm position (0.051), suggesting higher PD deposition at 0 mm position than the later (Figure S1e). In addition to parafilm, PD gradients were also generated on other substrates such as PDMS and PLLA nanofibers with the same procedure (Figure 2d and 2e) showing that PD gradient strategy could be easily replicated on other substrates. The relative amounts of PD coating on PDMS, as compared to that at the 0-mm position (i.e., 100%) were significantly lower at the 10-

12

mm (55.12 ± 7%) and 20-mm (40.59 ± 9%) positions (Figure 2d). Water contact angle measurements at each position also confirmed the successful generation of a PD gradient on the PDMS surface (Figure 2e). PD gradient generation on PLLA nanofibers was also achieved and which was confirmed by micro BCA analysis and water contact angle measurement. To demonstrate the versatility of this approach, a symmetrical PD gradient was fabricated on parafilm by folding it in a “V” shape at a defined angle (Figure S1f). As shown in Figure S1g, the quantified amounts of PD at the air-water interface showed the highest amounts of PD deposition (5.46 ± 0.1 and 5.42 ± 0.09 µg/cm2) at the edges of the parafilm (i.e., -15.0 mm and +15.0 mm) as compared to the middle position (0.0 mm → 4.71 ± 0.10 µg/cm2). A color change to dark brown serves as the primary evidence for dopamine polymerization. Hence, a gradual decrease in brown color intensity from the air-water interface to the bottom indicated a gradient PD coating on the parafilm surface. Previous studies have reported improved hydrophilicity on PD-coated substrates, which were attributed to phenolic hydroxyl groups and primary amines present in the PD structure.23,

25

In general, hydrophilic substrates favor cell

adhesion more than their hydrophobic counterparts.37 Hence, a PD gradient substrate with a wettability gradient may be able to regulate spatial cell adhesion. A change in surface roughness is one of the characteristics of PD coating caused by aggregation of dopamine particulates, and thus, a gradual change in surface roughness was evident on the parafilm, possibly due to the PD gradient generated on the surface.38 Previous studies have reported the presence of a nitrogen peak when the substrates are modified with polydopamine.39 Hence, a slight increase in the nitrogen peaks at 0 mm suggests the higher deposition of PD compared to 20 mm position. These results demonstrated that a PD gradient can easily be generated by positioning a substrate at a tilt angle, resulting in different oxygen levels during polymerization. In addition, a symmetrical PD gradient

13

could also be generated on foldable substrates, suggesting that this PD gradient fabrication technique is not only adaptable to linear gradient fabrication, but also to the generation of gradients from both sides. 3.3. Controlled cell adhesion To demonstrate PD gradient substrate capacity to induce cell adhesion, the PD-Parafilm gradient was incubated overnight with a suspension of human mesenchymal stem cells (hMSCs), and calcein staining was performed. The number of cells at the 0-mm position (34 ± 3) was significantly greater than at the 20-mm position (17.2 ± 1.5) with a linear decrease to the bottom (Figure 3a and 3c). Moreover, the adhered cells showed a well-spread cell morphology with welldeveloped actin filaments and focal adhesion points at the 0-mm position (2321.7 ± 698 µm/cm2), which gradually decreased toward the 20-mm position (652.9 ± 97 µm/cm2) (Figure 3b and 3d). It has been reported that differences in stem cell spreading area and focal adhesion formation are critical to control cell functions.40 Not surprisingly, hMSCs in the media lacking serum showed no gradient formation; the average cell number at the 0-mm position was 6.3 ± 1.5, which showed no significant difference from the cell numbers at the 10-mm (5.7 ± 0.9) and 20-mm (6.2 ± 2.06) positions (Figure 3c). These results are in good agreement with a previous report that cell adhesion on a PD-coated substrate is regulated by the interaction between the cells and adsorbed serum proteins.41 Similar to our results with hMSCs, human umbilical vein endothelial cells (HUVECs) also showed a gradient in cell adhesion and spreading thereby demonstrating the adaptability of our platform to multiple cell types (Figure 4a). The adhered cell number at 0 mm position was 71.6 ± 5.6, gradually reduced to the 10 mm region (38.7 ± 6.2) and lowest cell number was observed at

14

20 mm position (28.6 ± 2.5) (Figure 4b). The spreadability of HUVECs was also influenced by the PD gradient as 0 mm showed highest spreading morphology than lower positions (Figure 4c). A gradient distribution of cells is a common feature of physiological systems ranging from embryonic development to soft-hard tissue interfaces. Studies have also shown that differences in cell densities regulate the physical interactions of a single cell (cell-cell communication), which eventually affect multiple cell functions, including migration, proliferation, and differentiation, by influencing intercellular signaling pathways.42, 43 Hence, a gradient cell density platform can be used as a high-throughput in vitro assay tool for mimicking or modeling various cell and tissue functions. In addition, numerous studies have attempted to generate a cellular gradient platform, but many of them have relied on complex procedures with multiple steps. For example, an endothelial cell gradient was achieved on laminin gradient gold-coated substrate by generating an alkanethiol gradient.21 An MC3T3-E1 cell density gradient was formed by immobilizing varying densities of fibronectin onto protein-repellent poly(2-hydroxyethyl methacrylate) assemblies with gradually varying polymer surface coverage.44 However, our system is a stable one-step platform for generating cellular gradients of multiple cells that bypasses complex chemical processes. 3.4. RGD peptide gradient generation The spatially defined distribution of biological cues on a material substrate can be used to control the in vitro response of cells to in vivo tissue regeneration.11, 45 In this context, we tested the ability of the PD gradient to allow formation of a cell-interactive RGD peptide gradient containing a cysteine residue as the end amino acid (GGGRGDS). To prove the diverse utility of our system, we used a thin PDMS film as the substrate. The quantification of the immobilized RGD peptides from each gradient point showed the RGD gradient immobilization on the PDPDMS substrate (Figure 5b). RGD concentration, normalized to the 0 mm position, which showed

15

the highest concentration value, decreased gradually towards the 30 mm position (24%). When we fitted the values with a linear regression curve, the r2 value was 0.9377, showing the high linearity of the RGD gradient generated on the PD-PDMS gradient surface. To further demonstrate the RGD gradient on PD-PDMS substrate, we performed XPS analysis at three different positions of the gradient substrate (0 mm, 15 mm and 30 mm). Wide spectrum of samples indicates the presence of the nitrogen peak at 399 eV for 0 mm and 15 mm positions, and the peak intensity was not visible at 30 mm position (Figure S2a). High resolution nitrogen peaks showed the presence of nitrogen at all gradient positions, however, compared to 30 mm positon, peak intensity was higher at 15 mm position and 0 mm position showed the highest peak intensity (Figure S2b). It is possible that at 0 mm position where the PD coating was maximum, signals from primary amine of PD coating and amide bonding from RGD immobilization could be maximum, which may have collectively showed the maximum peak intensity than other two positions. Quantitative analysis of N/C atomic ratio showed the highest value of 0.16 at 0 mm position, which was reduced toward 15 mm (0.03) and 30 mm position (0.01) (Figure S2c). All these results suggest that RGD can be immobilized on PD-PDMS substrate in a gradient manner. The PD-RGD-PDMS gradient was also able to achieve hMSCs adhesion in a gradient manner when they were cultured under serum-free conditions (Figure 5a and 5c). The average cell number at the 0-mm position was 58 ± 5, and this was gradually reduced over a distance of 30 mm (34 ± 1 at the 30-mm position). It is worth noting that cells were seeded in the absence of serum in the media, and hence, the initial cell/material contact occurred only through the interactions between the immobilized RGD molecules and hMSCs. PD coatings have been used for bridging different biomolecules through covalent interactions between the catechol and amine or thiol groups present in the biomolecules.24, 46 Various substrates ranging from polymers to metals have been functionalized with different

16

biomolecules such as RGD peptide, fibronectin, laminin, or different growth factors through the deposited PD layer.25, 47, 48 The gradient distribution of biomolecules resembled the concentration gradient of biomolecules created during tissue regeneration. A number of studies have reported biomaterials with biomolecule gradients used to analyze different cellular events. For example, basic fibroblast growth factor concentration gradients generated on a glass slide or in microfluidic devices have been used to study the migration behavior of vascular smooth muscle or endothelial cells.49, 50 Similarly, an RGD concentration gradient substrate was used to analyze the influence of RGD concentration on stem cell fate determination and to study cell adhesion and focal contact formation of NIH/3T3 mouse embryonic fibroblasts.51, 52 Given that, our PD gradient platform will be amenable to gradient generation for various biomolecules including different growth factors, peptides, genes and among others. 3.5. siRNA gradient generation To confirm the gradual localization of siRNA-lipidoid nanoparticle (sLNP) complexes on a PD-PLLA nanofiber with PD concentration gradients, Dy647-conjugated siRNA was used to form fluorescently labeled sLNP complexes. Recently, we and our colleagues demonstrated that PD coating facilitated capturing sLNP complexes by covalently interacting with sLNP amine groups, providing a highly efficient siRNA reverse transfection platform.53 As expected, larger numbers of Dy647-labeled sLNP complexes were detected on the sites with higher concentrations of deposited PD (0 mm), and the numbers gradually decreased in areas with lower concentrations of deposited PD (20 mm) (Figure 6a and 6d). Then, siRNA transfection efficiency on the PDsLNP-PLLA gradient nanofibers was evaluated by delivering green fluorescence protein (GFP)-

17

targeting siRNA (siGFP) into GFP-overexpressing HeLa cells (GFP-HeLa). A full-length fluorescence image of HeLa cells cultured on the PD-gradient platform demonstrated spatially controlled GFP expression of the GFP-HeLa cells (Figure 6b). Specifically, the 0-mm position, where the deposited PD and immobilized sLNP complex was higher, showed lower numbers of GFP-overexpressing HeLa cells, and their numbers gradually increased toward the 20-mm position, which showed the lowest number of sLNP complexes. Quantitative analysis of GFPexpressing cells on the PD-sLNP-PLLA gradient nanofibers also confirmed the gradient genetic modification through siRNA delivery, as the 0-mm position showed 26.2 ± 9.2%, the 10-mm position showed 57.7 ± 6.6%, and the 20-mm position showed 77.5 ± 10.8% of GFP-positive cells (Figures 6c and 6e). Our results suggest that the PD gradient generated an immobilized sLNP complex gradient, and thereby induced spatially controlled specific gene expression silencing of cells grown on the substrate in a gradient manner. The ability to specifically silence genes using RNA interference has recently been applied to stem cell manipulation and tissue engineering.54, 55 When cells are transfected with siRNA, they can silence the expression of specific proteins by base pairing with their mRNA sequences. 56 Previous studies have reported that siRNA transfection into stem cells can control their commitment toward different lineages including osteogenic, chondrogenic, adipogenic, myogenic, or neuronal.57-60 Thus, gradient genetic modification of cells through siRNA transfection could provide an efficient strategy for guiding the gradual transition of cell fates and their behaviors. In addition, substrate-mediated reverse transfection has several advantages, such as spatially controllable transfection, higher probability of cell-siRNA contact, prolonged transfection, and applicability to implantable systems for therapeutic approaches, as compared to conventional forward transfection or a solution-mediated transfection system.53,

61

Collectively, our results

18

suggest that the PD-gradient system could contribute to establishing gradient reverse transfection platforms with sLNP gradients targeting different genes associated with cell differentiation to be used in different tissue engineering applications including interfacial tissue regeneration. 3.6. Preparation of gradient in biomineralization We next designed a triple-layered for preparation of gradient in biomineralization strategy by the controlled gelatin immobilization on a PD gradient applied to PLLA nanofibers. This served as a template for biomineralization (Figure 7a). Initially, the gelatin gradient was formed on PDPLLA nanofibers by exploring the interactions between catecholamines and the functional molecules present in the gelatin (amine or carboxyl groups) (Figure 7b). The quantified amount of gelatin at the 0-mm position was 9.3 ± 1.3 µg/cm2, and a decrease in the immobilized amount was observed over a distance of 30 mm (2.64 ± 1.5 µg/cm2) at the 30-mm position). Gelatin has previously been used for the biomimetic mineralization of different substrates where electrostatic interactions between carboxyl groups and calcium ions play a key role in mineralization.10 The diffraction patterns at all positions showed apparent peaks at 25.9°, 31.7°, 32.2°, and 32.9° corresponding to the (200), (211), (112), and (300) main reflection planes, which were identified as hydroxyapatite (JCPDS No. 09-0432). Peak intensity at the 0-mm position was much stronger than the intensity at other positions and decreased with position, indicating that the mineralization gradient was a function of position (Figure 7c). Scanning electron microscopy (SEM) images showed densely coated mineral structures at the 0-mm position and their numbers decreased as a function of distance from 0 mm. At the 30-mm position, very limited and not well formed mineral coating was observed (Figure 7e). Quantitative analyses from electron dispersive X-ray spectroscopy (EDX) also confirmed the mineralization gradient on the nanofiber surface (Figure 7d). A higher calcium content was found at the 0-mm position (10.5%) compared to the 15-mm

19

(4.4%) and 30-mm (1.5%) positions. Hence, our results demonstrate the controlled sequential immobilization of multiple factors for mineral gradient fabrication, all of which were regulated by the PD-gradient substrate, thereby establishing PD gradient platform adaptability. Interfacial tissues, such as ligament-to-bone, tendon-to-bone, and cartilage-to-bone, have graded calcified regions with hydroxyapatite as a major mineral component within the collagen matrix. No mineral content is found in the soft tissues, which gradually become mineralized toward the hard bone tissue regions.62 Attempts have also been made to mimic the hierarchical mineral content distribution to engineer these complex tissues in vitro.10, 63 Previous studies that have focused on mineral gradient fabrication on different materials mainly employed the syringe infusion pump method, where vials which contain the targeted substrates were filled at constant rates to create a linear gradient during incubation time.10, 63-65 While these approaches are highly efficient and reproducible, they often require initial surface modifications such as plasma treatment or specialized setups for gradient mineralization. Our approach is simple and applicable to a variety of materials, just requiring multiple dip coatings, which are relatively less laborious and more convenient when compared to previously reported approaches. 4. Conclusions In conclusion, a simple yet versatile method for generating a surface property gradient has been developed using spatially controlled dopamine polymerization. Our studies suggest that differences in diffused oxygen concentrations in a reaction solution across different depth scales could influence dopamine polymerization and additional surface coating with PD. Moreover, the PD gradient generated on substrates could be used to control adhesion of multiple cell types and immobilization of RGD peptide, siRNA, and biominerals. This method can be used as a simple,

20

universal method for generating a surface chemical gradient on various materials as a PD coating has already been proven to be a universal surface modification strategy. The most exciting feature of this strategy is that the PD layers contain reactive functional groups that allow for secondary immobilization with various molecules; hence, we anticipate that this method should have many promising applications in the fields of cell and tissue engineering. AUTHOR INFORMATION Corresponding Author * Email: [email protected] Author Contributions The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. ‡These authors contributed equally. Notes The authors declare no competing financial interest. ACKNOWLEDGMENT This work was supported by a National Research Foundation of Korea grant funded by the Korean government (MEST) (NRF-2016R1A2B3009936). Supporting Information Characterizations of PD gradient on parafilm such as AFM and XPS analysis, Characterization of a symmetrical PD gradient generation on parafilm, XPS analysis of PD-RGD-PDMS gradient are provided in the supporting information.

21

REFERENCES (1) Seidi, A.; Ramalingam, M.; Elloumi-Hannachi, I.; Ostrovidov, S.; Khademhosseini, A. Acta biomaterialia. 2011, 7, 1441-1451. (2) Michelmore, A.; Clements, L.; Steele, D. A.; Voelcker, N. H.; Szili, E. J. Journal of Nanomaterials. 2012, 2012, 8. (3) Wu, J.; Mao, Z.; Tan, H.; Han, L.; Ren, T.; Gao, C. Interface Focus. 2012, 6, 337-355. (4) Yu, S.; Gao, Y.; Mei, X.; Ren, T.; Liang, S.; Mao, Z.; Gao, C. ACS Applied Materials & Interfaces. 2016, 8, 29280-29288. (5) Torres, A. B. F.; Goren, T.; Ihalainen, T. O.; Guimond-Lischer, S.; Charnley, M.; Rottmar, M.; Maniura-Weber, K.; Spencer, N. D.; Reis, R. L.; Textor, M. ACS Applied Materials and Interfaces. 2015, 7, 2367-2375. (6) Barkefors, I.; Thorslund, S.; Nikolajeff, F.; Kreuger, J. Lab on a Chip. 2009, 9, 529-535. (7) Lamb, B. M.; Park, S.; Yousaf, M. N. Langmuir. 2010, 26, 12817-12823. (8) Hirschfeld-Warneken, V. C.; Arnold, M.; Cavalcanti-Adam, A.; López-García, M.; Kessler, H.; Spatz, J. P. European journal of cell biology. 2008, 87, 743-750. (9) Ghaemi, S. R.; Harding, F. J.; Delalat, B.; Gronthos, S.; Voelcker, N. H. Biomaterials. 2013, 34, 7601-7615. (10) Li, X.; Xie, J.; Lipner, J.; Yuan, X.; Thomopoulos, S.; Xia, Y. Nano letters. 2009, 9, 27632768. (11) Phillips, J. E.; Burns, K. L.; Le Doux, J. M.; Guldberg, R. E.; García, A. J. Proceedings of the National Academy of Sciences. 2008, 105, 12170-12175. (12) Chen, X.; Su, Y.-D.; Ajeti, V.; Chen, S.-J.; Campagnola, P. J. Cellular and molecular bioengineering. 2012, 5, 307-319. (13) Jeon, O.; Alt, D. S.; Linderman, S. W.; Alsberg, E. Advanced Materials. 2013, 25, 63666372. (14) Zhu, Y.; Mao, Z.; Gao, C. Biomacromolecules. 2013, 14, 342-349. (15) Mangindaan, D.; Kuo, W.-H.; Wang, M.-J. Biochemical engineering journal. 2013, 78, 198204. (16) Lee, S. J.; Khang, G.; Lee, Y. M.; Lee, H. B. Journal of colloid and interface science. 2003, 259, 228-235. (17) Zelzer, M.; Majani, R.; Bradley, J. W.; Rose, F. R.; Davies, M. C.; Alexander, M. R. Biomaterials. 2008, 29, 172-184. (18) Delalat, B.; Mierczynska, A.; Ghaemi, S. R.; Cavallaro, A.; Harding, F. J.; Vasilev, K.; Voelcker, N. H. Advanced Functional Materials. 2015, 25, 2737-2744. (19) Schwab, E. H.; Pohl, T. L.; Haraszti, T. s.; Schwaerzer, G. K.; Hiepen, C.; Spatz, J. P.; Knaus, P.; Cavalcanti-Adam, E. A. Nano letters. 2015, 15, 1526-1534. (20) Gallant, N. D.; Lavery, K. A.; Amis, E. J.; Becker, M. L. Advanced materials. 2007, 19, 965969. (21) Cai, K.; Dong, H.; Chen, C.; Yang, L.; Jandt, K. D.; Deng, L. Colloids and Surfaces B: Biointerfaces. 2009, 72, 230-235. (22) Meyyappan, S.; Shadnam, M.; Amirfazli, A. Langmuir. 2008, 24, 2892-2899. (23) Lee, H.; Dellatore, S. M.; Miller, W. M.; Messersmith, P. B. Science. 2007, 318, 426-430. (24) Lee, H.; Rho, J.; Messersmith, P. B. Advanced Materials. 2009, 21, 431-434. (25) Madhurakkat Perikamana, S. K.; Lee, J.; Lee, Y. B.; Shin, Y. M.; Lee, E. J.; Mikos, A. G.; Shin, H. Biomacromolecules. 2015, 16, 2541-2555.

22

(26) Ryu, J.; Ku, S. H.; Lee, H.; Park, C. B. Advanced Functional Materials. 2010, 20, 2132-2139. (27) Ball, V.; Del Frari, D.; Toniazzo, V.; Ruch, D. J Colloid Interface Sci. 2012, 386, 366-72. (28) Zhou, P.; Deng, Y.; Lyu, B.; Zhang, R.; Zhang, H.; Ma, H.; Lyu, Y.; Wei, S. PLoS One. 2014, 9, e113087. (29) Wei, Q.; Zhang, F.; Li, J.; Li, B.; Zhao, C. Polym. Chem. 2010, 1, 1430-1433. (30) Park, K. M.; Gerecht, S. Nature communications. 2014, 5. (31) Park, K. M.; Blatchley, M. R.; Gerecht, S. Macromolecular rapid communications. 2014, 35, 1968-1975. (32) Akinc, A.; Zumbuehl, A.; Goldberg, M.; Leshchiner, E. S.; Busini, V.; Hossain, N.; Bacallado, S. A.; Nguyen, D. N.; Fuller, J.; Alvarez, R. Nature biotechnology. 2008, 26, 561-569. (33) Cho, S. W.; Goldberg, M.; Son, S. M.; Xu, Q.; Yang, F.; Mei, Y.; Bogatyrev, S.; Langer, R.; Anderson, D. G. Advanced functional materials. 2009, 19, 3112-3118. (34) Kim, H. W.; McCloskey, B. D.; Choi, T. H.; Lee, C.; Kim, M.-J.; Freeman, B. D.; Park, H. B. ACS applied materials & interfaces. 2013, 5, 233-238. (35) Liu, Y.; Ai, K.; Lu, L. Chem Rev. 2014. (36) Shi, X.; Li, L.; Ostrovidov, S.; Shu, Y.; Khademhosseini, A.; Wu, H. ACS applied materials & interfaces. 2014, 6, 11915-11923. (37) Webb, K.; Hlady, V.; Tresco, P. A. Journal of biomedical materials research. 1998, 41, 422. (38) Shin, Y. M.; Lee, Y. B.; Shin, H. Colloids and Surfaces B: Biointerfaces. 2011, 87, 79-87. (39) Cho, I.; Gong, S.; Song, D.; Lee, Y.-G.; Ryou, M.-H.; Lee, Y. M. Scientific reports. 2016, 6. (40) Wang, X.; Hu, X.; Dulińska-Molak, I.; Kawazoe, N.; Yang, Y.; Chen, G. Scientific reports. 2016, 6. (41) Ku, S. H.; Ryu, J.; Hong, S. K.; Lee, H.; Park, C. B. Biomaterials. 2010, 31, 2535-2541. (42) Nelson, C. M.; Chen, C. S. Febs Letters. 2002, 514, 238-242. (43) Cichon, M. A.; Nelson, C. M.; Radisky, D. C. Cancer informatics. 2015, 1. (44) Bhat, R. R.; Chaney, B. N.; Rowley, J.; Liebmann‐Vinson, A.; Genzer, J. Advanced Materials. 2005, 17, 2802-2807. (45) Samorezov, J. E.; Alsberg, E. Advanced drug delivery reviews. 2015, 84, 45-67. (46) Lee, Y. B.; Shin, Y. M.; Lee, J.-h.; Jun, I.; Kang, J. K.; Park, J.-C.; Shin, H. Biomaterials. 2012, 33, 8343-8352. (47) Zongguang, L.; Shuxin, Q.; Jie, W. Progress in Chemistry. 2015, 27, 212-219. (48) Chien, C.-Y.; Tsai, W.-B. ACS applied materials & interfaces. 2013, 5, 6975-6983. (49) Wu, J.; Mao, Z.; Han, L.; Zhao, Y.; Xi, J.; Gao, C. Colloids and Surfaces B: Biointerfaces. 2014, 117, 290-295. (50) Barkefors, I.; Le Jan, S.; Jakobsson, L.; Hejll, E.; Carlson, G.; Johansson, H.; Jarvius, J.; Park, J. W.; Jeon, N. L.; Kreuger, J. Journal of Biological Chemistry. 2008, 283, 13905-13912. (51) Smith Callahan, L. A.; Policastro, G. M.; Bernard, S. L.; Childers, E. P.; Boettcher, R.; Becker, M. L. Biomacromolecules. 2013, 14, 3047-3054. (52) Lagunas, A.; Comelles, J.; Martínez, E.; Prats-Alfonso, E.; Acosta, G. A.; Albericio, F.; Samitier, J. Nanomedicine: Nanotechnology, Biology and Medicine. 2012, 8, 432-439. (53) Shin, J.; Cho, J. H.; Jin, Y.; Yang, K.; Lee, J. S.; Park, H. J.; Han, H. S.; Lee, J.; Jeon, H.; Shin, H. Small. 2016. (54) Cheema, S. K.; Chen, E.; Shea, L. D.; Mathur, A. B. Wound repair and regeneration. 2007, 15, 286-295. (55) Yao, Y.; Wang, C.; Varshney, R. R.; Wang, D.-A. Pharmaceutical research. 2009, 26, 263275.

23

(56) Tijsterman, M.; Plasterk, R. H. Cell. 2004, 117, 1-3. (57) Andersen, M. Ø.; Nygaard, J. V.; Burns, J. S.; Raarup, M. K.; Nyengaard, J. R.; Bünger, C.; Besenbacher, F.; Howard, K. A.; Kassem, M.; Kjems, J. Molecular Therapy. 2010, 18, 20182027. (58) Motomura, H.; Niimi, H.; Sugimori, K.; Ohtsuka, T.; Kimura, T.; Kitajima, I. Biochemical and biophysical research communications. 2007, 357, 997-1003. (59) Aziz, A.; Miyake, T.; Engleka, K. A.; Epstein, J. A.; McDermott, J. C. Developmental biology. 2009, 332, 116-130. (60) Yang, Y.; Li, Y.; Lv, Y.; Zhang, S.; Chen, L.; Bai, C.; Nan, X.; Yue, W.; Pei, X. Experimental cell research. 2008, 314, 2257-2265. (61) Rujitanaroj, P.-o.; Wang, Y.-C.; Wang, J.; Chew, S. Y. Biomaterials. 2011, 32, 5915-5923. (62) Moffat, K. L.; Sun, W.-H. S.; Pena, P. E.; Chahine, N. O.; Doty, S. B.; Ateshian, G. A.; Hung, C. T.; Lu, H. H. Proceedings of the National Academy of Sciences. 2008, 105, 7947-7952. (63) Lipner, J.; Liu, W.; Liu, Y.; Boyle, J.; Genin, G.; Xia, Y.; Thomopoulos, S. Journal of the mechanical behavior of biomedical materials. 2014, 40, 59-68. (64) Liu, W.; Lipner, J.; Xie, J.; Manning, C. N.; Thomopoulos, S.; Xia, Y. ACS applied materials & interfaces. 2014, 6, 2842-2849. (65) Lipner, J.; Shen, H.; Cavinatto, L.; Liu, W.; Havlioglu, N.; Xia, Y.; Galatz, L. M.; Thomopoulos, S. Tissue Engineering Part A. 2015, 21, 2766-2774.

24

25

Figure 1. Substrate fabrication with graded polymerized dopamine (PD). The photograph on the top left shows the actual setup for substrate preparation, where a substrate with dopamine solution was placed at a tilting angle on a triangular PDMS block, which created different reaction solution depths. The right schematic repesents the resulting PD gradient substrate and versatile generation of gradient in biomolceules and biomineralization. Oxygen partial pressure in reaction solution in a glass vial across different solution depths when there is a) no dopamine present and b) with dopamine. The top position denotes 0 mm from the air-water interface, and the middle and bottom positions represent 5 mm and 10 mm in depth from the air-water interface, respectively.

26

Figure 2. Characterization of substrates with PD gradients. a) Photographs of parafilm with a PD gradient and water droplets on the same parafilm. The black dotted line indicates the air-water interface, which is designated as position 0. Scale bar is 5 mm. b) Quantitative analysis of b) PD coating measured by micro BCA relative to the value determined at the 0-mm position and c) water contact angles. d) The quantified relative PD coatings at various positions on the PDMS and PLLA film with PD gradients. e) Water contact angles at various positions on the PDMS and PLLA film with PD gradients.

27

Figure 3. Controlled stem cell adhesion on a PD gradient platform. a) Images of hMSCs stained with calcein at different positions on a PD gradient on parafilm (scale bar is 100 µm). b) Confocal images of stem cells at different positions on parafilm with a PD gradient. Well-formed actin and focal adhesion points were observed at the 0-mm position as compared to those at the 10-mm and 20-mm positions. (Scale bars are 100 µm.) c) The number of stem cells on parafilm with a PD gradient and d) the spreading area of stem cells in the presence and absence of serum.

Figure 4. HUVECs cultured on parafilm with a PD gradient. a) Calcein staining of HUVECs present at different positions on the parafilm with PD gradient (the scale bar represents 100 µm). b) The number and c) the spreading area of HUVECs at each position on the graded substrates.

28

Figure 5. RGD peptide gradient generation on a PD gradient platform. a) Hoechst-stained cells show spatially controlled cell adhesion at different positions on the PD-RGD-PDMS gradient (scale bar is 100 µm). b) The percentage of immobilized RGD at different positions on the PDRGD-PDMS gradient. c) Quantified results of stem cell adhesion at different positions on the PDRGD-PDMS gradient.

29

Figure 6. Gradient siRNA reverse transfection on PD-gradient substrate. a) Immobilization of Dy647-labeled sLNP complexes on PD-gradient PLLA nanofibers (scale bar = 200 µm). b) Full-length fluorescence image showing spatially controlled gradient silencing of GFP expression in GFP-HeLa cells on siGFP-LNP-gradient PLLA nanofibers (scale bar = 2 mm). c) Magnified fluorescence images of reverse-transfected GFP-HeLa cells at several positions on the nanofibers

30

(scale bar = 200 µm). d) Quantification of the number of immobilized sLNP complexes at several positions on the surface of the PD-gradient PLLA nanofibers. e) Quantification of GFP-positive cells on the siGFP-LNP gradient PLLA nanofibers.

31

32

Figure 7. Nanofibers with graded mineral deposition on a PD-gelatin gradient generated using a step-by-step approach. a) Schematic of sequential gradient fabrication on PLLA nanofibers used to generate a mineral gradient. b) A gelatin gradient which formed on PLLA nanofibers with a PD gradient was quantified using micro BCA. c) X-ray diffraction showing the representative hydroxyapatite peaks from a PD gradient on PLLA nanofibers, indicating the successful formation of hydroxyapatite. Peak intensity was gradually reduced from the 0-mm position to the 30-mm position, thereby confirming mineral gradient generation on the PLLA nanofibers. d) Energy dispersive X-ray spectra at various positions along PLLA nanofibers with a mineral gradient. e) Representative SEM images from different positions on the biomineral gradient (scale bar represents 1 µm) .

33

Supplementary information

Graded functionalization of biomaterial surfaces using polydopamine chemistry Sajeesh Kumar Madhurakkat Perikamana a,b, ‡, Young Min Shin c, ‡, Jinkyu Lee a,b, Yu Bin Lee a,b

, Yunhoe Heo a,b, Taufiq Ahmad a,b, So Yeon Park d, Jisoo Shin e, Kyung Min Park f, Hyun Suk

Jung d, Seung-Woo Cho e and Heungsoo Shin a,b* *Corresponding author. Email: [email protected]

34

Figure S1. Characterization of substrates with PD gradients. The representative a) AFM images from the 0-mm, 10-mm, and 20-mm parafilm positions with the PD gradient (scale bar is 50 µm). b) Quantified surface roughness at different parafilm positions with a PD gradient. c) XPS spectrum of the PD-parafilm gradient at 0 and 20 mm. d) high resolution spectrum of nitrogen peaks and e) N/C molar ratio at 0 and 20 mm positions of PD-parafilm gradient. f) Photograph

35

showing water droplets on a symmetrical PD gradient on parafilm and e) the quantified amounts of PD at different positions.

Figure S2. XPS analysis of PD-RGD-PDMS gradient. a) XPS spectrum of the PD-RGD-PDMS gradient at 0 mm, 15 mm, and 30 mm. b) high resolution spectrum of nitrogen peaks and c) N/C molar ratio at 0 mm, 15 mm, and 30 mm positions of PD-RGD-PDMS gradient.

36