Heterodimeric Protein Surface‐Coupling Platform: Immobilization of Conformation Switchable and Functional αIIbβ3 Integrin

Many biotechnologies require direct contact between biomolecules and (semi)solid substrates. However, little information is available regarding site‐directed covalent surface immobilization of heterodimeric proteins. Integrins are heterodimeric, conformationally dependent membrane adhesion‐receptors, which are important in the (patho)biology of almost all human diseases. In this study, a biomimetic‐platform is developed for covalent and site‐directed immobilization of oriented integrin heterodimers onto most hydroxyl bearing surfaces, herein demonstrated on both glass‐slides and silica‐particles. This platform consists of a self‐assembled monolayer, upon which C‐terminal modified αIIbβ3 integrin extracellular domains are oriented and covalently anchored. As with cell surface integrins, the conformation of immobilized αIIbβ3 is switchable and can be modulated to the active ligand‐binding conformation by divalent cations. Furthermore, the αIIbβ3‐coupled silica particles display platelet‐mimetic hemostat function and co‐aggregate with platelets from both wild‐type and fibrinogen/von Willebrand factor double deficient mice, facilitating αIIbβ3‐non‐classical ligand discoveries. This work provides a biomaterial platform for functional multimeric protein‐substrate coupling, which should have broad impact on multiple fields of biology, biotechnology, and clinical diagnosis/therapy.


Introduction
[3] A common, versatile, and relatively facile method to couple biomolecules to substrates is the use of organosilane self-assembled monolayers (SAM).[5] They are commonly utilized in applications across many scientific and engineering disciplines such as the development of fouling resistant surfaces, [4][5][6] improved substrate biocompatibility, [6,7] biosensing, [8,9] and cellular signaling and interactions. [10]Literature reports of the covalent immobilization of monomeric proteins and nucleic acids are quite common, however, there are no literature reports of the sitedirected, oriented, covalent coupling of heterodimeric proteins to organic SAMs.This is due to the difficulty of immobilizing two separate, non-identical amino-acid chains in a manner that retains structure and function.As such, many technologies revert to the use of mimetic peptides or antibodies to reproduce complex multimeric protein function.However, multimeric proteins usually participate in multiple biological interactions and mimetics can generally only reproduce one of these interactions, which would therefore limit the functionality of the biotechnology.Furthermore, a mimetic peptide may possibly cause undesired biological effects by interacting with an undesired receptor.
[19][20] All integrins are non-covalent heterodimers consisting of an  and  subunit.[14] Integrin activity is structure dependent and can switch conformations from an inactive bent conformation to an activated high-affinity ligand-binding conformation (Figure 1b). [13,21,22]As a result, applications that necessitate integrin-functionalized substrates should preserve the ligand binding conformation in order to function.
One integrin that has gained considerable research and therapeutic attention is platelet integrin IIb3 (GPIIbIIIa). [17,20,23][26] A key platelet function is the formation of a platelet plug through the aggregation of adjacent platelets mediated by the binding of IIb3 to the distal bivalent AGDV sites of fibrinogen), [27] a platelet aggregation theory that has lasted more than half century.However, unexpectedly, as we reported, even in the absence of fibrinogen, and both fibrinogen and von Willebrand factor (VWF), platelet aggregation persists mediated by IIb3 and unknown crosslinking ligands, termed fibrinogen/VWFindependent platelet aggregation. [28,29]Identifying these nonclassical IIb3 ligands is important for enhancing hemostasis and preventing thrombosis.Furthermore, platelet IIb3 and its ligands are emerging as key mediators in cancer metastasis. [30]herefore, developing a reliable method to study IIb3 and its ligands are crucial not only for the basic sciences but also to control these major human diseases.
Herein, we report a facile to fabricate biomaterial platform for the immobilization of heterodimeric integrins, and other multimeric polypeptides.The platform is composed of a combination of the facile and robust modification of the underlying hydroxyl bearing surface with a N-(3-Trimethoxysilylpropyl) diethylenetriamine (DETA) SAM [31] and the reliable modification of the coupled polypeptide [32] leading to the first report of a sitedirected, oriented, and covalently immobilized integrin onto a SAM.This heterodimeric protein immobilization platform could have a widespread impact and constitutes a significant addition to the advancement of science in the areas of biology, biotechnology, as well as clinical diagnosis and therapy.

Preparation and Characterization of 𝜶IIb𝜷3 Surface Coupling
The immobilization platform was prepared on two different substrates: 5 mm 2 glass slides and ≈1 μm (average, platelet size) diameter ferromagnetic silica particles; following the three-step reaction scheme depicted in Figure 1c.The  and  subunits of IIb3 ectodomain were C-terminal modified to feature a negatively-charged acidic residue rich tail on the 3 subunit and a positively-charged tail followed by a HIS-tag on the IIb subunit to increase purification efficiency and overall heterodimer stability as described by Takagi et.al., [32] a strategy that was further utilized in several structural studies of integrins. [33,34]To confirm the IIb3 heterodimer does assemble in solution, HIS column purified IIb3 ectodomains using the HIS-tag on the IIb subunit were assessed with SDS-PAGE (Figure S2, Supporting Information) and found both subunits at approximately an equal molar ratio.This suggests that both subunits form the IIb3 heterodimer in solution.The integrin complex was covalently coupled to the DETA SAM via the acidic residue rich 3 tail and the distal primary amine of the DETA SAM.Integrin-SAM cross-coupling is performed in a one-step process using EDC/NHS chemistry to activate the carboxylic acid containing sidechains of the acidic residues that are primarily present on the engineered acidic residue rich tail of the 3 subunit.The process is performed in one-step to decrease the undesired competing process of NHS hydrolysis. [35]Once coupled, IIb3 can be activated to the high-affinity binding conformation by incubation with divalent cations (Figure 1b). [21]ray photoelectron spectroscopy (XPS) was used to monitor the characteristic elements of the substrate (O and Si), DETA SAM (Si, C, and N), and IIb3 integrin (C and N) at each step of the fabrication process for both glass slides and ferromagnetic silica substrates.Glass substrate XPS survey spectra and atomic abundance data are presented in Figure 2a-d.Bare glass, as expected for SiO 2 , was found to have an O to Si ratio of ≈2:1.Low levels of unavoidable adventitious carbon and nitrogen contamination were also observed on the bare substrate.Following DETA SAM deposition, both the C and N signals increased with a concurrent decrease in O and Si signals, which is consistent with a C and N rich organosilane atop the SiO 2 glass substrate.Also, if the adventitious C and N signals of the bare substrate are subtracted, the atomic ratio of C and N is consistent with that of DETA.Furthermore, when the Si 2p narrow scans are examined (Figure 2e), the peak binding energy shifts from 102.9 ± 0.2 eV for the bare glass surface to 102.1 ± 0.1 eV for the DETA-coated surface.This binding energy shift corresponds to a transition from a predominantly Si-O 2 -containing bare substrate to a surface that now bears the Si-OH/Si-O-C species as a result of DETA deposition.[36,37] Upon integrin coupling to the DETA SAM, the silicon signal was further attenuated while that of C, O, and N remained relatively unchanged.
Originally, S was intended to be used as a unique element to indicate the presence of IIb3, as it contains multiple disulfide bonds.However, due to the low relative abundance of sulfur no such peak was detected in the XPS spectra.Despite this, the presence of IIb3 is demonstrated by the N 1s (Figure 2f) binding energy peak shift from 399.0 ± 0.2 eV for DETA SAM coated glass, to 400.0 ± 0.3 eV when IIb3 is immobilized.The DETA peak arises due to contributions primarily from the DETA terminal NH 2 and partially from internal NH species and corresponds to previously reported values. [38]The binding energy shift upon the addition of IIb3 is due to positively charged nitrogen species [38] that are commonly found in amino acids; indicating that IIb3 integrin is indeed present.The same conclusions as above can be drawn from the ferromagnetic silica particle XPS data Figure S1 (Supporting Information) (atomic abundance of both substrates is tabulated in Tables S1 and S2, Supporting Information).To further evaluate the efficacy of IIb3 immobilization, the intrinsic fluorescence of the immobilized protein was evaluated.IIb3 coated particles showed a maximum emission of 420 ± 0.5 nm indicating the presence of protein, while DETA do not show a significant fluorescence emission, Figure S2 (Supporting Information).

Evaluation of Immobilized 𝜶IIb𝜷3 Integrin
At each stage of the platform fabrication process (bare, DETA monolayer, pre-activation IIb3 immobilized and activated IIb3 immobilized substrates), both glass and silica substrates were analyzed for interactions with conformation independent anti-3 monoclonal antibody (mAb) PSI-E1 [39] and active-conformation dependent endogenous ligand, fibrinogen [29] (Figure 3a).Glass slides were evaluated using fluorescence confocal microscopy (Figure 3) and flow cytometry was used for analysis of ferromagnetic silica particles (Figure 4).Both pre-activated and activated IIb3 substrates can be   recognized by mAb PSI E1, indicating the presence of the integrin on the surface (Figures 3b,c and 4a,b).Furthermore, 4G6 IIb specific antibody bound to activated IIb3 silica particles (Figure 4c), this in conjunction with the PSI-E1 binding data confirm that both subunits are present on the particle surface.Im-portantly, the DETA SAM was able to markedly resist the nonspecific adsorption of fibrinogen when compared to the bare substrates.Varying concentrations of fluorescently labeled fibrinogen were incubated with DETA and activated IIb3 silica particles, and the mean fluorescence intensity determined using flow cytometry.Dose response curves were plotted (Figure 4f) and fit to a one-site binding model producing a K d apparent of 78 ± 7 nM for surface activated IIb3.Furthermore, only activated IIb3 substrates were able to specifically bind fibrinogen (Figures 3b,d and 4d,e), suggesting that the SAM-coupled integrin can switch to the high-affinity ligand binding conformation.As negative controls, FITC-conjugated anti-GPIb and anti-CD62p antibodies (both antibodies targeting other platelet surface receptors) were exposed to DETA SAM, pre-activated or activated IIb3-immobilized silica particles, and no segnificant non-specific binding was detected (Figure 4k-n).
To further confirm the integrin conformation switch, preactivation IIb3 and activated IIb3 substrates were assessed for binding to PAC-1, a mAb specific for an activation-induced conformational epitope on human IIb3. [40]Substrates were incubated in 50 nM FITC conjugated PAC-1 mAb for 1 h at RT, washed and analyzed.It was revealed that PAC-1 mAb bound to activated substrates and particles (Figures 3b,e and 4h-j) and the analysis of PAC-1 flow cytometry binding data revealed that the integrin activation process modulates the particle population ≈20% active, in the pre-activation state, to ≈91% active when activated with divalent cations (Figure 4i,j).Interestingly, similar to the integrins on platelet surface, treatment of activated IIb3 particles with 1mM ethylenediaminetetraacetic acid (EDTA) can switch off IIb3 activation and reduce fibrino-gen binding (Figure 4g).After EDTA treatment the positive fibrinogen binding population decreased from ≈85% to ≈45%.
While the XPS and ligand-binding data demonstrates that IIb3 is immobilized and can adopt the high-affinity ligandbinding conformation, the data does not directly reveal whether IIb3 is covalently bound to DETA coated substrates.To evaluate this, ferromagnetic silica particles were prepared with and without the EDC/NHS coupling step producing covalently immobilized IIb3 particles and non-covalent physisorbed IIb3 particles, respectively (Figure 5a,b).Flow cytometry analysis for PSI-E1 binding (Figure 5c) revealed that both the covalently immobilized and physisorbed particles loaded approximately equal amounts of IIb3.However, after the addition of 2% SDS, the physisorbed particle fluorescence intensity was reduced by half while the covalently immobilized IIb3 particles signal remained unchanged, indicating that IIb3 was indeed covalently bound to DETA-coated surface (Figure 5c).This result is consistent with other literature examples where 20% SDS was not able to remove the bulk of physisorbed protein form various surfaces. [41]In addition, the physisorbed and covalently bound IIb3 particles were activated and analyzed for fibrinogen binding.Despite similar loading of IIb3, covalently bound and activated particles bound more fibrinogen, while the covalently bound pre-activation IIb3 particles showed less non-specific binding to fibrinogen than the physisorbed pre-activated IIb3 particles (Figure 5d).These data indicate that covalent oriented attachment of IIb3 enhances the active binding conformation of the integrin and reduces non-specific adsorption in the preactivation state.

Bioactivity of 𝜶IIb𝜷3-Coupled Particles
It is imperative that the immobilized IIb3 be biologically active and that the SAM can tolerate biologically relevant matrices.As we previously demonstrated, platelet surface IIb3 integrin plays a vital role in hemostasis by mediating platelet aggregation through the binding of fibrinogen and other yet unknown ligands, which leads to the crosslinking of adjacent platelets and the formation of a platelet plug. [17,29]To evaluate the biological activity of DETA-immobilized integrin, DETA SAM and activated IIb3 ferromagnetic 1 μm silica particles were labeled with Alexa Fluor 488, for co-aggregation with wild type (WT) C57 murine platelets.To confirm that the Alexa Fluor 488 labeling process does not interfere with IIb3 function, Alexa Fluor 488labeled activated IIb3 and DETA SAM particles were independently incubated with purified human fibrinogen, washed, and imaged (Figure 6a).The addition of fibrinogen caused large aggregates of activated IIb3 particles while the DETA monolayer particles looked no different to labeled particles in the absence of fibrinogen (Figure 6a).
Co-aggregation experiments were performed by initiating thrombin induced platelet aggregation between Alexa Fluor 488 labeled DETA or activated IIb3 particles with gel purified WT murine platelets.Upon imaging the resulting aggregates, it was unclear whether the DETA or activated IIb3 particles interacted with platelets as both seemed to be surrounded by platelet aggregates.The platelet aggregate surrounded particles were magnetically pulled down and imaged to decipher if the particles actively interact with platelets.The resulting images (Figure 6b) clearly illustrate that the IIb3-coupled particles interact directly with platelets and were incorporated into the platelet aggregates while the DETA particles did not interact with the murine platelets and are visibly separated from one another after pulldown.Furthermore, it was tested whether IIb3 coupled particle incorporation also occurs in the absence of fibrinogen/VWF, as we observed occlusive thrombi still occurred in fibrinogen/VWF double deficient mice. [28,29]Both DETA and IIb3 particles were co-aggregated with fibrinogen/VWF −/− murine platelets, where aggregation was induced by thrombin.Figure 6c unambiguously demonstrates that the particles do in fact interact with the yet unidentified ligands that mediate fibrinogen/VWFindependent platelet aggregation. [28,29]o further investigate the integrin coated particle performance in biologically relevant media, murine fibrinogen/VWF −/− plasma pulldown experiments were performed to determine if known activation-dependent IIb3 ligands could be identified.Pre-activated and activated IIb3 particles were incubated in murine fibrinogen/VWF −/− platelet poor plasma.The particles were pulled down and then analyzed by liquid chromatographytandem mass spectrometry (LC-MS/MS) to identify bound proteins.The quantitative values (normalized total peptide spectra) of identified ligands are depicted in Figure 6d (and tabulated in Table S4, Supporting Information).The well established IIb3 ligands of fibronectin and vitronectin were identified as well as recently identified ligands apolipoprotein A-IV [42] and complement factor H. [43] Novel ligands were also identified, including plasma kallikrein, which has been recently found to enhance platelet aggregation induced by sub-physiological doses of ADP leading to the irreversible second wave of aggregation. [44]The activated IIb3 particles had the highest quantitative value for each ligand, indicating these are IIb3 activation specific ligands.It is clear, by the data presented, that DETA-coupled IIb3 retains the active conformation in biologically relevant matrices.A property that makes our protein multimer immobilization platform ideal for use in many biotechnologies, cell mimetic or clinical diagnostic applications.

Platform Fabrication and Evaluation
Although conventional methods for (multimeric) protein surface coupling strategies (such as physisorption [42] or capture by surface immobilized antibodies [45] can be effective, they also possess drawbacks that can hinder protein conformation and thus prevent the full function of the protein receptor.For instance, all lack the ability to control the orientation of the immobilized protein, which can lead to higher variability and lower reproducibility in applications that require a multimeric protein such as integrin to be functional.Also, some immobilization methods are complex, requiring multiple reaction steps that can lead to reduced reproducibility. In this study, we describe a robust yet facile three-step surface chemistry for the oriented coupling of functional proteins using the heterodimeric IIb3 integrin as an example (Figure 1c).The foundation of this platform is the highly versatile organosilane DETA linker-based SAMs that can be formed on many ─OH bearing substrates under standard atmospheric conditions.The advantage of using DETA over other organosilane linkers to generate the SAM, is its exceptional anti-fouling capabilities.][55] The fouling resistance of fibrinogen is thought to be linked increased hydrophilicity of the surface after DETA functionalization, leading to hydrogen bonding between DETA primary amines and water molecules increasing surface hydration, which is a property linked to fouling resistance. [6,47,49]owever, to fully access the antifouling properties of the DETA coating further studies evaluating the fouling of complex protein mixtures (e.g., blood serum) utilizing mass sensitive sensors (e.g.bulk acoustic wave sensors [56] ) or fluorescence microscopy methodologies would need to be performed.Covalent integrin coupling to DETA is facilitated by the reaction between the EDC/NHS activated acidic residue rich tail of the recombinant 3 subunit and the primary amine terminated DETA SAM without other possibly destabilizing modifications that are necessary for targeted oriented immobilization. [32,33]he conformation of DETA coupled IIb3 was modulated from the low affinity to the high affinity state by incubation with divalent cations, as confirmed through the binding of conformation dependent endogenous ligand, fibrinogen (Figures 3b,d  and 4c-e) and conformation specific PAC-1 mAb (Figure 3b,e  and 4g-i).Notably, treatment with cation chelating EDTA can switch integrins from the active conformation back to the inactive conformation, a function similar to integrins on the cell surface (Figure 4g).
Oriented covalent coupling of IIb3 onto DETA dramatically enhanced the binding of fibrinogen compared to particles fabricated with IIb3 physisorbed onto DETA (Figure 5d), even though both fabrication methods loaded the same amount of integrin onto the particle surface (Figure 5c).Furthermore, preactivated covalently bound particles resisted non-specific fibrinogen binding far better than physisorbed particles (Figure 5d).This is likely due to the covalent site-specific immobilized integrin being more uniformly distributed and oriented upward on the surface compared to physisorbed IIb3, providing more solvent penetration into the coating, which is a putative fouling resistance mechanism determined through neutron reflectometry and molecular dynamics calculations. [5,6]iterature reported binding affinities for the interaction between IIb3 and fibrinogen vary widely with a reported K d range from ≈50 nM to >2000 nM. [45,57,58]It is postulated that the greatly varying literature K d values reflect the uncertainty in the activation state of the integrin probe, [45] with methods that retain the integrin in the solid phase, such as cell surface analysis, produce tighter binding affinities.The K d apparent obtained from direct binding experiments of immobilized IIb3 is 78 ± 7 nM (Figure 4f), on the higher affinity end of the literature range.We postulate that following the oriented covalent coupling of the 3 subunit, the positively charged c-terminal tail of the IIb subunit becomes anchored to the substrate through interactions with the negatively charged acidic tail of the 3 subunit and that these non-covalent interactions between the positively and negatively charged tails provide the structural flexibility necessary for integrin activation and the stabilization.

Biological Activity of Surface Coupled Integrin 𝜶IIb𝜷3
The biological activity of immobilized heterodimeric IIb3 integrin was assessed through ex vivo pulldown and platelet co-aggregation experiments that were performed as part of a study of fibrinogen/VWF independent platelet aggregation and thrombosis.The proteomic analysis identified that activated IIb3 coupled particles pulled down known activation specific binding partners: fibronectin, vitronectin, apolipoprotein A-IV, and complement factor H (Figure 6d).The assay also identified previously unknown IIb3 ligands including plasma kallikrein; a factor that has been recently identified to enhance platelet aggregation. [44]This data demonstrates that our platform is powerful for the construction of molecular probes to identify novel ligands, such as plasma kallikrein, and the yet unidentified IIb3 integrin "X-ligands" that mediate novel fibrinogen-independent platelet aggregation, [28,29] a discov-ery that could have broad implications in the treatment of thrombosis, bleeding disorders, cardiovascular diseases, and tumor metastasis. [24,30][61] Construction is based around a nano or microparticle core (i.e., liposome, latex, albumin, PLL-PLGA), which is then decorated with proteins or peptides. [60]One aspect that all platelet mimetic hemostats share, is that they are void of integrins on the surface and that primary functions of platelet integrin IIb3 and 21 (collagen receptor) are achieved by utilizing mimetic peptides and antibodies, a trait that is common with many biotechnologies.The drawback of this approach is that integrins bind to more than one receptor or ligand and often serve secondary roles which are not emulated by the peptide/protein mimic. [62]The integrin coupled particles presented in this study possess platelet mimetic hemostat function as they aggregate with platelets.Specifically, activated IIb3 coupled ferromagnetic silica particles were incorporated into aggregates of both WT and fibrinogen/VWF −/− murine platelets (Figure 6), indicating retention of biological activity, that can be useful for therapeutic applications.Based on the surface area of the 1 μm ferromagnetic silica particles, we estimate that a 50x higher density of immobilized IIb3 can be achieved compared to platelets.This further demonstrates that the platform can be developed into a topically applied hemostat to help arrest bleeding from a severe injury.This is similar to our previously developed mesoporous silica particle-based topical hemostat that reduced bleeding time. [63,64]The use of oriented functional integrin ectodomains as surface decoration is a significant advancement in platelet mimetic hemostat technology as this heterodimeric single receptor provides the multi-role capabilities that mimetics cannot deliver.
Our novel IIb3-coupled particles have the potential to make an impact in the diagnosis of auto-and allo-immune responses against platelets that cause thrombocytopenia and bleeding disorders.For example, Immune thrombocytopenia (ITP) is a common bleeding disorder where autoantibodies target platelet surface receptors.Approximately 70-80% of ITP mediating autoantibodies target integrin IIb3 and ≈20-40% of the autoantibodies target another platelet surface receptor GPIb.[67] The gold standard assay MAIPA is rarely clinically performed due to assay complexity and lack of analytical sensitivity, leaving diagnosis to be made by response to treatment. [67]Due to their high density of IIb3, our integrin coated particles can markedly enhance the avidity between low affinity autoantibodies and IIb3.They could also be adapted to determine the patient's autoantibody specificity and assist physicians to select the appropriate course treatment the first time rather than relying on response to therapy.Furthermore, platelets have been shown to support tumor growth and metastasis through a variety of interactions wherein platelet IIb3 integrin is a key player. [30]Our SAM-IIb3 integrin coupling could be applied to a biocompatible hydroxide (─OH) bearing particles engineered to release a drug cargo.Our functional integrin surface coupling would allow the particles to interact with tumor cells and release their drug inducing tumor cell death.
In summary, our heterodimeric protein biomaterial immobilization platform is comprised of a combination of the modification of a hydroxyl bearing surface with a self-assembled monolayer (SAM) and the modification of the polypeptide probe.IIb3 integrin was implemented to demonstrate this platform and the integrin was switchable the biologically active highaffinity ligand binding conformation and bind endogenous ligands with extremely low background signal.Considering the sequence homology between different integrins and the ubiquity in the use of oppositely charged tails to stabilize recombinant integrin ectodomains, the oriented and covalent immobilization strategy described here can be easily applied to all other integrins and more generally to heterodimeric surface proteins.Furthermore, we propose that our immobilization platform can be extended beyond dimers to other protein multimers, such as the trimeric SARS-CoV-2 spike protein, an advancement of the current repertoire of available chemical tools for the functional oriented immobilization of complex proteins.Therefore, the methodology established in the current study can be applied to and have a broad impact on fundamental biology, biomaterials, biosensors and nano-and biotechnology that can advance ods for both diagnosis and therapy.

Experimental Section
Unless otherwise specified, all reagents were purchased from Sigma-Aldrich and used as received.Furthermore, all buffers and aqueous solutions were prepared using ultrapure distilled deionized water (ddH 2 O) with a measured resistivity ≥18.0 MΩ•cm.Ferromagnetic silica particles and platelets were counted using a Beckman Coulter Multisizer 4e Particle Analyzer.
DETA Silanization of Glass Slides: Glass slides (5 mm 2 ) were cleaned by individually soaking for 45 min in piranha solution (3/1 (v/v) mixture of concentrated H 2 SO 4 and 30% H 2 O 2 ) at 90 °C.The glass slides were rinsed with ddH 2 O (x3) followed by spectrograde methanol (x3).Next, the slides were sonicated in spectrograde methanol for 2 min, and individually placed in vials to be dried at 180 °C for 2 h.After drying, the glass substrates were individually stored in screw top vials until needed but no longer than 24 h.For SAM formation the cross-linker N-(3-Trimethoxysilylpropyl) diethylenetriamine (DETA) was used.To prevent any undesired reaction of DETA molecules with the walls of the test tubes used during silanization, all glassware was pre-treated with octadecyltrichlorosilane (OTS) (1/20 (v/v) solution in anhydrous toluene).A 1% (v/v) solution of neat DETA diluted in anhydrous toluene was prepared.The cleaned glass slides were individually soaked in 1 mL of the DETA solution overnight.Afterward, the glass slides were rinsed twice with dry toluene, sonicated in toluene for 5 min, and finally rinsed again with toluene.The previous rinsing procedure was repeated with spectrograde methanol.After a final rinse with spectrograde methanol, the slides were dried under a gentle stream of N 2 gas.
DETA Silanization of Ferromagnetic Silica Particles: Unless otherwise specified rinsing of ferromagnetic silica particles consists of magnetically pelleting the particles, removing the supernatant and resuspending in new solution.1 μm average diameter ferromagnetic silica particles were purchased from Magna Medics.The silica particles were activated by treatment with piranha solution for 45 s, upon which NaOH was added to immediately neutralize the piranha solution.The particles were then rinsed (3x) in ddH 2 O.The rinsing procedure was repeated with spectrograde methanol and anhydrous toluene.Following the rinsing procedure, a 1% (v/v) solution of neat DETA was diluted in anhydrous toluene in an OTS treated glass vial.Ferromagnetic silica particles were subsequently im-mersed in the DETA solution to a final volume equal to that which was originally aliquoted from the activated bead stock, in a screw top glass vial pretreated with OTS.The vial was capped and incubated at room temperature on a bench top oscillator overnight.Afterward, the freshly silanized particles were rinsed (3x) with anhydrous toluene.The rinsing procedure was repeated with spectral grade methanol and finally PBS.Particles were resuspended in PBS (10 mMm sodium phosphate, 140 mM NaCl, 5 mKCl pH 7.4) to a final concentration of 1 × 10 8 particles/mL.
Coupling of Human IIb3 Ectodomain to DETA Silanized Glass Slides or Silica Particles: Recombinant human ectodomain IIb3 expressed in CHO cells modified with a positively charged amino acid tail on the IIb subunit and a negatively charged acidic amino acid rich tail on the 3 subunit (R&D Systems) was freshly immobilized onto silanized glass slides or particles as required.Immobilization buffer was prepared (4 mM EDC, 10 mM sulfo-NHS, 10 mm sodium phosphate, 140 mM NaCl, 5 mM KCl pH 7.4) and recombinant human extracellular IIb3 was taken up in this buffer to a final optimized concentration of 250 μg mL −1 (optimization of integrin loading depicted Figures S5 and S6, Supporting Information).
Glass slide IIb3 Coupling: Freshly silanized glass slides were immersed in the IIb3 containing immobilization solution and incubated overnight (16 h) on a bench top oscillator at 4 °C.Following the incubation, the glass slides were rinsed with PBS buffer and incubated in borate buffer (10 mM borate, 140 mM NaCl pH 8.5) for 2 h at 4 °C to hydrolyse any remaining protein coupled NHS esters. [35]The glass slides were washed (3x) with copious amounts of PBS.The slides were dried under a gentle stream of N 2 gas at room temperature and stored in screw top vials at 4°C until needed.
Ferromagnetic Silica Particle IIb3 Coupling: Freshly silanized particles were taken up in the integrin containing immobilization buffer to a final concentration of ≈1 × 10 8 particles/mL and incubated at 4 ˚C overnight (16 h) on a bench top oscillating shaker.Following incubation, the particles were pelleted, washed with borate buffer (10 mM borate, 140 mM NaCl pH 8.5) and incubated in borate buffer for 2 h at 4 °C to hydrolyse any remaining protein coupled NHS esters.The integrin coupled particles were then pulled down, rinsed (3x) with copious amounts of PBS and suspended in PBS at a concentration of ≈1 × 10 8 particles/mL and stored at 4 ˚C.
Activation of IIb3 SAMs Activation of SAM immobilized integrin IIb3 was achieved by 72 h incubation of either coupled glass slides or ferromagnetic silica particles in activation buffer (1 mMmeach of CaCl 2 , MnCl 2 and MgCl 2 taken up in PBS).
X-Ray Photoelectron Spectroscopy: X-ray photoelectron spectroscopy (XPS) to evaluate substrate silanization (SAM formation) and subsequent IIb3 immobilization was performed with a Theta probe XPS Instrument (ThermoFisher Scientific) at Surface Interface Ontario (University of Toronto, Toronto, Ontario, Canada).Glass surfaces and ferromagnetic silica particles were analyzed with monochromated Al K X-rays at takeoff angle of 60°relative to the normal.The binding energy scale was calibrated to the C1s signal at 285 eV.Peak fitting and data analysis were performed with the Avantage Data System software package (ThermoFisher Scientific) provided with the instrument.Complete XPS atomic abundance data for both glass slides and ferromagnetic silica particles are tabulated in Tables S1 and S2 (Supporting Information).
Confocal Fluorescence Microscopy of Glass Slides: Freshly prepared glass slides were incubated with the desired concentration, 50 nM, of fluorescently labeled antibody or ligand in PBS at a volume sufficient to completely submerse the glass slides and incubated for 1 h.Following the incubation period, the slides were rinsed (3x) with copious amounts of PBS and dried with a gentle stream of N 2 gas.The glass slides were imaged using a Zeiss LSM 700 confocal microscope and the images analyzed with the ImageJ software package.
Flow-Cytometric Analysis of Ferromagnetic Silica Particles: Unless otherwise stated, all flow cytometry experiments were conducted under the same conditions using either a BD FACS Calibur or BD Fortessa X20 flow cytometer.3 μL of 1×10 8 particles/mL were mixed with the desired concentration of fluorescently labeled antibody or ligand (5 to 370 nM fibrinogen, 50 nM PAC-1/PSI-E1, and 100 nM anti CD62p/anti GPIb) to a final volume of 100 μL in PBS and incubated for 1 h.Following incubation, the samples were diluted to a final volume of 1 mL and flow cytometrically analyzed.Samples that were compared with one another were run on the same instrument under the same instrumental conditions (signal gain, flow rate etc.).Any differences in absolute mean fluorescence intensity (MFI) value between different experiments was due to variance between the two flow cytometers utilized.Since the experiments were performed using synthetic particles, the forward scatter and side scatter plots depicted only a single population, which was gated (Figure S4, Supporting Information), and from which the fluorescence intensity was calculated.The data was analyzed using the FlowJo VX software package (BD).
Platelet Particle Co-aggregation Assay: Gel filtered platelets were prepared as previously described. [29]DETA or activated IIb3 particles were co-aggregated with wild type (WT) or VWF/fibrinogen −/− platelets by incubating equal volumes of 1 × 10 8 /mL particles with 1 × 10 8 /mL gel filtered platelets in PIPES buffer (25 mMm 1,4-Piperazinediethanesulfonic acid (PIPES), 140 mM NaCl, 4 mM KCl, 5.5 mM D-glucose pH 7.0) to a final volume of 500 μL.The platelets were activated by the addition of human thrombin to a final concentration of 5 units/mL, followed by incubation at RT for 45 min on a vertical rotating mixer at max speed.The particles were magnetically pulled down, washed 3x and imaged under a Zeiss Axiovert 200 inverted fluorescence microscope at 40x magnification.
Liquid Chromatography Tandem Mass Spectrometry Analysis of IIb3 Coupled Particles Incubated in Fibrinogen/ VWF −/− Platelet Poor Plasma: Platelet poor plasma (PPP) from fibrinogen/VWF −/− mice was prepared as previously described. [29]DETA, pre-activation and activated IIb3 particles (at a final concentration of 2.5 × 10 7 particles/mL) were each incubated in 100 μL of PPP overnight (16 h) at 4 °C.The particles were pulled down, washed (3x) with PBS, and resuspended in a final volume of 50 μL PBS.Next, the particles were transferred to the SPARC BioCentre (Hospital for Sick Children, Toronto, Ontario, Canada) where the pulled down proteins were cleaved off the particles with trypsin and proteomically analyzed by liquid chromatography tandem mass spectrometry.
Data Fitting and Statistical Analysis: Data was tabulated, plotted, and analyzed using the GraphPad Prism 8 software package (GraphPad Software).Data were presented as mean ± standard deviation.Statistical significance was assessed by an unpaired, two-tailed Student's t-test.The apparent dissociation constants (K d apparent ) were determined by plotting the dose response curve and fitting to the stated binding model within Graph-Pad Prism.70]

Figure 1 .
Figure 1.Integrin coupled surface fabrication scheme: Schematic depictions of a) self-assembling monolayer formation, b) integrin activation states, and c) reaction scheme of the IIb3 integrin coupled surface.The formation of the IIb3 integrin containing coating is a facile three step process where a DETA SAM is formed on an ─OH bearing substrate (Step 1), covalent immobilization of the integrin (Step 2) and activation of the integrin into the high-affinity ligand binding conformation (Step 3).

Figure 2 .
Figure 2. X-Ray Photoelectron Spectroscopy (XPS) Characterization of IIb3 Integrin SAM-Coupling.XPS survey spectra a) bare cleaned, b) DETA SAM coated, and c) IIb3-coupled 5 mm 2 glass slides.d) Changes in relative atomic percentage of carbon, nitrogen, oxygen, and silicon at each stage of the surface fabrication process.e) Silicon 2p narrow scans of bare and DETA coated glass substrates depicting the binding energy shift due to silinization of the DETA SAM.f) Nitrogen 1s narrow scan of DETA coated and IIb3 coupled glass slides depicting the binding energy shift caused by protein immobilization.n ≥ 3 data presented as mean ± SD, analyzed with a two-tailed t-test, *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Figure 3 .
Figure 3. Characterization of integrin-coupled glass substrates: a) Depiction of the integrin conformation states and the approximate binding locations of conformation-independent anti-3 antibody PSI E1, conformation-dependent antibody PAC-1 and conformation-dependent endogenous ligand, fibrinogen.b) Fluorescence microscopy images of glass substrates at each stage of the manufacturing process incubated with fluorescently labeled PSI E1, fibrinogen, and PAC-1.The fluorescence intensity of the microscopy images of the glass slides were quantified using the ImageJ software package for c) PSI E1, d) fibrinogen, and e) PAC-1 binding.n ≥ 3 data presented as mean ± SD, analyzed with a two-tailed t-test, *p < 0.05, **p < 0.01, and ***p < 0.001.

Figure 4 .
Figure 4. Flow cytometry characterization of IIb3-coupled particles: Ferromagnetic silica particles at each stage of the coating fabrication process were analyzed with flow cytometry for binding to a,b) PSI E1, c) 4G6 d,e) fibrinogen.Representative flow cytometry data and corresponding mean fluorescence intensity (MFI) plotted in bar graphs of 5 nM Fibrinogen.f) Fibrinogen dose-response curves of the binding of varying concentrations (5-370 nM) of Alexa fluor 488-fibrinogen binding to activated IIb3 and DETA (control) particles.The DETA signal was subtracted from the activated IIb3 signal.The activated IIb3 dose response curve was fit to a one site specific binding model producing a K d of 78 ± 7 nM.g) Flow cytometry dot plots and corresponding MFI bar graphs of DETA and activated IIb3 particles binding fibrinogen before and after treatment with 1 mM EDTA.h,i) DETA, Pre-activation and activated IIb3 particles analyzed for PAC-1 binding.j) Bar graph of DETA, Pre-activation and, activated IIb3 particles in the activate integrin population.k-n) DETA, Pre-activation IIb3 and activated IIb3 particles were evaluated against negative control antibodies, that target other platelet surface receptors, k,l) anti-CD62p and m,n) anti-GPIB.n ≥ 3 data presented as mean ± SD, analyzed with a two-tailed t-test, ns = not significant *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Figure 5 .
Figure 5. IIb3 integrin is covalently immobilized onto the SAM and covalent immobilization increases ligand binding efficiency: Schematics depicting IIb3 a) covalently immobilized and b) physiosorbed onto the DETA SAM.c) Covalently immobilized and physisorbed IIb3-coupled silica magnetic particles are flow cytometrically evaluated for FITC-PSI E1 binding before (left) and after (right) treatment with 2% SDS.d) Covalently coupled and physisorbed IIb3-coupled ferromagnetic silica particles are flow cytometrically analyzed for Alexa Fluor 488-fibrinogen binding in the pre-activation and activated integrin states.n ≥ 3 data presented as mean ± SD, analyzed with a two-tailed t-test, ns = not significant and ***p < 0.001.

Figure 6 .
Figure 6.IIb3 integrin-coupled ferromagnetic silica particles co-aggregate with murine WT and fibrinogen/VWF -/-particles demonstrating hemostat function: DETA and activated IIb3 particles were labeled with Alexa Fluor 488, and a) incubated with fibrinogen and visualized with bright field and fluorescence microscopy to ensure the labeling process did not diminish IIb3-fibrinogen binding.b) Alexa Fluor 488-labeled DETA and activated IIb3 particles were incubated at a 1:1 ratio with wild type C57 murine platelet and aggregation was induced with thrombin.The particles were pulled down washed and visualized with bright field and fluorescence microscopy.c) DETA and activated IIb3 particles were incubated at a 1:1 ratio with fibrinogen/VWF −/− murine platelets, aggregation was induced with thrombin.The particles were pulled down washed and visualized with bright field microscopy.White arrows indicate platelets and black arrows indicate either DETA and activated IIb3 particles.d) Liquid chromatography tandem mass spectrometry analysis DETA, pre-activation IIb3 and activated IIb3 particles incubated with fibrinogen/VWF −/− murine plasma.The analysis identified the known activation specific IIb3 ligands of fibronectin, vitronectin apolipoprotein A-IV, and complement factor H as well as novel ligand, plasma kallikrein.