(Received for publication, February 11, 1997)
From the Center for Immunology and the Department of Pathology, Washington University School of Medicine, Box 8118, St. Louis, Missouri 63110
The interaction of the T cell glycoprotein CD2
and its ligand CD58 is important for T cell interaction with
antigen-presenting and target cells. The binding interaction is of low
affinity and has a fast off-rate (>5 s1) in
solution. However, solution measurements may not accurately predict the
behavior of molecules in an adhesive contact area. Interaction between
T cells that express CD2 and glass-supported planar bilayers containing
purified and fluorescently labeled CD58 leads to accumulation of CD58
(fluorescence) in the cell/bilayer contact area. CD58 molecules
accumulated within the contact area in excess of the CD58 density in
the bilayer outside the contact area can be considered as bound by cell
surface CD2. Here, this phenomena and fluorescence photobleaching
recovery were utilized to determine whether CD2-CD58 bonds are
transient in contact areas. Fluorescent CD58 molecules accumulated in
the T cell-bilayer interface were completely bleached. The bleached
CD58 molecules accumulated in the contact area were rapidly replaced by
fluorescent CD58 that diffused into the contact area from adjacent
bilayer regions outside the contact area. Rapid recovery of the
accumulated fluorescence directly demonstrates that the CD2-CD58 bonds
are dissociating and that the dissociation leads to partner exchange,
rather than rebinding of the same CD2-CD58 pairs. This suggests that
the solution off-rate provides an accurate description of CD2-CD58
interaction in contact areas. Accumulated fluorescent IgG in contacts
between K562 cells expressing low affinity Fc receptors and planar
bilayers with fluorescent IgG bound to hapten-derivitized phospholipids displayed slower recovery than CD58 by a factor of 10. This suggests that the Fc receptor-IgG interaction has a longer lifetime than the
CD2-CD58 interaction. These findings have implications for the
mechanism of signaling by CD2 and the mechanism of cell detachment from
large numbers of transient interactions.
Specific binding of membrane receptors across a small gap between neighboring cells is an exquisitely specific mode of communication. In the immune response, the fundamental regulatory and effector functions of T lymphocytes require contact with antigen presenting cells or target cells, respectively (1, 2). Recent studies on the molecular basis of these cell-cell interactions emphasize the role of low affinity, transient receptor-ligand interactions, both for cell-cell adhesion and direct cell-cell signaling functions (3, 4). However, these measurements have been made in solution and may not accurately predict the behavior of molecules in contact areas. Therefore, it is important to directly determine the properties of transient molecular interactions in contact areas.
The T lymphocyte glycoprotein CD2 and its ligand CD58 play an important
role in stabilizing adhesion between T cells and antigen-presenting or
target cells (3, 5). CD2 and CD58 are closely related members of the
immunoglobulin superfamily (5). The accumulation of CD58 in contacts
between T cells and bilayers containing fluorescently labeled CD58 has
been used to determine the density of CD2-CD58 complexes as a function
of free CD58 density. The two-dimensional Kd
calculated from this data is 21 molecules/µm2 (6). This
suggests that CD2-CD58 interactions operate efficiently at
physiological densities of CD2 and CD58 that exceed 100 molecules/µm2. The solution Kd for the
CD2-CD58 interaction is 5 µM, and the off-rate is >5
s1 (3). Since the three-dimensional and two-dimensional
Kd values for the interaction of CD2 and CD58 have
been determined under similar conditions, it is possible to calculate
the third dimension for the interaction: the third dimension is
calculated at 5 nm (6). This is the size of a single Ig domain. We
would interpret this number as the range of oscillations in the
distance between membranes and in adhesion molecule tilting that
generate the third dimension for the interaction. When these bonds
dissociate there may be a significant probability of rebinding of the
same pair of molecules, since they are not immediately removed from each other's presence except by lateral diffusion (7). Therefore, the
effective off-rate of CD2-CD58 bonds, the rate at which pairs of
molecules actually diffuse free of each other, may be much lower than
the intrinsic off-rate deduced in solution from the kinetic
off-rate.
The CD2-CD58 system provides an opportunity to test the hypothesis that low affinity bonds are effectively transient in contact areas. In the CD2-CD58 system, the interaction can drive striking accumulation of CD58 in the contact area between a cell and an artificial bilayer containing fluorescently labeled CD58, which is observed by fluorescence microscopy. Here, a brief laser pulse was used to eliminate CD58 fluorescence in the entire cell-bilayer contact area, and digital fluorescence imaging was used to monitor the spatially resolved fluorescence intensity prior to and at various times after the laser pulse. If the bonds are transient and the CD2 and CD58 molecules exchange partners rapidly, the fluorescence in the contact area should recover rapidly. If the bonds are transient but dissociated pairs rebound rapidly without partner exchange, or if the bonds became stable through a contact-specific mechanism, then the recovery of fluorescence would be slow or nonexistent. The results show that bond turnover is rapid at low and high free CD58 densities and in nonactivated and activated cells. The low affinity Fc receptor on K562 cells displayed slower recovery, suggesting a longer effective bond lifetime, which is consistent with solution measurements. These results demonstrate that the rapid off-rate of the CD2-CD58 bond is evident in contact areas with constrained diffusion of molecules
The TS2/9, TS2/18, and Jurkat cell lines were obtained from ATCC. The CD2.1 antibody was a gift of D. Olive or was purchased from Immunotech (Westbrook, ME). Fluorescein isothiocyanate (FITC)1 was purchased from Molecular Probes (Eugene, OR). Other reagents were obtained from Sigma or Fisher.
Microscope ConfigurationThe optical path is diagrammed in
Fig. 1. A 1-mm beam from a 2-watt argon laser (Coherent, Palo Alto, CA)
tuned to the 488-nm line was expanded with a 25-mm focal length
plano-convex lens (25.2-mm diameter). The diverging beam was then
focused through one 400-mm focal length acromat (40-mm diameter) and
one 500-mm focal length acromat (63-mm diameter) that were positioned
on a rail such that they could be moved independently along the axis of
the beam. Movement of these lenses changed the focal point of the laser
beam with respect to the sample. The focused beam passed through a
45° beam splitter that was used to merge the laser and xenon arc lamp
light paths. The xenon arc lamp output was focused by a quartz
condenser. Both of these beams pass through an adjustable aperture that
acted as a field diaphragm for the arc lamp, the outline of which was
imaged on the sample plane when the objective was focused at the
coverslip/media interface. The field diaphragm did not interfere with
the laser transmission, since it was focused to a small point at this
position. The final lens element before the objective was a 200-mm
focal length acromat that was positioned 100 mm from the dichroic
mirror used for epi-illumination. The images were formed with Zeiss
(Thornwood, NY) infinity-corrected optics on a cooled CCD chip (Eastman
Kodak Co. KAF1400, grade 2) of a Photometrics (Tuscon, AZ) PXL camera.
The excitation filter, dichroic mirror, and emission filters were
independently selected with filter wheels and single mirror sliders.
For most experiments, the dichroic mirror was a 505-nm-long pass, and
the emission filter was a 530-nm band pass. Fluorescence excitation was
performed with a 485-nm band pass filter. Interference-reflection
images were performed with 530-nm excitation and neutral density
filters. Although the 530-nm light was not efficiently reflected by the dichroic mirror, more than enough light for the CCD was reflected, and
neutral density filters of 1-2 OD units were required on the excitation path to obtain usable exposure times. Images were acquired directly into a Macintosh computer using IP lab software (Signal Analytics, Vienna, VA). Fluorescence images were usually acquired with
10 × 10 camera pixels/image pixel (10 × 10 binning) to
increase sensitivity and to allow image acquisition without significant photobleaching. A camera pixel was 68 nm square with × 100 objective. IP lab was also used to control the excitation filter wheel
and shutter timing through a serial port-linked control box (Ludl, Hawthorne, NY). The microscope system was designed in collaboration with and assembled by Yona Microscopes (Silver Spring, MD).
Confocal Microscopy
Cells were fixed after adhesion to bilayers with 1% paraformaldehyde, and slides were mounted under a solution of 0.2% N-propylgallate in phosphate-buffered saline. Confocal images were obtained with a Molecular Dynamics (Sunnyvale, CA) Confocal system interfaced with a Zeiss Axioscope.
Purification of CD58 and Fluorescent LabelingCD58 was purified from human erythrocytes by immunoaffinity chromatography on TS2/9 IgG linked to CNBr-activated Sepharose CL-4B (Pharmacia Biotech Inc.) (8). The purified protein was bound again to TS2/9-Sepharose CL-4B and labeled with 8 mM FITC in bicarbonate buffer pH 9 (6). The labeled protein was eluted at pH 3 and reconstituted into unilamellar liposomes by detergent dialysis (9). Glass-supported planar bilayers were then formed by incubating the liposome suspensions (0.4 mM) with clean glass coverslips (10). The density of molecules in the resulting bilayers was determined by radiometric binding assays with iodinated TS2/9 IgG of known specific activity. Radiometric assays were done with bilayers formed on one side of a 12-mm, round coverslip in wells of 24-well cluster plates. Prior to elution of bound iodine, the coverslips were transferred under buffer to a clean 24-well plate to avoid detection of liposomes bound to the plastic well.
Observation of CD58 Redistribution in BilayersBilayers
containing FITC-CD58 or unlabeled CD58 were formed on glass coverslips
in a parallel plate flow cells (Bioptechs, Butler, PA). The bilayers
were blocked with bovine serum albumin, and the experiments were
performed in Hepes-buffered saline (10 mM Hepes, pH 7.4, 137 mM NaCl, 5 mM KCl, 0.5 mM
Na2PO4, 5 mM D-glucose, 1% bovine serum albumin). The cells were allowed to settle on the
bilayers and were observed to form contacts over a period of 20 min as
described previously (6). When images were acquired with the × 100 objective and 2 × 2 binning, a bleaching correction of 0.8 was applied to allow comparison of images acquired over time. Images
acquired with 10 × 10 binning could be acquired sequentially without significant bleaching. Flat field images were acquired by
averaging together 10 images of a uniform 2%
NBD-phosphatidylethanolamine, 98% egg phosphatidylcholine planar
bilayer. The fluorescence spectrum of NBD overlaps with that of
fluorescein, and this has been the fluorophore of choice for labeling
phospholipids due to its small size. The flat field and bias images (no
excitation light) were used to correct the experimental images:
corrected = ((experimental bias)/(flat field
bias)) × average intensity of flat field.
The laser spot size was adjusted using NBD-containing bilayers either to image the beam or to bleach holes in the bilayer. The spot size was adjusted so that it would bleach an object the size of a cell/bilayer contact area (50-100 µm2). The duration of bleaching and the laser power were determined empirically to be the lowest necessary to achieve complete elimination of accumulated fluorescence. Typically, the laser output at 488 nm was 400-500 milliwatts, and bleach duration of 20-50 ms was used. Prebleach and a series of timed postbleach images were acquired with 10 × 10 binning. IRM images were acquired before and after the series and were sometimes acquired during the recovery period to determine whether the bleach pulse altered the contact area. The bleaching and recovery data were expressed relative to Ib, the specific intensity of the fluorescent adhesion molecules in the bilayer. This allows comparison of bleaching data where the density of free CD58 in the bilayers varies over a 20-fold range.
Image AnalysisThe bright fluorescent contact areas in the prebleach images were identified using the segmentation feature of IP-lab. This mask was then used to analyze the prebleach and recovery images in which the targeted cell was initially nonfluorescent. The contact area intensities were measured at different time points. The bilayer fluorescence intensities were extracted by comparison with images of unlabeled bilayers acquired with the same settings. Cellular autofluorescence was variable but was minimized when the Jurkat cells were maintained in log phase growth. The autofluorescence on each day and at different times during the course of experiments was monitored by imaging cells on bilayers with unlabeled CD58. The autofluorescence was found to be partially bleached and did not recover over 550 s. Therefore, an appropriate autofluorescence correction was applied to the data acquired with fluorescent CD58 based on the behavior of autofluorescence on that day using the same bleach and imaging parameters.
An imaging system was designed to perform FPR experiments (11) with a large bleach spot to eliminate the fluorescence in an entire cell/substrate contact area of up to 100 µm2. The microscope system used for digital imaging FPR is similar to the system utilized by Kapitza et al. (12) for video FPR. The beam from a 2-watt argon laser was expanded and focused on the sample plane. The focal point of the laser could be shifted with respect to the image plane so that the diverging (defocused) beam could be used for bleaching (13). The diverging beam is still gaussian on average but contains interference patterns. The interference patterns did not alter the effectiveness of bleaching (not shown). The laser beam was merged with the output from a xenon arc lamp using a beam splitter (Fig. 1). A field diaphragm could be used without interfering with the laser beam intensity. This was important for use of interference reflection microscopy, where the field diaphragm is used to restrict scattered light (14). The laser beam and arc lamp output were independently controlled by fast electronic shutters. The sample was imaged using wide field epi-illumination and the cooled CCD camera.
The bleaching system was characterized using the NBD-PE-containing
planar bilayers. The diffusion coefficient for NBD-PE in an egg
PC-based planar bilayer is 1 µm2/s (6). Fig.
2a shows images of an experiment with NBD-PE
in an egg PC-based planar bilayer. The half-time for recovery for small, medium, and large bleach spots were <3, 16, and 24 s,
respectively (Fig. 2b). The effective beam radii
(w), calculated from the relationship t1/2 = w2/4D,
where D is the diffusion coefficient (15), were 3.5, 8, and
10 µm for these bleach spots. These are estimates, since monitoring was performed with wide field illumination rather than a gaussian beam
laser. However, these estimated beam radii were in agreement with the
size of the dark central areas of the bleach spots. Therefore, this
system could be used to measure relative changes in two-dimensional diffusion in larger regions than classical FPR.
The CD2-CD58 interaction occurs through a complementary interaction of
two charged faces of the membrane distal Ig domain. The experiments
below required that the lysyl residues in the binding interface be
protected from modification with fluorophores. Therefore, CD58 was
labeled with fluorescein isothiocyanate (through lysyl residues) while
bound to the adhesion-blocking antibody TS2/9. This was previously
shown to result in full retention of activity in adhesion assays (6).
The FITC-CD58 was reconstituted into small unilaminar liposomes of egg
PC by detergent dialysis and matched liposomes of egg PC only were
prepared in parallel. Glass-supported planar bilayers were formed by
incubating the liposomes suspensions with clean glass coverslips (10).
The site density of CD58 was measured by radioimmunoassay with the TS2/9 antibody. It was found that different uniform densities of CD58
could be achieved by diluting the FITC-CD58-containing egg PC liposomes
with matched egg PC liposomes and then forming the planar bilayers as
usual. The site density measure by radiometric assay was linearly
related to the proportion of the CD58 liposomes in the mixture (Fig.
3). The resulting bilayers show uniform fluorescence. Apparently, the CD58 liposomes and the CD58 free liposomes have an
equal potential to fuse to form planar bilayers, and after fusion, the
CD58 redistributes by diffusion. This method was used to prepare
bilayers with 750 and 35 molecules/µm2 of CD58. Large
spot photobleaching on the CD58 bilayers revealed a diffusion
coefficient similar to that for NBD-PE, consistent with the concept
that glycolipid-anchored CD58 interacts with the bilayer in the manner
of a phospholipid (not shown).
The human T lymphoma cell line Jurkat has a high level of CD2
expression (70,000-80,000 CD2 molecules/cell). Jurkat cells adhered
efficiently to bilayers with 750 molecules/µm2 of CD58
and with somewhat reduced efficiency to bilayers with 35 molecules/µm2. Adhesive contact areas that were
detectable by interference reflection microscopy rapidly accumulated
FITC-CD58 molecules based on fluorescence imaging. Contacts initially
expanded in area and often included a main central contact and
satellite contacts that appeared to be formed by radially arranged
membrane protrusions. These satellites of fluorescent CD58 frequently
merged into the main contact area after 20 min at room temperature
(Fig. 4). After 20 min, the shape of the contacts did
not change.
Cell surface CD2 is laterally mobile and would also be expected to
redistribute into contact areas (16). CD2 redistribution was followed
using the nonadhesion-blocking anti-CD2 mAb, CD2.1 (6F10) (17). CD2.1
was labeled with FITC. Jurkat cells were preincubated with FITC-CD2.1
to saturate all CD2 sites and then were washed and incubated on
CD11a-CD18 (LFA-1) or CD58 containing planar bilayers (Fig.
5). LFA-1 binds to CD54 (ICAM-1) on the Jurkat cells and
was not expected to disturb the distribution of CD2. After 1 h,
the cells were fixed, and the distribution of CD2 was examined by
confocal microscopy. CD2 displayed striking redistribution to the
contact area with CD58 bilayers, but not LFA-1 bilayers. Many
FITC-CD2.1-labeled Jurkat cells adhering to CD58 still showed specific
fluorescence in noncontact membranes. This suggests that free CD2 was
distributed over the cell surface.
FPR experiments were performed to determine whether CD2-CD58 bond
dynamics could be visualized through fluorescence recovery after
bleaching of FITC on CD58 molecules in the contact area. It was assumed
that photobleaching does not alter the function of the CD58 molecules.
This was strongly supported by the observation that the contact area,
as measured by the independent IRM method, does not change over the
course of the FPR experiment (not shown). If the CD58 were functionally
altered, then some acute change in the contact area would be expected.
Fig. 6a shows image data from a bleaching
experiment at a free CD58 density of 750 molecules/µm2.
Under these conditions, recovery was nearly complete in 2-3 min. Since
all the fluorescence accumulated in the contact area was eliminated as
shown by the first postbleach image, it was not possible to account for
this recovery in terms of diffusion of the CD2-CD58 bonds within the
contact area as was possible in our prior study with conventional FPR
(6). Therefore, the fast off-rate of the CD2-CD58 interaction was
manifested as rapid partner exchange between CD2 and CD58 in the
contact area.
The time required to restore the accumulated fluorescent CD58 to the
prebleach level may be dependent on the free CD58 density. This effect
would be expected based on the dependence of the flux rate for
FITC-CD58 molecules through the contact area on the free FITC-CD58
density. However, an additional component would be the density of free
CD58 molecules that could act as competitors to block rebinding of
recently dissociated CD2-CD58 bonds. FPR data from experiments with
different free FITC-CD58 densities are expressed in terms of free CD58
equivalents. Thus for 750 or 35 molecules/µm2, 1 Ib unit = 750 or 35 molecules/µm2, respectively. When the free CD58 density
was 750 or 35 molecules/µm2, the time required to restore
the accumulated fluorescence to one free CD58 equivalent over the
bilayer level was nearly identical (Fig. 6, b-d). The time
required for recovery of one Ib unit was nearly
always 30-40 s in the early stages of recovery, very similar to the
half time for recovery of the NBD-PE (Fig. 2). Nearly identical results
were obtained when cells were treated with phorbol myristate-acetate to
increase CD2 avidity (18) or with the CD2.1 IgG that can activate cells
in concert with CD58 (17). Part of this early recovery is due to the
restoration of the bleached free-CD58 in and around the contact area,
but it was clear that the accumulated CD58 in the contact area showed
recovery even before the surrounding free CD58 had fully recovered
(Fig. 5a, 13-s image). Thus, it appears that the turnover of
the CD2-CD58 bonds in the contact area was nearly diffusion-limited.
This was consistent with the fast off-rate of the CD2-CD58 bonds in
solution of >5 s1 (19). Therefore, we conclude that the
dissociation of CD2 from CD58 was followed by diffusion of the free
CD58 away from the CD2 receptor so that it can bind another CD2
molecule or diffuse out of the contact area. This property was not
regulated by cell activation or parallel ligation of CD2 with the CD2.1
antibody.
To determine if this kind of dynamic behavior was associated with
another low affinity receptor system, we studied the low affinity Fc
receptor II expressed on K562 cells. Bilayers were prepared with 2%
DNP-PE in egg PC and were labeled with FITC-rabbit anti-DNP IgG. Fig.
7 shows that the recovery of accumulated fluorescence is
not as rapid with Fc receptor II-IgG as with CD2-CD58 but that recovery
was clear at later time points. The recovery rate of bound IgG was at
least 10-fold slower than bound CD58 based on the apparent half-time
extrapolated from the data acquired. Therefore, the recovery time
course for accumulated fluorescence in cell-bilayer contacts was
clearly dependent on the molecular details of the receptor-ligand
system.
The novel finding of this report is that the CD2-CD58 bond
displays striking dynamics in contact areas. This result resolves the
question of whether the fast off-rate of bonds in solution is observed
in cell-cell contact areas where rebinding of dissociated receptors and
ligands might suppress observable turnover of transiently bound ligand
molecules (7). A degree of this dynamic behavior was also displayed by
the low affinity Fc receptor of K562 cells interacting with IgG on a
planar substrate. The major off-rate for the interaction of low
affinity Fc receptors with IgG in solution is 1.4 s1,
significantly slower than the off-rate for interaction of CD2 and CD58,
>5 s
1 (20). The exchange of CD58 molecules appeared to
be nearly diffusion-limited, but the exchange of IgG molecules engaged
by low affinity Fc receptors was much slower. It was critical to use
imaging-based FPR in these experiments in contrast to spot or line-scan
FPR experiments. By this method, we could establish that all the
accumulated CD58 or Fc receptor molecules had their FITC bleached and
that recovery occurred by an exchange reaction with unbleached
FITC-labeled molecules from outside the contact area.
Transient binding is likely to be prevalent for Ig family members and many other low affinity adhesion molecules. However, transient binding may not be universal for all adhesion systems. For example, integrins display regulated high affinity binding that may lead to formation of a smaller number of stable bonds in contact areas (21). Integrins in focal contacts do not show recovery after photobleaching (22). Some integrins may achieve irreversible binding that is only overcome by cellular force rather than spontaneous dissociation (23). It will be of interest to establish a parallel system for integrins where the diffusion of a laterally mobile ligand can be used to monitor binding and turnover. Such a system would allow a more systematic test of the relative importance of low and high affinity conformations of integrins than has been possible to this point.
In addition to adhesion molecules, other types of cellular processes may operate by transient binding in nearly two-dimensional systems. While the bonds formed are transient, the low affinity of the CD2-CD58 interaction has a remarkable ability to redistribute both CD58 (Fig. 4) and CD2 (Fig. 5) in the membrane; i.e. the low affinity interaction of CD2 and CD58 has a high biological affinity when used in its normal physiological context. We show here that this enhanced binding is not due to suppression of the off-rate in the contact area. The single most important factor in determining how efficiently a low affinity interaction can operate in a contact area is the confinement of the interaction to a very small space between the aligned cell membranes (see below). Similarly low affinities may be used to cluster receptors in clathrin-coated pits. Along these lines, it has been shown by FPR that the accumulation of receptors in coated pits can occur through transient interactions (24, 25). Furthermore, the distance from the membrane of binding motifs in the cytoplasmic tails of some endocytosed receptors is critical for their effectiveness (26), a property expected of interactions that operate by minimization of a confinement region. Recently, it has been shown that glycosyltransferases in the Golgi apparatus are also retained by transient interactions (27). Therefore, many biological systems have probably evolved to take advantage of the two-dimensional organization of fluid phospholipid bilayers to generate transient, but effectively high affinity, binding systems.
CD2 ligation can activate T cells (28). Pairs of anti-CD2 mAb are a
very potent stimulus for T cell activation (28, 29). This signaling
requires expression of the TCR complex (30). CD2 engagement by CD58
alone does not activate T cells (17, 31). The kinetic proofreading
model for T cell activation addresses the issue of bond duration and
signaling in theoretical terms (32). A correlation exists between
activation of T cells and TCR off-rates of less than 1 s1
(4). Therefore, CD2-CD58 interaction, with an off-rate of >5
s
1, is not likely to be able to assemble a signaling
complex (3). Since there are no known higher affinity ligands for CD2,
it is unlikely that CD2 ligation alone signals T cells. It is more
likely that CD2 plays a critical supporting role in T cell activation. This supportive role may be 3-fold. First, since CD2-CD58 complexes span the same gap between membranes as the TCR-peptide-MHC complex (33), the interaction of many CD2-CD58 molecules would position the T
cell and antigen-presenting cell at an optimal distance for the T cell
receptor-ligand interaction. Second, the large number of CD2-CD58
interactions can be used to overcome charge repulsion between cells,
thus eliminating bond strain on the TCR-ligand interaction. Finally,
the large, highly conserved cytoplasmic domain of CD2 may recruit
cytoskeletal and signaling molecules into the contact cap (34-36).
The dynamic nature of the CD2-CD58 interaction supports the application of equilibrium binding analysis to this problem. Recently, my collaborators and I (6) used the method of Scatchard to analyze equilibrium binding data for interaction of CD2 and CD58 in the T cell to planar bilayer system. The Scatchard analysis assumes that all CD2 is in the contact area and that the free CD2 is also concentrated in this region (37). Our results show that much of the CD2 does become concentrated in the cell-planar bilayer contact area. However, there is also CD2 evident in noncontact regions (Fig. 5b). If the CD2 outside the contact area represents laterally mobile free CD2, then the Scatchard method would overestimate the concentration of free CD2 and also would then overestimate the two-dimensional Kd by up to 10-fold (given that the contact area is approximately one-tenth the total cellular area). Therefore, the actual two-dimensional Kd for the CD2-CD58 interaction may be closer to 2 molecules/µm2 than the 21 molecules/µm2 that was determined with the Scatchard analysis. The confinement region may be on the order of 0.5 nm, rather than 5 nm as calculated previously (6). This suggests a remarkably precise alignment of the interacting membranes.
Recent experiments examining the hydrodynamic flow-driven rolling of
small beads coated with rat CD48 on substrates of rat CD2 show that
these bonds are transient (38). In some respects, these experiments
have better time resolution than earlier plasmon resonance experiments.
The off-rate under 10-piconewton force is 7.8 s1. Our
experiments do not allow direct measurement of an off-rate, since the
recovery rate in our experiment is a complex product of diffusion,
on-rate, and off-rate and may also be influenced by active cellular
processes. However, the experiments reported here demonstrate bond
dynamics under physiological conditions relevant to cell adhesion and
signaling and therefore make a unique contribution to our understanding
of this system and adhesion mediated by low affinity interactions in
general.
These experiments also have the advantage of having CD2 supplied on a cell. This allowed the study of the effects of cell activation on the CD2-CD58 interaction. Phorbol ester treatment increases the strength of CD2-mediated adhesion (18). However, we observed no change in the half-time for CD58 turnover in the contact area when Jurkat cells were treated with phorbol esters or an activating anti-CD2 monoclonal antibody. Thus, it is unlikely that these agents act by altering the affinity or off-rate of the CD2-CD58 interaction.
These results have implications for the mechanism of cell detachment. The bonds formed by the CD2-CD58 mechanism are individually transient, but when ~50,000 bonds form it is not likely that cells spontaneously detach given any amount of time (6). It is also unlikely that the cell can simply overpower these bonds (39, 40). Therefore, some mechanism is needed to get the CD2 or CD58 molecules out of the contact area. The CD2-CD58 mechanism is profoundly regulated by the surface charge (41). Interestingly, some of the major charge-carrying molecules of the cell, such as CD43, are rearranged by the cell cytoskeleton (42). Large sialoglycoproteins such as CD43 could be used as molecular "crowbars" to work the CD2-CD58 interactions apart. This is similar to the role that has been proposed for the form of neural cell adhesion molecule with polysialic acid (43). The insertion of large charged molecules into the contact area would increase the confinement region, increase the two-dimensional Kd, and decrease equilibrium bond formation. It will be important to study the effects of the glycocalyx on the two-dimensional Kd of the CD2-CD58 interaction and the observed bond dynamics.
I thank E. R. Unanue and S. Teitelbaum for generous support of my nascent effort. I owe special thanks to E. Elson for the extended loan of the argon laser and for valuable discussions. I thank D. E. Golan, T. A. Springer, D. M. Zhu, and Y. Nissim for valuable discussions and encouragement. I am grateful to J. M. Miller and R. Houdei for excellent technical assistance.