Fast dissociation kinetics between individual E-cadherin fragments revealed by flow chamber analysis

Abstract

E-cadherin is the predominant adhesion molecule of epithelia. The interaction between extracellular segments of E-cadherin in the membrane of opposing cells is homophilic and calcium dependent. Whereas it is widely accepted that the specificity of the adhesive interaction is localized to the N-terminal domain, the kinetics of the recognition process are unknown. We report the first quantitative data describing the dissociation kinetics of individual E-cadherin interactions. Aggregation assays indicate that the two outermost domains of E-cadherin (E/EC1-2) retain biological activity when chemically immobilized on glass beads. Cadherin fragment trans-interaction was analysed using a flow chamber technique. Transient tethers had first-order kinetics, suggesting a unimolecular interaction. The unstressed lifetime of individual E-cadherin interactions was as brief as 2 s. A fast off rate and the low tensile strength of the E-cadherin bond may be necessary to support the high selectivity and plasticity of epithelial cell interactions.

Fig. 1. Diagram of a flow chamber. I, inlet; O, outlet; 1, chamber cavity; 2, mica sheet forming the chamber floor; 3, steel plate; S, screw; G, toric gasket.

Fig. 2. Characterization of E/EC1–2 and E/W2A fragments. (A) Coomassie Blue stain of a 15% acrylamide gel showing lysates from transformed BL21(DE3) bacteria, with (lane 1) or without (lane 2) IPTG induction, and purified E/EC1–2 (lane 3) and E/W2A (lane 4) fragments. (B) Western blot detection of purified E/EC1–2 and E/W2A fragments using a polyhistidine antibody. (C) To test protein folding, both fragments were trypsin digested in the presence of Ca²⁺ or EDTA for the indicated times.

Fig. 3. Specific E/EC1–2 bead self-aggregation. (A) Representative field of E/EC1–2-coated beads in the presence of CaCl₂ or EDTA (bar = 5 µm). (B) Beads were allowed to aggregate in the presence of Ca²⁺ or EDTA. The aggregation index was high for E/EC1–2 beads, but not for E/W2A or control beads (casein or BSA). E/EC1–2 bead aggregation was Ca²⁺ dependent. Data are mean values of at least two independent experiments (∼4500 beads counted for each condition).

Fig. 4. Specific binding of E/EC1–2 beads to E-cadherin-expressing cells. (A) Representative view of E-RBL cells following incubation with E/W2A or E/EC1–2 beads (bar = 5 µm). (B) E-RBL cells were incubated with E/EC1–2- or E/W2A-coated beads or with control beads. The numbers of cells with bound beads were recorded as a percentage of the number of total cells in the field. Data show the result of one experiment (at least 600 cells for each condition). Nearly 70% of E-RBL cells bound E/EC1–2 beads; only 30% of the cells bound E/W2A or control beads.

Fig. 5. Characterization of the flow chamber interfaces. (A) Schematic representation. Coated beads under hydrodynamic flow can interact with fragments fixed and oriented on mica (bottom of the chamber) by the histidine tag. (B) Three-dimensional representation of confocal images of mica sheets covered with Alexa Fluor^® 488–E/EC1–2 fragments. Fluorescence intensities for two E/EC1–2 densities are reported. For coating solution diluted 1/1000 (leading to a site density of 50–100 E/EC1–2 molecules/µm²) or more, the distribution was homogeneous. The scale represents fluorescence intensity in arbitrary units (au) (bar = 10 µm).

Fig. 6. A typical trajectory. The position of an E/EC1–2-coated bead driven along an E/EC1–2-coated surface is shown (top curve). The bead exhibited three arrests of 0.78, 1.24 and 0.38 s, respectively. The bead velocity was 6.84 µm/s before the first arrest, 6.6 µm/s between the first and the second arrest and 6 µm/s between the second and the third arrest. The bottom curve corresponds to the bead area.

Fig. 7. Adhesion specificity of E/EC1–2 fragments in the flow chamber. Mean values ± SDs for at least four independent experiments are reported. The tryptophan side chain analogue I3A inhibits E/EC1–2 interaction to values similar to those for E/W2A-coated beads or using an irrelevant protein.

Fig. 8. Dissociation kinetics and dependence of dissociation rate on site density. (A) The motion of E/EC1–2-coated spheres along E/EC1–2-coated surfaces, and the duration of binding events were recorded. The number of bound particles was plotted against time after the initial stop. As shown for a representative example, the curve was linear on the time interval (0–1 s), consistent with first-order dissociation kinetics. (B) Mica surfaces were coated with various dilutions of an E/EC1–2 coating solution and used for dynamic study of interaction with E/EC1–2-coated spheres (wall shear rate 8/s). Each data point corresponds to 100–150 trajectory measurements; error bars are within the size of data points. At high protein density on mica (dilution 1/250), long arrest durations were observed (>3 s) corresponding to multiple binding interactions. At dilutions of 1/1000 (a site density of 50–100 E/EC1–2 molecules/µm²) or more, the dissociation rate was constant, indicating measurements of interactions between single E/EC1–2 fragments for binding site densities of 50–100 to 5–10 molecules/µm².

Fig. 9. Effect of the wall shear rate on the dissociation kinetics of the E/EC1–2 interaction. The dissociation rate of bead–surface attachments was determined for different values of wall shear rate and for site densities from 5–10 to 50–100 molecules/µm². The force applied on bonds responsible for particle arrest was calculated as described in Materials and methods. Results are consistent with an exponential dependence of k_off on the tension applied on tethers.