Monomer dimer dynamics and distribution of GPI-anchored uPAR are determined by cell surface protein assemblies

Abstract

To search for functional links between glycosylphosphatidylinositol (GPI) protein monomer-oligomer exchange and membrane dynamics and confinement, we studied urokinase plasminogen activator (uPA) receptor (uPAR), a GPI receptor involved in the regulation of cell adhesion, migration, and proliferation. Using a functionally active fluorescent protein-uPAR in live cells, we analyzed the effect that extracellular matrix proteins and uPAR ligands have on uPAR dynamics and dimerization at the cell membrane. Vitronectin directs the recruitment of dimers and slows down the diffusion of the receptors at the basal membrane. The commitment to uPA-plasminogen activator inhibitor type 1-mediated endocytosis and recycling modifies uPAR diffusion and induces an exchange between uPAR monomers and dimers. This exchange is fully reversible. The data demonstrate that cell surface protein assemblies are important in regulating the dynamics and localization of uPAR at the cell membrane and the exchange of monomers and dimers. These results also provide a strong rationale for dynamic studies of GPI-anchored molecules in live cells at steady state and in the absence of cross-linker/clustering agents.

Figure 1.

Distribution and diffusion of cell surface uPAR-G in serum-plated HEK293 cells. (A, left) Distribution of uPAR-G in apical membranes of live HEK293 cells. The magnified image of the boxed area shows two regions (positions 1 and 2) in which fluorescence intensity traces were acquired (bottom right). The scheme (top right) illustrates the typical two-photon volume selected for FCS measurements. (B, left) Distribution of uPAR-G in the basal membranes of the cells shown in A. The magnified image of the boxed area illustrates the region (position 3) in which the third FCS measurement of this example was taken. The trace acquired in position 3 (bottom right) shows an initial photobleaching (red rectangle subset). The diffusion coefficient of the residual mobile receptors was derived from the data at plateau (black rectangle subset). (C) Normalized autocorrelation functions (ACFs) derived from the fluorescence intensity traces acquired in position 2 in A and in position 3 in B. Green lines, curves fitted according to Eq. 1 (Fig. S2, available at http://www.jcb.org/cgi/content/full/jcb.200702151/DC1). (D) Diffusion coefficients and anomality coefficients (α; inset) of uPAR-G in apical and basal membranes. Box-whisker plots show minimum, 25% percentile, median, 75% percentile, and maximum values. Three to four ACFs per cell were analyzed. The number of independent experiments (n) and statistical significance are indicated. Cells were seeded in serum-containing medium at 37°C, and fluorescence images and FCS measurements were recorded at 27°C 24–48 h later. kcps, kilocounts per second.

Figure 2.

Distribution and diffusion of cell surface uPAR-G in HEK293 cells adhered to Vn or Fn matrices. (A and D) Distribution of uPAR-G in apical and basal membranes of HEK293 cells seeded on Vn (A) and Fn (D) matrices. (B and E) Representative normalized autocorrelation functions (ACFs) in one apical and one basal region are shown in B for cells on Vn and in E on Fn matrices. Green lines, curves fitted according to Eq. 1 (Fig. S2, available at http://www.jcb.org/cgi/content/full/jcb.200702151/DC1). (C and F) Diffusion coefficients and anomality coefficients (α; inset) of uPAR-G in apical and basal membranes are shown in C for cells on Vn and in F on Fn matrices. Box-whisker plots show minimum, 25% percentile, median, 75% percentile, and maximum values. Three to four ACFs per cell were analyzed. The number of independent experiments (n) and statistical significance are indicated. Cells were seeded in serum-free medium on dishes coated with 2.5 μg/ml Vn or 10 μg/ml Fn and were allowed to recover for 2–4 h at 37°C before the recording of fluorescence images and FCS at 27°C. kcps, kilocounts per second.

Figure 3.

Fluorescence lifetime analysis of HEK293/G cells by Phasor-FLIM. (A) The universal circle of the phasor representation showing a single lifetime phasor (position a) and a phasor arising from a complex lifetime decay (position b; see Materials and methods for the phasor transformation of fluorescence lifetime decays). (B) Representative phasor distributions. TCSPC-transformed decay data at each pixel are shown in contour plots for the basal membranes of two HEK293/G cells at high (position 1) and low (position 1′) uPAR-G expression. The basal membrane of an untransfected HEK293 cell (cell autofluorescence) is represented by the contour plot in position 2. HEK293/R cells behave as untransfected HEK293 under identical experimental conditions (not depicted). The green line is the trajectory joining the mean values of the phasor distributions in HEK293/G at different contributions of cell autofluorescence. The black lines are the calculated FRET trajectories describing all possible positions of the uPAR-G phasors in the presence of quenching as a result of FRET for the basal membranes imaged in positions 1 and 1′ having a cell autofluorescence contribution of 1% and 37%, respectively. The FRET trajectories are computed using the phasor of the donor in the absence of FRET (positions 1 and 1′) and the phasor of the background (position 2), which were determined independently on untransfected cells, and applying the classic definition of FRET efficiency, [1 − (τ_{donor-acceptor})/τ_donor], to determine the phasor corresponding to the quenched donor. All possible phasors that are quenched with different FRET efficiencies (from 0 to 100%) describe a curved trajectory (black line) in the phasor plot. The experimental position of the phasor of a given sample along the trajectory determines the amount of quenching and, therefore, the FRET efficiency. The contributions of the background and of the donor without acceptor is evaluated using the rule of the linear combination with the background phasor (which was determined independently) and the donor unquenched (which was also determined independently). As an example, the circles 3 and 3′ mark the positions corresponding to the phasor distributions 1 and 1′ with 50% FRET quenching. Thus, the red line marks the position of uPAR-G phasors at 50% FRET and having all possible cell autofluorescence contributions (0–100%).

Figure 4.

Distribution of uPAR-GR dimers at the apical and basal membranes of HEK293/GR cells seeded in the presence of serum. (A and B) A representative phasor-FLIM experiment on a HEK293/G cell imaged at the basal (A) and apical (B) membranes shows the fluorescence intensity image, the phasor plot, and localization in the FLIM image (pink masks) of the pixels comprised in the selected area (black circles) of the phasor plot. The black circles in this example select 98% of the total pixels of the image. The fluorescence lifetime of uPAR-G does not depend on its local concentration, as pixels with high and low intensity are included in the phasor selection. (C and D) The parallel experiment on a HEK293/GR cell is reported in C for the basal and D for the apical membranes. Two phasor subsets are shown in the phasor plots, and the correspondent pixels are localized in the FLIM images: pixels included in the 8–24% (top) and <8% (bottom) FRET efficiency ranges. Cells were seeded in serum-rich medium and kept at 37°C and 5% CO₂ for 48 h. Fluorescence images and FLIM measurements were performed at 27°C.

Figure 5.

Distribution of uPAR-GR dimers at the apical and basal membranes of HEK293/GR cells seeded on Vn or Fn matrices. (A–D) Representative phasor-FLIM experiments on HEK293/GR cells on Vn matrix (A and B) and on Fn matrix (C and D). Each panel shows the fluorescence intensity image and two phasor subsets (black circles in the phasor plots). The correspondent FLIM images illustrate the localization of the pixels selected in each subset: 8–24% (top) and <8% (bottom) FRET efficiency ranges. Cells were seeded in serum-free medium on dishes coated with 2.5 μg/ml Vn or 10 μg/ml Fn and allowed to recover for 2–4 h at 37°C before recording fluorescence images and FLIM at 27°C.

Figure 6.

Statistical analysis of mean FRET efficiencies in HEK293/GR cells. (A and B) Mean FRET efficiencies derived from the phasor distributions acquired in basal and apical membranes of HEK293/GR cells grown in serum-containing medium (A) or seeded on purified Vn or Fn (B) matrices. The box-whisker plots show minimum, 25% percentile, median, 75% percentile, and maximum values. The number (n) of independent experiments is indicated. Statistical significance is either indicated in the plot or reported in Table S1.

Figure 7.

Loss of uPAR-GR dimers in HEK293/GR cells exposed to uPA–PAI1. (A and B) Representative phasor-FLIM experiment on a HEK293/GR cell, which was grown and maintained in serum-rich medium, exposed to 8 nM uPA–PAI1 for 30 min at 37°C and imaged at 27°C on the basal (A) and apical (B) membranes. Each panel shows the fluorescence intensity image and the phasor plot in which the black circles select phasor subsets of <8% FRET efficiency. The black circles in this example include 95% of pixels of the images. The correspondent FLIM panels show the localization of these pixels in the images (pink masks). (C) The mean FRET efficiency measured in HEK293/G exposed to uPA–PAI1 is reported. The box-whisker plots show minimum, 25% percentile, median, 75% percentile, and maximum values. The data in the absence of uPA–PAI1 are shown for comparison and are reproduced from Fig. 6 A. The number of independent measurements (n) and statistical significance are shown.

Figure 8.

Loss of uPAR-G dimers in HEK293/G cells exposed to uPA–PAI1 and diffusion of the uPAR-G/uPA–PAI1 complexes at the cell surface. (A) Plot of the relative frequency of the brightness obtained by local PCH analysis in apical membranes of HEK293/G cells exposed to uPA–PAI1 as described in Fig. 7. Brightness was derived taking intervals of 2.5 s on the fluorescence fluctuation traces (see Materials and methods; Chen et al., 2002). Each curve is an independent measurement (i.e., a fluorescence intensity trace; n = 30). (inset) Mean brightness of monomeric EGFP in solution as a function of the concentration expressed as the mean number of molecules (Chen et al., 1999). Each EGFP solution was tested in triplicate. The mean brightness of monomeric EGFP in solution was 4,746 cpsm (minimum 4,077 cpsm, maximum 5,210 cpsm). (B) Plot of the relative frequency of brightness in the absence of uPA–PAI1. The brightness was determined as in A. Each curve is an independent measurement (i.e., a fluorescence intensity trace; n = 32). (C) Representative normalized autocorrelation functions (ACFs) in apical regions of HEK293/G cells before (black line) and after (green line) exposure to 8 nM uPA–PAI1 for 30 min at 37°C. Red lines, curves fitted according to Eq. 1 (Fig. S2, available at http://www.jcb.org/cgi/content/full/jcb.200702151/DC1). (D) Diffusion coefficients and anomality coefficients (α, inset) of uPAR-G in apical membranes of HEK293/G cells exposed to uPA–PAI1. Box-whisker plots indicate minimum, 25% percentile, median, 75% percentile, and maximum values. Three to four ACFs per cell were analyzed. The number of total measurements (n) and statistical significance are shown. The data from HEK293/G not exposed to uPA–PAI1 are shown for comparison and are reproduced from Fig. 1 D. Cells were grown and maintained in serum-rich medium and imaged at 27°C.

Figure 9.

Recovery of uPAR-GR dimers in HEK293/GR cells after endocytosis and recycling. (A and B) Representative phasor-FLIM experiment on the basal (A) and apical (B) membrane of a HEK293/GR cell grown on serum-coating exposed to uPA–PAI1 for 30 min and then left to recycle for 2 h at 37°C in fresh, serum-rich medium lacking uPA–PAI1 before imaging at 27°C. Each panel shows the fluorescence intensity image and two phasor subsets (black circles in the phasor plots). The correspondent FLIM images illustrate the localization of the pixels selected in each subset: 8–24% (top) and <8% (bottom) FRET efficiency. (C) The mean FRET efficiency measured in HEK293/G after recycling is reported. The box-whisker plots show minimum, 25% percentile, median, 75% percentile, and maximum values. The numbers (n) of independent measurements and statistical significance are indicated. The data in the presence of uPA–PAI1 (i.e., in the absence of recycling) are shown for comparison and are reproduced from Fig. 7 C.

Figure 10.

uPAR monomer–dimer dynamics on the cell surface. (A) At steady state, Vn in the extracellular matrix (left) recruits and stabilizes uPAR dimers in the basal membrane. uPAR dimers are the dominant form in the basal membrane. As a result of the interaction with Vn, 40–50% of uPAR is immobile, and the rest diffuses more slowly and more anomalously than in apical sides. Receptors not engaged in matrix contact (i.e., receptors in the apical membrane) are in equilibrium between dimeric and monomeric forms, the latter being the most prominent; both forms display fast diffusion in this membrane side. When cells are seeded on Fn (right), a different picture is observed: there is no polarized distribution of uPAR between the basal and apical membranes, dimers are generally rare, and fast, lateral diffusion is observed throughout the cell membrane. (B) Commitment to endocytosis (left) through uPA–PAI1, binding depletes dimeric uPAR from the cell surface and destabilizes adhesion to Vn. The monomerization induced by commitment to endocytosis is reversible, and recycling is followed by the de novo formation of uPAR dimers and reestablishes the firm adhesion (right).