Coagulation Sensor

On Silicon Photonic Sensors

Sara M. Medfisch1,3, Ellen M. Muehl1,3, James H. Morrissey2* & Ryan C. Bailey1* Blood coagulation is regulated through protein–protein and protein–lipid interactions that occur at the sub-endothelium following vascular damage. Soluble clotting proteins bind to membrane components in a phosphatidylserine (PS) dependent manner to assemble multi-protein complexes that regulate clot formation; however, PS is of limited abundance physiologically. In this manuscript, we investigate synergy between PS and phosphatidylethanolamine (PE)—a lipid of much higher abundance naturally. Using a label-free, silicon photonic technology, we constructed arrays of Nanodiscs having variable lipid composition and probed the binding interactions of seven different clotting factors with GLA domains that have never been studied in tandem experiments before.

Figure: System Model & Simulation Flow for Coagulation Sensor

The factors studied were prothrombin, activated factor VII, factor IX, factor X, activated protein C, protein S, and protein Z. Equilibrium dissociation constants (Kd) for each coagulation factor binding to Nanodiscs with unique compositions of PE and PS were determined. While all factors showed greater binding affinities in the presence of PS and PE, the most dramatic improvements in binding were observed when PS quantities were lowest. This demonstrates that synergy is effective in promoting coagulation factor binding under physiological lipid compositions, as opposed to the artificially high PS content probed in most in vitro activity studies.

Protein Z

The phospholipid content of the cell membrane is a key regulator of the blood coagulation ­cascade1,2. Exposure of tissue factor and the negatively charged lipid, phosphatidylserine (PS), via cell membrane damage is one of the initiating factors of the ­cascade3,4. The majority of steps in the clotting cascade occur at the membrane surface and involve multiple PS lipid-binding ­proteins4. The most common PS-binding domain found in both pro- and anti-coagulant factors is the γ-carboxyglutamic acid-rich GLA domain located at the N-terminus of these pro- teins [Procoagulants: factor VII (fVII); factor IX (fIX); factor X (fX); and prothrombin (PT). Anticoagulants:

Www.Nature.Com/Scientificreports/

activated protein C (aPC); protein S (PrS); and protein Z (PrZ).] The GLA domain reversibly binds to PS lipids in a calcium dependent ­manner4. While the structures of the GLA domains are highly conserved, they display membrane binding affinities that vary by more than two orders of ­magnitude2,4, which suggests differences in the lipid binding preferences of each.

Even though PS lipids are necessary for optimal activity of many clotting factors, they only constitute ~ 12% of the phospholipid content of the cell ­membrane5, which is much lower than what is required for optimal pro- coagulant activity of activated fVII (fVIIa) in vitro (~ 30% PS in liposomes)1. Previous studies have shown that lipid composition influences the binding and activity of GLA domain-containing clotting proteins. For example, even though GLA domains of clotting proteins are structurally and compositionally homologous, aPC and fVIIa bind phosphatidic acid (PA) lipids much more tightly than PS ­lipid2. Maximal rates of activation of fX by fVIIa bound to tissue ­factor6–8, of activation of prothrombin by fXa bound to factor ­Va9, and of inactivation of factor Va by ­aPC10, have been reported to require much less PS when membranes also contain phosphatidylethanola- mine (PE). This phenomenon has been termed PE–PS synergy. The “Anything But Choline” (ABC) hypothesis explains PE–PS synergy by stating that almost any phospholipid, other than those with choline head groups, can work cooperatively with PS lipids to greatly reduce the amount of PS required for binding and activation of clotting ­proteins8. PE comprises ~ 25% of the phospholipid content of the plasma membrane and is normally sequestered, along with PS, in the inner leaflet of the ­bilayer5. Cell membrane damage exposes both PE and PS to support of blood coagulation reactions, showing physiological relevance for PE–PS synergy.

Since most of the previous work on PE–PS synergy focused on enzyme activity assays using membrane-bound enzyme, cofactors and substrates, it is not clear which of the proteins within these multi-protein complexes actu- ally exhibited enhanced membrane binding in the presence of PE. This is especially important since some of the proteins in these complexes (e.g., factors Va and VIIIa), bind to PS in membrane surfaces via discoidin-type C2 domains, not GLA ­domains4. In fact, to date, the only GLA domain-containing protein whose binding to membrane surfaces has been directly shown to be enhanced by PE is fX, raising the question of how generally PE–PS synergy promotes the binding of GLA domain-containing proteins.

The aim of this study was to comprehensively and quantitatively evaluate the degree to which PE–PS lipid synergy enhances membrane binding of all seven GLA domain-containing proteins of the clotting cascade in tandem experiments. As a model membrane system, Nanodiscs were constructed to present well-defined, vari- able lipid compositions. Nanodiscs are small lipid bilayer discs held together by two membrane scaffold proteins (MSPs) that offer a high degree of control over lipid ­composition11,12. Nanodiscs have proven to be a useful tool in the study of membrane-lipid interactions of the blood coagulation ­cascade1,2,8. Various Nanodisc were spa- tially arrayed onto a label-free and highly multiplexable silicon photonic detection platform to allow in series determination of binding constants (Fig. 1). Recently, we demonstrated the ability to utilize this combination of Nanodiscs and microring resonator sensors for high-throughput interrogation of protein binding to model membrane ­surfaces13,14, including coagulation factors, PT, fX, fVIIa, and aPC. This work demonstrated that bind- ing interactions can be rapidly probed simultaneously with reduced time and reagent consumption compared to conventional methods. Herein, we further extend this technology to probe the binding interactions of all seven GLA domain-containing coagulation factors to Nanodiscs presenting defined ratios of PS and PE. These experiments were performed in series on a single array of Nanodiscs to eliminate environmental changes day- to-day that can result in variation in array formation. Thus, providing an optimal surface for direct comparison of binding improvements due to lipid composition.

More specifically, we determined Kd values for fVIIa, fIX, fX, PT, aPC, PrS, and PrZ binding to Nanodiscs at eight distinct lipid compositions. We directly compared Kd values of binding to Nanodiscs of 10, 25, 40 and 50% PS to Nanodiscs containing 10% PS/40% PE, 25% PS/25% PE, and 40% PS/10% PE, or 50% PE. The balance in all Nanodiscs was made up with non-interacting phosphatidylcholine (PC) lipids. Utilizing silicon photonic microring resonators as our sensing platform, we have shown that sequential titrations can be performed on a Figure 1.   (A) Schematic of Nanodisc physisorption to a microring resonator. The MSP (purple) stabilizes the lipids (PC: yellow; PS: red; and PE: blue) to form the Nanodisc bilayer mimetic. Note The schematic is not to scale. The microring resonator is 30 µm in diameter while the Nanodiscs are about 13 nm in diameter and 5 nm in thickness. (B) Spotting layout of the sensor chip for protein titrations. The flow path for experiments is shown in light blue. Figure created using Adobe Illustrator CC2020 (https​://www.adobe​.com/produ​cts/illus​trato​r.html).

Www.Nature.Com/Scientificreports/

Nanodisc array ­surface14. Thus, experiments with each clotting factor were performed sequentially on the same Nanodisc array surface to eliminate day-to-day variation observed with other experimental setups. We found that PE had a significant effect on Kd, driving much tighter binding of all factors to the model mem- brane surface. Using these calculated Kd values, we quantitated the fold-change in binding due to the addition of PE to PS. Interestingly, the magnitude of binding improvement was greatest at the lowest PS contents, consistent with physiological lipid abundances. This synergistic improvement in binding was more striking for the clotting proteins that showed poorest binding to PS alone which have not been shown to observe this synergy—fVIIa, PrZ, and aPC. These proteins yield very high (or unmeasurable) Kd values (poor binding) at near physiologi- cal PS content and therefore PE–PS synergy appears essential for regulation of coagulation. This work offers a comprehensive and quantitative view of the synergistic influence of PE on binding of clotting cascade proteins across all seven GLA domain-containing clotting proteins and is a high-throughput complement to activity-based studies of their pro- and anticoagulant activities modulated by protein–lipid interactions.

Materials And Methods

Chemicals and reagents. Lipids presenting phosphocholine (POPC; 1-palmitoyl-2-oleoyl-sn-glycero- 3-phosphocholine), phosphoserine (POPS; 1-palmitoyl-2-oleoyl-sn-glycero-2-phosphoserine), and phosphoe- thanolamine (DOPE; 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine) head groups were purchased from Avanti Polar Lipids (Alabaster, AL). MSP1E3D1 was expressed in E. coli and purified as described ­previously12.

Human fVIIa, fX, PT, aPC, and PrZ were purchased from Enzyme Research Laboratories Inc. (South Bend, IN). Human fIX and PrS were purchased from Haematologic Technologies (Essex Junction, VT). Amberlite XAD-2 hydro-phobic beads and all other chemicals were purchased from Sigma Aldrich (St. Louis, MO) and used as received unless otherwise noted. Buffers were prepared with 18.2 MΩ water and sterile filtered prior to use.

Solution preparation. Nanodisc solutions were prepared in a TBS buffer (20 mM Tris–HCl, 100 mM NaCl, and 0.01% (w/v) ­NaN3; pH 7.4). PT, fIX, and fX solutions were prepared in a HEPES buffer (10 mM HEPES, 150 mM NaCl, 50 μM EDTA, 2.5 mM ­CaCl2, 0.1% (w/v) PEG 8000; pH 7.4). Solutions of fVIIa, aPC, PrS, and PrZ were prepared in HEPES buffer with 0.2% (w/v) BSA. The HEPES rinse buffer (HEPES(−)) for surface regeneration was made without ­CaCl2.

Nanodisc preparation and purification. Nanodisc preparation and purification has been described in detail ­previously11,12,15. Briefly, lipids solubilized in chloroform were measured into test tubes and dried under nitrogen. For Nanodiscs containing mixtures of POPS, DOPE, and/or POPC, the lipids were mixed at defined ratios prior to drying. After drying, lipids were placed in a desiccator, under vacuum overnight. Once com- pletely dry, lipids were dissolved in TBS buffer with 100 mM deoxycholate to give a final molar ratio of 2:1 dexoycholate:phospholipids. Dissolved lipids were then combined with MSP1E3D1 in TBS to give a final molar ratio of 135:1 phospholipid:MSP. The solution of MSP and lipids was actively mixed at 4 °C for approximately 1 h. Amberlite XAD-2 hydrophobic beads were then added to the MSP/lipid solution with the amount of bead solution being half the total of the MSP/lipid solution volume. This combined solution was then mixed for approximately 1.5 h at 4 °C. Amberlite XAD-2 hydrophobic beads were then removed by filtering through a 0.22 μm syringe filter. Nanodiscs were then purified using size exclusion chromatography using a Superdex 200 column (GE).

Silicon photonic microring resonators. The Maverick M1 optical scanning instrumentation and micro- ring resonator sensor chips were purchased from Genalyte, Inc. (San Diego, CA). The operation of the instru- ment has been previously ­described16–20. The sensor chips were each 4 mm × 6 mm and contained 128, 30-μm diameter active sensor microrings arranged in clusters of four, plus four temperature control microrings and two dedicated to detecting leaks from the microfluidic gasket positioned atop the sensor chip during microring detection experiments.

Sensor chip array functionalization. Prior to use, sensor chips were placed in a vial of acetone for 2 min with gentle agitation. Sensor chips were then transferred to a vial of isopropanol for 2 min. After the sensor chips had been dried with ­N2, between 0.1 and 0.2 μL of each type of Nanodisc were spotted at 0.5 μM. A spotting map showing the arrangement of the Nanodisc solutions on the sensor substrate is shown in Fig. S1. After spotting, chips were stored in a humidity chamber at 4 °C for at least 1 h before use.

Protein binding titrations. Laser-cut Mylar gaskets to direct fluid flow across the chip were aligned onto the functionalized sensor chips, assembled into a Teflon cartridge, and loaded into the sensor scanner instru- ment. A 2% solution of BSA in HEPES (−) buffer was first flowed across the chip surface at 10 μL/min to prevent the non-specific binding of proteins. For Kd determination titrations, the proteins were flowed across the chip in increasing concentrations at 10 μL/min and the response allowed to approach steady state before the next solu- tion injection. Following each titration, all of the coagulation factors were released from the surface by flowing HEPES (−) buffer solution. The titration was then performed for the next protein.

Data analysis. Data analysis was performed using custom R scripts in RStudio. Sensor traces were cor- rected for temperature fluctuations and any residual non-specific binding by subtraction of response of 100% PC Nanodiscs. The maximum shift for each protein concentration was calculated by subtracting the relative shift (Δpm) at the equilibration point for HEPES buffer (at 10 min) from the relative shift (Δpm) at equilibration for

Related Journal Articles & DOI Links

Selected peer-reviewed publications relevant to 12 Lead ECG Acquisition. Click the DOI to access the full paper (may require institutional access).

Why Choose Us?

Bangalore guidance for robotics, Spectre and autonomous systems projects.