An adhesion frequency assay for measuring receptor-ligand interaction kinetics when both molecules are anchored on the surfaces of the interacting cells is described. This mechanically-based assay is exemplified using a micropipette-pressurized human red blood cell as adhesion sensor and integrin αLβ2 and intercellular adhesion molecule-1 as interacting receptors and ligands.
The micropipette adhesion assay was developed in 1998 to measure two-dimensional (2D) receptor-ligand binding kinetics1. The assay uses a human red blood cell (RBC) as adhesion sensor and presenting cell for one of the interacting molecules. It employs micromanipulation to bring the RBC into contact with another cell that expresses the other interacting molecule with precisely controlled area and time to enable bond formation. The adhesion event is detected as RBC elongation upon pulling the two cells apart. By controlling the density of the ligands immobilized on the RBC surface, the probability of adhesion is kept in mid-range between 0 and 1. The adhesion probability is estimated from the frequency of adhesion events in a sequence of repeated contact cycles between the two cells for a given contact time. Varying the contact time generates a binding curve. Fitting a probabilistic model for receptor-ligand reaction kinetics1 to the binding curve returns the 2D affinity and off-rate.
The assay has been validated using interactions of Fcγ receptors with IgG Fc1-6, selectins with glycoconjugate ligands6-9, integrins with ligands10-13, homotypical cadherin binding14, T cell receptor and coreceptor with peptide-major histocompatibility complexes15-19.
The method has been used to quantify regulations of 2D kinetics by biophysical factors, such as the membrane microtopology5, membrane anchor2, molecular orientation and length6, carrier stiffness9, curvature20, and impingement force20, as well as biochemical factors, such as modulators of the cytoskeleton and membrane microenvironment where the interacting molecules reside and the surface organization of these molecules15,17,19.
The method has also been used to study the concurrent binding of dual receptor-ligand species3,4, and trimolecular interactions19 using a modified model21.
The major advantage of the method is that it allows study of receptors in their native membrane environment. The results could be very different from those obtained using purified receptors17. It also allows study of the receptor-ligand interactions in a sub-second timescale with temporal resolution well beyond the typical biochemical methods.
To illustrate the micropipette adhesion frequency method, we show kinetics measurement of intercellular adhesion molecule 1 (ICAM-1) functionalized on RBCs binding to integrin αLβ2 on neutrophils with dimeric E-selectin in the solution to activate αLβ2.
1. RBCs isolation from the whole blood
Note: Step 1.2 should be performed by a trained medical professional such as a nurse, with an Institutional Review Board approved protocol.
2. RBCs biotinylation
3. Functionalizing the biotin-linked ligands* on RBCs
*If your protein has no biotin link you can use one of the commercially available kits for protein biotinylation (for example, Thermo Scientific # 21955 EZ-Link Micro NHS-PEG4-Biotinylation Kit) or use biotinylated capturing antibodies as an intermediate step as shown in the video.
4. Quantification of receptor and ligand densities
5. Preparation for micropipette and cell chamber
6. Micropipette adhesion frequency assay
7. Data analysis
8. Representative Results:
Figure 1 Determination of integrin αLβ2 site density on neutrophils. Neutrophils were first incubated with 1μg/ml of E-selectin-Ig for 10min to match the experimental condition used in Figures 3,4 or without E-selectin-Ig, and then with saturating concentrations (10μg/ml) of PE-conjugated anti-human CD11a mAb (Clone HI111, see Table of specific reagents and equipment) or irrelevant mouse IgG1 for control, washed, and analyzed immediately. Samples were read on BD LSR flow cytometer with standard QuantiBRITE PE calibration beads. Panel A shows fluorescence histograms of calibration beads (pink) together with those of E-selectin-Ig treated (blue color) or untreated cells (green color). Specific CD11a mAb staining is shown in solid curves and irrelevant isotype-matched control antibody staining is shown in dotted curves. Cells treated with E-selectin-Ig (presence in all washing steps and in FACS buffer) did not affect the CD11a density as seen from the comparison with untreated cells. Panel B shows the process of density quantification. Log10 was calculated for the mean fluorescent intensity (FI) of each peak value of four calibration bead histograms from Panel A (pink circles) and for the lot-specific PE molecules per bead (from the manufacturer). A linear regression of Log10 PE molecules per bead against Log10 fluorescence was plotted. For E-selectin treated cells the Log10 FI (y) values equal 3.99 (blue solid circle) and 2.23 (blue open circle) for specific mAb and control antibody, respectivly. We solved the linear equation for x (values are plotted as green and blue circles on Panel B). x = Log10 PE/cell and, as PE:mAb ratio was 1:1, the total number of αLβ2 on neutrophils was calculated as 9587. Surface density was calculated to be 43 molecules/μm2, using 8.4μm as the neutrophil diameter22. Density of ICAM-1 was similarly measured by flow cytometery using PE-anti-human CD54 mAb, which equaled 65 mol/μm2.
Figure 2 Micropipette system schematics. Our micropipette system was assembled in house and consists of three subsystems: an imaging subsystem to allow one to observe, record and analyze movements of the micropipette-aspirated cell; a micromanipulation subsystem to enable one to select the cells from the cell chamber, and a pressure subsystem to allow one to aspirate the cells into micropipettes. The central piece of the imaging subsystem is inverted microscope (Olympus IMT-2 IMT2) with a 100x oil immersion 1.25 N.A. objective. The image is sent to a video cassette recorder through a charge couple device (CCD) camera. A video timer is coupled to the system to keep track of time. Each micropipette can be manipulated by a mechanical drive mounted on the microscope and finely positioned with a three-axis hydraulic micromanipulator. Mechanical manipulators from Newport could be used as well. One of the micropipette holders is mounted on a piezoelectric translator, the driver of which is controlled by a computer LabView code (available upon request) and the signal translates through a DAQ board via a voltage amplifier (homemade) to the piezo actuator. This allows one to move the pipette precisely and repeatably in an adhesion test cycle. The pressure regulation subsystem is used to control suction during the experiment. A hydraulic line connects the micropipette holder to a fluid reservoir. A fine mechanical positioner allows the height of the reservoir to be precisely manipulated.
Micropipettes are generated using KIMAX melting point borosilicate glass capillary tubes (with outside diameter of 1.0±0.07 mm and an inside diameter of 0.7±0.07 mm. First, the micropipettes from capillary tubes are pulled using PN-30 Narishige’ Magnetic Glass Microelectrode Horizontal Puller (Sutter Instruments Micropipette Puller is another puller option). Second, Microforge system (built in house) is used to cut the micropipettes to desired size opening. Commercial models of Microforge systems are available as well.
To avoid vibration of the micropipettes during the experiment, the microscope, along with the micromanipulators, is placed on an air suspension table.
Figure 3 Running adhesion frequency Fi for specific (A) and nonspecific (B) binding at 1s (red) and 10s (blue) contact times measured from repeated adhesion test cycles between RBCs coated with ICAM-1 (A) or hIgG (B) with human neutrophils expressing integrin αLβ2. Fi = (X1 + X2 +…+ Xi)/i (1 ≤i ≤ n), where i is the test cycle index, Xi equals “1” (adhesion) or “0” (no adhesion). Fn (n=50) was used as the best estimate for adhesion probability.
Figure 4 Kinetics of ICAM-1 binding to neutrophil integrin αLβ2 (). Adhesion probability measured as shown in Figure 3 for three cell pairs at each contact time is averaged and plotted versus contact time. The chamber medium was HBSS with 1mM each of Ca2+ and Mg2+ plus 1μg/ml of dimeric E-selectin-Ig to upregulate αLβ2 binding. To capture ICAM-1-Ig on RBCs, an intermediate step was added to incubate RBCs with 10μg/ml capture antibody (biotinylated goat-anti-human Fc antibody, eBioscience) after the streptavidin incubation step. To control for nonspecific binding two different conditions were used: 1) RBCs coated with the anti-human-Fc capture antibody and incubated with human IgG instead of ICAM-1-Ig (O) and 2) neutrophils binding to RBCs not coated with the capture antibody (Δ). 10μg/ml human IgG was added to the medium to minimize binding of E-selectin-Ig in solution to the capture antibody on the RBC surface. Nonspecific binding recorded as human IgG control curve was used to obtain a specific adhesion probability curve () using Eq. 2. Fitting specific adhesion probability curve with Eq. 1 (solid line) returned effective binding affinity AcKa = 1.4•10-4 μm4 and koff = 0.3 s-1.
Reagent | MW (g/mol) | Concentration (mM) | Amount (g) |
---|---|---|---|
Adenine | 135.13 | 2 | 0.27 |
D-glucose (dextrose) | 180.16 | 110 | 19.82 |
D-Mannitol | 182.17 | 55 | 10.02 |
Sodium Chloride (NaCl) | 58.44 | 50 | 2.92 |
Sodium Phosphate, Dibasic (Na HPO ) | 141.95 | 20 | 2.84 |
L-glutamine | 146.15 | 10 | 1.46 |
Table 1. EAS-45 buffer preparation (1L).
25 mg biotin-XHS in 550 μl of DMF | 0.1M biotin solution |
1:10 dilution of 0.1 biotin w/DMF | 0.01M biotin solution |
1:100 dilution of 0.1 biotin w/ DMF | 0.001M biotin solution |
Table 2. Preparation of the biotin solution.
biotin final concentration (μM) | 4 | 10 | 20 | 50 | 100 | 160 |
---|---|---|---|---|---|---|
RBCs pellet stock (μl) | 10 | 10 | 10 | 10 | 10 | 10 |
1x PBS (μl) | 179.2 | 178 | 176 | 179 | 178 | 176.8 |
0.1M borate buffer (μl) | 10 | 10 | 10 | 10 | 10 | 10 |
0.01M biotin solution (μl) | 1 | 2 | 3.2 | |||
0.001M biotin solution (μl) | 0.8 | 2 | 4 |
Table 3. Biotinylation of RBCs.
To successfully use the micropipette adhesion frequency assay one should consider several critical steps. First, make sure to record the specific interaction for the receptor-ligand system of interest. Nonspecific control measurements (cf. Fig. 3, 4) ensure the specificity. Ideally, nonspecific adhesion probabilities should be below 0.05 for all contact time durations and to have a significant difference between the specific and nonspecific adhesion probabilities for each time point. Different methods could be used to couple the ligands to the RBCs surface. It was shown that chromium chloride coupling method17 gave a higher level of nonspecific binding than biotin-streptavidin coupling.
Second, an adhesion probability for specific interaction should be in the middle range. This requirement may be met by varying the densities of receptors and ligands (or only ligands on RBCs if receptors are constitutively expressed on cells the density of which is hard to change). For this purpose, when testing a new system with unknown receptor-ligand binding affinity, prepare a range of biotinylated RBCs (see Table III as an example) to test a variety of ligand densities. The steady-state adhesion probability should be no more than 0.8 on average as cell-to-cell receptor density variation usually brings the adhesion level of some cells to 1 if the average ligand density is too high. During initial testing for specificity, if some cells have adhesion levels of 0 or 1, the ligand density needs to be adjusted. Nonspecific control measurements usually follow the specific measurements and here one would like to have adhesion probability as close to zero as possible.
Third, find the correct range for the contact times. If the contact time is long enough the exp(- kofft) term in Eq. 1 goes to zero and the effective 2D affinity can be calculated from the plateau level of the adhesion curve1:
The off-rate koff determines the transition phase or how quickly the adhesion curve reaches to a plateau level. To estimate the off-rate accurately requires measurements of several points at a plateau level as well as enough time points in the transition phase. Accounting for the three critical steps described above one will obtain binding curves similar to Fig. 4.
The remaining task is to use a receptor-ligand binding kinetics model to interpret the experimentally measured binding curve and to estimate of 2D kinetic parameters from fitting the model prediction to the data. It should be noted that Eq. 1 represents a simplest model of second-order forward, first-order reverse, single-step, reversible kinetics between a single receptor-ligand species1. More complex kinetic processes have been described, including the cases of dual receptor-ligand species3,4, two-stage binding without14 or with19 trimolecular interactions, and adhesion kinetics limited by active site formation instead of bond kinetics10. In these cases, more involved mathematical models are required to relate the adhesion probability vs. contact time curve and the receptor-ligand binding kinetics.
The sustained interest in the kinetics of receptor-ligand interaction stems from the fundamental hypothesis that kinetics parameters play a role in determining the downstream signaling events inside the cell. The micropipette adhesion frequency assay presented here is one of the very few methods that allow in situ measurements of the two-dimensional (2D) binding kinetics. Two-dimensional means that both receptors and ligands are on the cell surfaces, as naturally occurs in many cell-cell interactions inside the organism. The 2D kinetic rate constants of receptor-ligand binding provide information for how rapidly cells bind to each other or to the extracellular matrix, how long they remain bound, and how many bonds will form. By comparison, in the Surface Plasmon Resonance (SPR) method23 one of the interacting molecules is in the fluid phase, hence called three-dimensional (3D) binding. Because both interacting molecules are purified and isolated from the cellular environment, the kinetic parameters obtained in 3D measurement could be drastically different from those obtained in 2D measurements even for the same receptor-ligand pair17.
The adhesion frequency method analyzes 2D kinetics on living cell membrane and thus provides an opportunity for one to analyze the biophysical and biochemical regulations of the cellular environment. These include the membrane microtopology5, membrane anchor2, molecular orientation and length6, carrier stiffness9 and curvature20, impingement force20, and modulators of the cytoskeleton and membrane organization where the interacting molecules reside15,17.
Because cross-junctional receptor-ligand interaction requires direct physical contact between two cells and results in physical linkage between two cells, the chemical reaction kinetics of molecular interaction can be analyzed by a mechanical assay that puts the cells in contact and detects binding by the effect of force. Although we exemplified the adhesion frequency assay using a micropipette-aspirated RBC as an adhesion sensor, other force techniques can be used, including atomic force microscopy24, biomembrane force probe8,17, optical tweezers25, and the integrated micropipette and cantilever26.
Other mechanically-based 2D assays have been developed. These include the thermal fluctuation assay8, centrifugation assay27,28, rosetting assay29, and flow chamber assay30,31.
The limitation of the adhesion frequency assay is the slow and labor-intensive nature of the assay due to the repeated serial cycles with a single pair of cells tested one contact at a time. It becomes difficult for receptor-ligand interactions with slow off-rates because long contact times would be required for the binding curve to reach steady-state, making the experiment inhibitively long.
The force transducer constructed by a micropipette-aspirated RBC is capable fo detecting piconewton-level forces, which is an order of magnitude lower than the typical strength of a noncovalent receptor-ligand bond1,32. However, receptor-ligand dissociation could occur even at zero forces. Any weak adhesion that goes undetected leads to an underestimation of the binding affinity and on-rate1.
The adhesion frequency method assumes that each adhesion test is identical and independent from the others. This requirement could be violated as was shown for some systems33, where current adhesion increased or decreased the probability of the next adhesion. The Matlab code for checking if the requirement is met for the recorded sequence of adhesion events is available upon request.
The authors have nothing to disclose.
This study was supported by NIH grants R01HL091020, R01HL093723, R01AI077343, and R01GM096187.
Name of the reagent | Company | Catalogue # | Comments |
---|---|---|---|
10x PBS | BioWhittaker | 17-517Q |
Dilute to 1x with deionized water prior to use |
Vacutainer EDTA | BD | 366643 | RBCs isolation |
10ML PK100 | |||
Histopaque 1077 | Sigma-Aldrich | 10771 | RBCs isolation |
Adenine | Sigma-Aldrich | A2786 | EAS-45 preparation |
D-glucose (dextrose) | Sigma-Aldrich | G7528 | EAS-45 preparation |
D-Mannitol | Sigma-Aldrich | 6360 | EAS-45 preparation |
Sodium Chloride (NaCl) | Sigma-Aldrich | S7653 | EAS-45 preparation |
Sodium Phosphate, Dibasic (Na₂HPO₄) | Fisher Scientific | S374 | EAS-45 preparation |
L-glutamine | Sigma-Aldrich | G5763 | EAS-45 preparation |
Biotin-X-NHS | Calbiochem | 203188 | RBCs biotinylation |
Dimethylformamide (DMF) | Thermo Scientific | 20673 | RBCs biotinylation |
Borate Buffer (0.1M) | Electron Microscopy Sciences | 11455-90 | RBCs biotinylation |
Streptavidin | Thermo Scientific | 21125 | Ligand functionalizing |
BSA | Sigma-Aldrich | A0336 | Ligand functionalizing |
Quantibrite PE Beads | BD Biosciences | 340495 | Density quantification |
Flow cytometer | BD Immunocytometry Systems | BD LSR II |
Density quantification |
Capillary Tube 0.7-1.0mm x 30" |
Kimble Glass Inc. | 46485-1 | Micropipette pulling |
Mineral Oil | Fisher Scientific | BP2629-1 | Chamber assembly |
Microscope Cover Glass | Fisher Scientific | 12-544-G | Chamber assembly |
PE α-human CD11a Clone HI 111 |
eBioscience | 12-0119-71 | Reagent for Fig.1 |
PE anti-human CD54 | eBioscience | 12-0549 | Reagent for Fig.1 |
Mouse IgG1 Isotype Control PE | eBioscience | 12-4714 | Reagent for Fig.1 |
hydraulic micromanipulator | Narishige | MO-303 | Micropipette system |
Mechanical manipulator | Newport | 461-xyz-m, SM-13, DM-13 | Micropipette system |
piezoelectric translator | Physik Instrumente | P-840 | Micropipette system |
LabVIEW | National Instruments | Version 8.6 | Micropipette system |
DAQ board | National Instruments | USB-6008 | Micropipette system |
Optical table | Kinetics Systems | 5200 Series | Micropipette system |