HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 52, 50255-50262, December 27, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/52/50255    most recent M209822200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Salas, A. Articles by Springer, T. Search for Related Content PubMed PubMed Citation Articles by Salas, A. Articles by Springer, T. Social Bookmarking                      What's this?

Transition From Rolling to Firm Adhesion Is Regulated by the Conformation of the I Domain of the Integrin Lymphocyte Function-associated Antigen-1^*

Azucena Salas, Motomu Shimaoka, Shuqi Chen, Christopher V. Carman, and Timothy Springer

From the Center For Blood Research and Departments of Pathology and Anesthesia, Harvard Medical School, Boston, Massachusetts 02115

Received for publication, September 24, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The integrin lymphocyte function-associated antigen-1 (_L2), which is known for its ability to mediate firm adhesion and migration, can also contribute to tethering and rolling in shear flow. The _L I domain can be mutationally locked with disulfide bonds into two distinct conformations, open and closed, which have high and low affinity for the ligand intercellular adhesion molecule 1 (ICAM-1), respectively. The wild type I domain exists primarily in the lower energy closed conformation. We have measured for the first time the effect of conformational change on adhesive behavior in shear flow. We show that wild type and locked open I domains, expressed in _L2 heterodimers or as isolated domains on the cell surface, mediate rolling adhesion and firm adhesion, respectively. _L2 is thus poised for the conversion of rolling to firm adhesion upon integrin activation in vivo. Isolated I domains are surprisingly more effective than _L2 in interactions in shear flow, which may in part be a consequence of the presence of _L2 in a bent conformation. Furthermore, the force exerted on the C-terminal -helix appears to stabilize the open conformation of the wild type isolated I domain and contribute to its robustness in supporting rolling. An allosteric small molecule antagonist of _L2 inhibits both rolling adhesion and firm adhesion, which has important implications for its mode of action in vivo.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Two distinct adhesive modalities are required for leukocyte accumulation at inflammatory sites and lymphocyte homing. Rolling adhesion greatly increases the time a cell spends in a post-capillary venule and enables surveillance of endothelium for activating signals such as chemoattractants. Firm adhesion results in the arrest of the leukocyte in the postcapillary venule, and sets the stage for diapedesis.

The integrin _L2 (lymphocyte function-associated antigen-1 (LFA-1)¹) mediates adhesion and migration of leukocytes in immune and inflammatory processes by binding to intercellular adhesion molecules (ICAMs), which are members of the Ig superfamily (1). Dynamic regulation of ligand-binding activity by _L2 and other integrins in response to signals transmitted from inside the cell (inside-out signaling) activates ₂ integrin adhesiveness in response to engagement of the antigen receptor on T lymphocytes in immune responses, and in response to chemoattractant binding to G-protein-coupled receptors in leukocyte adhesion to endothelium (2-4). ₂ integrin-mediated arrest of rolling leukocytes within the vasculature occurs on a second time scale, enabling arrest to occur in the same postcapillary bed where chemoattractant is encountered.

The ₂ integrins are far more facile in mediating firm adhesion than rolling adhesion. In many in vitro and in vivo systems in which ₂ integrins mediate firm adhesion and selectins mediate rolling adhesion, ₂ integrins are not seen to mediate tethering in shear flow or rolling (2, 3). The ₄ integrins have been known for some time to mediate rolling as well as firm adhesion (5, 6), although they do not support rolling as efficiently as selectins (7).

Recently, ₂ integrins have also been found to be capable of contributing to rolling in vivo and in special cases to support on their own rolling in vitro. When multiple adhesion pathways are blocked in vivo, i.e. selectins together with ₂ integrins or ICAM-1, or LFA-1 together with ₄ integrins, ₂ integrins and in particular _L2 can be seen to contribute to accumulation of rolling cells, the stability of rolling, and the velocity of rolling cells (8-10). _L2 can mediate tethering in flow of leukocytes to ICAM-2 on platelets in the absence of selectin-mediated interactions (11). An important ligand-binding domain of _L2, the inserted (I) domain of the _L subunit, has been expressed on the cell surface in isolation from other integrin domains and found to support rolling on immobilized ICAM-1 under shear flow (12). It was suggested that the I domain represents a transient ligand-binding domain and that cooperation with other ligand-binding domains was required for firm adhesion. However, recent studies have shown that the isolated I domain, when stabilized in a conformation that has high affinity for ligand, is sufficient for the same amount of adhesiveness in static binding assays as maximally activated _L2 (13), and binds with the same kinetics and affinity as activated _L2 in real-time soluble ligand-binding assays (14). _L2 expressed in K562 cells, which shows little basal activity in static adhesion assays (13, 15), has recently been shown to support rolling on ICAM-1 in shear flow, whereas _L2 expressed in Jurkat cells, which shows basal adhesion to ICAM-1 in static adhesion assays that is further inducible with activation, supported weak adhesion without rolling (16).

In integrins that contain I domains, conformational changes within the I domain regulate ligand binding (17). A downward movement of the C-terminal -helix of the I domain is linked to structural rearrangements in the metal ion adhesion site and surrounding loops that constitute the ligand-binding site of the I domain, as shown by crystallographic studies of _M and ₂ (18, 19). Movements in the same regions of the _L I domain occur in the presence of an ICAM-1 fragment as shown by NMR chemical shift experiments (20). Two conformations of the I domain termed open and closed have been shown to have high and low affinity for ligand, respectively. An engineered disulfide bond in the _L I domain that locks the loop between the C-terminal -strand and -helix into the open conformation has been shown to activate cell adhesiveness in static binding assays and to increase the affinity of the locked open I domain in soluble ligand-binding assays 9,000-fold relative to the wild type I domain (13-15). The increase in affinity was the result of a 50-fold increase in on-rate to 140,000 M¹s¹ and a 200-fold decrease in off-rate from 5 s¹ to 0.025 s¹. A mutant I domain that was analogously locked in the closed conformation was similar in affinity to the wild type I domain, and all evidence suggests that the closed conformation is lower energy than the open conformation and is the predominant conformation assumed by isolated I domains.

Small molecule antagonists directed to the I domain of _L2 have been developed. Crystal and NMR structures show that they bind to the closed conformation of the I domain, between the C-terminal -helix and the body of the domain, distal from the ligand-binding site on the "top" of the I domain (21, 22). These antagonists are allosteric modulators that stabilize the closed conformation of the I domain, as confirmed by failure to antagonize I domains locked in the open conformation with disulfide bonds (15).

The effect of conformational alterations in integrins on rolling interactions has not been studied, although it is known in general that activation can convert rolling to firm adhesion for both ₄ and ₂ integrins. It has long been hypothesized that rapid kinetics for bond association and dissociation are important in rolling (23), and that slower bond dissociation would favor firm adhesion. Although there exists now an extensive body of literature on the kinetics of bond dissociation for receptors that mediate rolling (24), there is no data on how conformational change, with accompanying alterations in bond association and dissociation kinetics, would affect adhesive behavior in shear flow. Recent crystal structure, NMR, and electron microscopic studies have revealed that integrins, including _L2, assume a highly bent conformation in the resting state, and that activation results in a dramatic switchblade-like opening (25-27). In the bent conformation the headpiece is close to the membrane, whereas in the active, extended conformation it moves ~15 nm upwards and into an orientation much more favorable for ligand binding. Furthermore, these conformations are in rapid equilibrium, and activation should be viewed as a shift in the equilibrium rather than fixing a particular conformation. The I domain, although not directly visualized in these studies, is connected to the headpiece and would become dramatically more accessible to ligand in the extended conformation. To examine the effect of only conformational change within the I domain, which is likely to be linked to global conformational change in native integrin heterodimers, we study I domains expressed in isolation from other integrin domains on the cell surface. We demonstrate that _L integrin I domain conformation, as influenced by disulfide bonds that lock in specific conformations, or binding of a small molecule antagonist that stabilizes the closed conformation, regulates rolling interactions in shear flow. Furthermore, we compare for the first time the efficacy of cell surface _L2 heterodimers and isolated _L I domains in rolling assays, and describe some surprising differences.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Cell Lines and Antibodies-- K562 cells stably transfected with _L2 containing wild type, locked open (K287C/K294C), or locked closed (L289C/K294C) I domains, or these I domains expressed in the absence of other integrin domains using a platelet-derived growth factor receptor (PDGFR) transmembrane domain and the first five amino acid residues of the PDGFR cytoplasmic domain were described previously (13). All mutant constructs were verified here to be expressed at similar levels on the cell surface as previously shown (13, 15). _L2 and I domain transfected cells were maintained in RPMI 1640 medium containing 10% fetal bovine serum, penicillin/streptomycin, and 3 µg/ml puromycin or 100 µg/ml hygromycin, respectively. The mouse anti-human _L monoclonal antibodies TS2/6 (28) and MHM24 (DAKO, Carpinteria, CA) were used to block LFA-1-mediated interactions. A non-binding mouse IgG1 (X63) as control and two different anti-human _L I domain monoclonal antibodies, TS1/11 and TS1/12, were used to determine surface expression of the transfectants by immunofluorescence flow cytometry (29).

Cell Adhesion to Immobilized ICAM-1 under Static Conditions-- Adhesion of K562 cell transfectants to ICAM-1 purified from human tonsil (30) and coated at 6 µg/ml on 96-well plates was assayed as described (29). Cells were labeled with 2',7'-bis-(carboxyethyl)-5-(and -6)-carboxyfluorescein, acetoxymethyl ester, and resuspended in Hank's balanced salt solution, 10 mM HEPES, pH 7.4, 0.5% bovine serum albumin containing either 1 mM Ca²⁺ + 1 mM Mg²⁺, or 2 mM Mg²⁺ + 1 mM EGTA. Cells were added to the ICAM-1-coated wells and incubated for 30 min at 37 °C. After incubation wells were washed and the fluorescence read.

Cell Adhesion to Immobilized ICAM-1 or HUVECs Under Shear Flow-- Three different forms of ICAM-1 were immobilized on substrates. Human tonsil ICAM-1 was directly coated on polystyrene Petri dishes for 1 h at 37 °C in coating buffer (phosphate-buffered saline, 20 mM bicarbonate, pH 9.0). Substrates were washed and blocked with 2% human serum albumin in coating buffer for 1 h at 37 °C. Soluble IC1-5/IgA chimera containing the five Ig domains of human ICAM-1 fused to the Fc portion of IgA (ICAM-1-Fc) was described previously (31). ICAM-1-Fc (10 µg/ml in coating buffer) was spotted on a dish previously coated with 20 µg/ml goat anti-human IgA in coating buffer (Zymed Laboratories Inc., San Francisco, CA) and blocked with 2% human serum albumin in coating buffer. ICAM-1-Fc (10 µg/ml in coating buffer unless noted otherwise) was spotted on a dish previously coated with protein A (20 µg/ml) in coating buffer for 1 h at 37 °C and blocked with 2% human serum albumin in coating buffer (32). HUVECs (American Type Culture Collection) were maintained in medium 199 modified Earle's salt solution containing 20% fetal bovine serum, 100 mg/ml endothelial growth supplement (Sigma), 1% Nutridoma-NS (Roche), and 100 µg/ml heparin at 37 °C in humidified air containing 5% CO₂. Cells were grown on polystyrene cell culture dishes pre-coated with 10 µg/ml fibronectin and used for no more than five passages. For flow experiments, HUVECs were seeded onto fibronectin-coated wells of six-well cell culture dishes at 90% confluency and cultured for 3-4 days prior to use. Cells either received no treatment or were activated with TNF- (100 ng/ml) for 5 or 24 h prior to each experiment. ICAM-1 surface expression was determined by flow cytometry using IC1/12, a mouse anti-human ICAM-1 monoclonal antibody (33) directly conjugated with Alexa488, and CBRM1/23-Alexa488, an anti-human _M monoclonal antibody (34) as a negative control.

ICAM-1 substrates or HUVEC monolayers were assembled as the lower wall in a parallel wall flow chamber and mounted on an inverted phase-contrast microscope (2). Cells were washed twice with Ca²⁺ and Mg²⁺-free Hank's balanced salt solution, 10 mM Hepes, pH 7.4, 5 mM EDTA, 0.5% bovine serum albumin and resuspended at 5 × 10⁶/ml in Ca²⁺ and Mg²⁺-free Hank's balanced salt solution, 10 mM Hepes, 0.5% bovine serum albumin (buffer A) and kept at room temperature (22 °C) throughout the experiment. Cells were diluted to 5 × 10⁵/ml in buffer A containing 1 mM Ca²⁺ + 1 mM Mg²⁺ or 2 mM Mg²⁺ + 1 mM EGTA immediately before infusion in the flow chamber using an automated syringe pump. Images were captured using a CCD camera mounted on an inverted microscope with a 10× objective and recorded on Hi-8 videotape.

Accumulation in Shear Flow and Rolling Velocity-- Cells were allowed to accumulate at 0.3 dynes/cm² for 30 s. Shear stress was increased every 10 s up to 36 dynes/cm². Rolling velocity at each shear stress was calculated from the average distance traveled by rolling cells in 3 s. The number of cells interacting for more than 3 s with the coated surface was measured at each shear stress. To avoid confusing rolling with small amounts of movement due to tether stretching or measurement error, a velocity of 1.5 µm/s, which corresponds to a movement of 1/2 cell diameter during the 3 s measurement interval, was the minimum velocity required to define a cell as rolling instead of firmly adherent.

Shear Detachment Assays-- To evaluate the strength of the _L2 and I domain interactions with ICAM-1, cells were infused into the flow chamber and allowed to settle onto the substrate in stasis for 2 or 5 min as indicated. Flow was initiated at a shear stress of 0.3 dyn/cm² and increased every 10 s. The number of cells that remained attached at the end of each shear interval was counted.

The Effect of LFA703 on LFA-1-ICAM-1 Interactions-- The statin-like LFA-1 antagonist LFA703 (35) was kindly provided by Novartis (Basel, Switzerland). LFA703 (100 mM in Me₂SO) was diluted in assay buffer (Hank's balanced salt solution/Hepes/bovine serum albumin). Cells were preincubated with LFA703 (0-100 µM) at room temperature for 15-30 min prior to infusion in the flow chamber. The 0 µM LFA703 or Me₂SO control used the same concentration of Me₂SO (0.1%) as in the highest LFA703 concentration.

The Effect of Latrunculin A on LFA-1 and I Domain-ICAM-1 Interactions-- Wild type and high affinity _L2 and I domain expressing K562 cells were incubated with 1 µM latrunculin A (Calbiochem) or Me₂SO for 20 min at RT. (22 °C).

Cells were resuspended in Buffer A containing 2 mM Mg²⁺. Cells were infused at 10⁷/ml and accumulated as described above. Shear stress was increased every 10 s, and the number of cells adherent at the end of each 10 s interval was counted.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Comparison of Static and Shear Flow Adhesion Assays-- Binding of transfected cells expressing wild type _L2 heterodimer or the wild type or locked closed isolated _L I domain to substrates coated with native ICAM-1 purified from tonsils is barely detectable if at all in conventional static adhesion assays in Ca²⁺ and Mg²⁺ (Fig. 1A) (13). However, locking the I domain in the open conformation by mutational introduction of a disulfide bond, or activation of wild type _L2 heterodimers with Mg²⁺/EGTA results in marked binding to ICAM-1 (Fig. 1A) (13). Contrasting results in Ca²⁺ and Mg²⁺ were obtained when cells were allowed to settle in stasis for 5 min on native ICAM-1 substrates and then subjected to controlled shear flow forces. Laminar flow was initiated at a wall shear stress of 0.3 dyn/cm² and then incremented every 10 s. The number of cells that were initially in the field of view at stasis was determined, and then the percentage remaining at the end of each 10 s step was determined (Fig. 1B). In agreement with the data in the static adhesion assays in Ca²⁺ and Mg²⁺, K562 transfectants expressing locked open _L2 heterodimers and the locked open _L I domain were able to bind to ICAM-1. In both cases the cells were firmly adherent; the average velocity of all adherent cells was close to 0 µm/s (Fig. 1C), and detachment from the substrate was not preceded by rolling. In further agreement with the results of the static assays, transfectants with closed I domains or closed heterodimers were not adherent, and transfectants with wild type _L2 heterodimers were also not adherent. However, in marked contrast to the results of the static assays, transfectants expressing the wild type isolated I domain were adherent in the shear flow assay (Fig. 1B). Furthermore, this was accompanied by a contrasting adhesive behavior: the cells rolled (12) (Fig. 1C). Excluding the highest shear stress, at which few cells remained, the cells rolled at a velocity of 15 to 30 µm/s, corresponding to several cell diameters per second.

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Binding of K562 transfectants to immobilized ICAM-1. K562 transfectants expressed the indicated I domain mutants within intact L2 or as isolated L I domains linked to the transmembrane and first 5 amino acid residues of the cytoplasmic domain of the PDGF receptor. All assays in this figure are with tonsil (native) ICAM-1 adsorbed to substrates at 6 µg/ml. A, static binding assays. Cells were allowed to bind in the indicated cations for 30 min at 37 °C to ICAM-1 in 96-well plates, and wells were washed by aspiration. Results are mean ± S.D. of duplicate samples from three different experiments. B and C, cells were infused into the flow chamber in medium containing 1 mM Ca2+ + 1 mM Mg2+ and allowed to settle in stasis for 5 min onto a substrate coated with ICAM-1. Flow was then initiated at a wall shear stress of 0.3 dyn/cm2 and increased every 10 s to the indicated values. The number of cells that remained bound at the end of each interval (B) and the average rolling velocity of all rolling and firmly adherent cells (C) was measured.

Effect of Presentation of ICAM-1 on the Substrate on Adhesion in Shear Flow-- Although we found that native ICAM-1 adsorbed to a substrate did not support adhesion by wild type _L2 transfectants, these transfectants have been reported to support rolling adhesion on ICAM-1-Fc chimeras bound to protein A substrates. Therefore, we compared different types of ICAM-1 substrates for support of adhesion in shear flow in the presence of 1 mM Ca²⁺ + 1 mM Mg²⁺. The initial cell binding to the substrate occurred in shear flow at 0.3 dyn/cm² rather than in stasis. ICAM-1 directly adsorbed to substrates or ICAM-1 fused to the Fc portion of IgA (ICAM-1-Fc) and immobilized by binding to anti-IgA on a substrate supported adhesion in shear flow and rolling of cells expressing the _L I domain, but not adhesion in shear flow and rolling of cells expressing _L2 (data not shown). However, ICAM-1 fused to the Fc portion of IgG (ICAM-1-Fc) and immobilized by binding to protein A on a substrate supported rolling of both I domain and _L2 transfectants (16) (Fig. 2). Rolling through _L2 was faster in velocity and less shear-resistant than rolling through the _L I domain (Fig. 2). The ability of unactivated _L2 to support rolling was not nearly as robust as the ability of activated _L2 to support firm adhesion but was consistent with its ability to contribute, in combination with other adhesion pathways, to tethering and rolling in vivo and on cellular substrates in vitro (see Introduction). The great sensitivity of native wild type _L2 but not the isolated _L I domain to the mode of ICAM-1 presentation on the substrate is consistent with adoption of the bent conformation by resting _L2, in which the I domain would have limited accessibility to ligand (26).

View larger version (20K): [in this window] [in a new window]   Fig. 2.   Rolling of K562 transfectants expressing wild type isolated L I domain or L2 on ICAM-1. Substrates consisted of ICAM-1-Fc/protein A coated on a plastic surface as described in "Experimental Procedures" or a monolayer of HUVECs stimulated with 100 ng/ml TNF- for 5 h. Cells were infused into the flow chamber and allowed to accumulate for 30 s at 0.3 dyn/cm2. Further accumulation or detachment occurred as the wall shear stress was increased in steps every 10 s. A, total adherent cells (rollingly and firmly adherent). B, average velocity of the adherent cells. Values show mean ± S.D. for three independent experiments.

To compare interactions mediated by the isolated I domain and _L2 using a more physiologically relevant presentation and density of ICAM-1, we used a monolayer of HUVECs either in basal conditions or after stimulation with TNF- for 5 or 24 h. In basal conditions, HUVECs express low amounts of ICAM-1 on their surface (6.6 M.F.I. measured with IC1/12 monoclonal antibody directly conjugated with Alexa488) and support no interaction in shear flow with K562 cells expressing the isolated I domain or _L2 (not shown). After incubation with 100 ng/ml TNF-, ICAM-1 expression increased in a time-dependent manner (57.3 mean fluorescence intensity at 5 h and 238.9 M.F.I. at 24 h) as previously shown (36).

The isolated I domain expressed on K562 cells efficiently interacted with TNF--stimulated endothelial cells (Fig. 2). The number and velocity of cells rolling on HUVEC stimulated 5 h with TNF- was comparable with the number and velocity of rolling cells observed in 10 µg/ml ICAM-1-Fc/protein A-coated substrates (Fig. 2).

Compared with cells expressing the isolated I domain, K562 cells expressing _L2 bound less efficiently to stimulated HUVEC monolayers. The number of adherent _L2-K562 cells on HUVEC and ICAM-1-Fc/protein A-coated substrates was comparable (Fig. 2A). Most cells adhered to the substrate during flow at 0.3 dyn/cm² because there was little additional adhesion after the flow rate was increased (Fig. 2A). The K562 cells expressing _L2 rolled when flow was incremented to 0.4 dyn/cm² but then quickly became firmly adherent (Fig. 2B). This reflects a difference in behavior on HUVEC compared with purified substrates that may reflect the ability of TNF-stimulated HUVEC to activate adhesion of _L2 on K562 transfectants. However, the main point for this study is that adhesion and rolling of I domain transfectants is very similar on ICAM-1-Fc/protein A and TNF-stimulated HUVEC substrates.

Controls in all experiments showed that rolling and firm adhesion were dependent on the I domain, as demonstrated by complete abrogation by two different antibodies to the I domain, TS2/6 and MHM24 (data not shown). Furthermore, both rolling and firm adhesion were completely abolished by EDTA (data not shown).

Effect of I Domain Conformation on Adhesion in Shear Flow-- To examine in the context of both _L2 and the isolated _L I domain the effect of I domain conformation on adhesive behavior in shear flow, transfectants were allowed to accumulate on ICAM-1-Fc/protein A substrates for 30 s at 0.3 dyn/cm² and as the wall shear stress was increased every 10 s (Fig. 3). The overall behavior in shear flow was similar for the isolated I domain and _L2 transfectants, except the number of cells that tethered to the substrate in shear flow was greater for I domain than _L2 transfectants (note difference in scale between Fig. 3, A-C and D-F).

View larger version (23K): [in this window] [in a new window]   Fig. 3.   Effect of locking the I domain in a closed or an open conformation on interactions under shear stress. ICAM-1-Fc was coated at 100 µg/ml (A and D) or 10 µg/ml (B, C, E, and F). K562 transfectants expressing the indicated wild type or mutant isolated L I domains or L2 heterodimers were infused into the flow chamber in medium containing 1 mM Ca2+ + 1 mM Mg2+ and allowed to accumulate for 30 s at a wall shear stress of 0.3 dyn/cm2 on a substrate coated with ICAM-1-Fc/protein A. Thereafter, shear was increased every 10 s, and the number of rollingly adherent cells (white bars) or firmly adherent cells (gray bars) was enumerated at each shear interval. Only cells interacting for  3 s were counted.

Cells expressing the locked closed low affinity conformation of the I domain showed very little tethering to the ICAM-1 substrate. At the standard coating concentration of 10 µg/ml ICAM-1-Fc used elsewhere in this manuscript and in Fig. 3, B-C and E-F, no tethering of the closed I domain was seen. 100 µg/ml was required to detect tethering of the locked closed I domain (Fig. 3A), and even this concentration did not support tethering of locked closed _L2 (Fig. 3D). The few cells that tethered showed rolling interactions, and the rolling cells were detached at higher shears (Fig. 3A).

Transfectants expressing wild type _L2 or the _L I domain tethered in shear flow and the majority of adherent cells rolled (Fig. 3, B and E). The percentage of rolling cells increased with shear, so that the vast majority of cells were rolling at  0.8 dyn/cm², both for I domain and _L2 transfectants.

Cells expressing the high affinity, open mutation of the I domain tethered ~2-fold more efficiently compared with wild type (Fig. 3, C and F). However, in marked contrast to wild type, > 95% of the tethered cells were firmly adherent even at the highest shears tested. Furthermore, the open mutant transfectants were more resistant to detachment at higher wall shear stresses. The behavior of cells bearing the open mutation in the isolated I domain or in _L2 was qualitatively similar.

We examined the requirement of the disulfide bond in the open mutant I domain for firm adhesion in shear flow. Previous studies showed that treatment of the open mutant I domain with a reducing agent abolished its increased affinity for ICAM-1 (14) and also abolished adhesion of isolated I domain transfectants to ICAM-1-coated plates under static conditions (13). Thus the disulfide bridge is required to lock the mutant I domain in the high affinity, open conformation. Very few transfectants expressing the open mutant I domain rolled on ICAM-1 over a range of shear stresses (Fig. 4), confirming the results in Fig. 3C. However, after treatment with the reducing agent DTT for 20 min at 37 °C, K562 cells expressing the open mutant I domain rolled as efficiently as cells expressing the wild type I domain (Fig. 4). Furthermore, DTT treatment of K562 cells expressing the wild type I domain had no effect on rolling. Thus, the effect of locking the I domain in the open conformation with a disulfide bond is reversible with DTT treatment, and the effect of DTT treatment on adhesive behavior in shear flow, i.e. conversion of firm adhesion to rolling, mirrors its effect on I domain affinity for ICAM-1.

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Effect of disulfide bond reduction on adhesive behavior in shear flow mediated by the open mutant isolated I domain. K562 transfectants expressing wild type or open mutant I domains (K287C/K294C) were incubated with or without 10 mM DTT for 20 min at 37 °C. Cells were infused into the flow chamber at a wall shear stress of 0.3 dyn/cm2, and the shear stress was incremented every 10 s. The percentage of the total adherent cells that were rollingly adherent was determined.

Differential Effect of Divalent Cations on _L2 and _L I Domain Interactions in Shear Flow with ICAM-1-- The presence of Mg²⁺ and absence of Ca²⁺, i.e. Mg²⁺/EGTA, activates adhesion through wild type _L2. Interactions mediated by _L2 with a locked open I domain were efficient in Ca²⁺/Mg²⁺ as shown above, and were not affected by removal of calcium (data not shown). We compared the effect of removal of Ca²⁺ on adhesive behavior in shear flow of wild type _L2 and wild type _L I domain transfectants (Fig. 5). In the presence of Ca²⁺/Mg²⁺, interactions through wild type _L2 were low in number (Fig. 5B); chelation of Ca²⁺ by EGTA greatly increased the number of cells that tethered and accumulated in shear flow (Fig. 5D). Furthermore, removal of Ca²⁺ changed the character of the interactions, because all of the increase was accounted for by cells that were firmly adherent. By contrast, for cells that expressed isolated I domains, removal of Ca²⁺ did not affect the percentage of rolling versus firmly adherent cells (Fig. 5, A and C). Nonetheless, the number of isolated I domain-expressing cells that accumulated and remained interacting was higher in Mg²⁺/EGTA than in Ca²⁺/Mg²⁺ (Fig. 5, A and C). These results suggest that activation of firm adhesion through _L2 by removal of Ca²⁺ requires a domain of the integrin other than the I domain, such as the ₂ I-like domain (15).

View larger version (23K): [in this window] [in a new window]   Fig. 5.   Effect of divalent cations on adhesive behavior in shear flow mediated by wild type isolated I domains or L2 heterodimers. K562 transfectants expressing the wild type isolated L I domain or L2 heterodimer in medium containing either 1 mM Ca2+ + 1 mM Mg2+ (A and B) or 2 mM Mg2+ + 1 mM EGTA (C and D) were infused for 30 s at 0.3 dyn/cm2 on a substrate coated with ICAM-1-Fc/protein A. Thereafter, shear was increased every 10 s and the numbers of rollingly adherent cells (white bars) and firmly adherent cells (gray bars) was determined.

Role of the Actin Cytoskeleton in the Interactions Mediated by _L2 and Isolated I Domain Transfectants-- Could the actin cytoskeleton be involved in regulating the transition between rolling and firm adhesion by wild type _L2? The isolated I domains are expressed using a PDGFR transmembrane and a truncated PDGFR cytoplasmic domain, which are not expected to interact with the actin cytoskeleton. _L2 and I domain transfectants were treated with latrunculin A, which associates specifically with actin monomers, preventing them from polymerizing into filaments (37). Disruption of actin filaments did not significantly affect the percentage of cells that mediated rolling versus firmly adherent interactions for wild type or open _L2 or wild type or open _L I domain transfectants (Fig. 6). However, treatment with latrunculin A did significantly increase the total number of _L2 transfectants that interacted with the ICAM-1 substrates, both for the wild type and the open mutant. This is consistent with the generally observed enhancing effect of actin disruption on adhesion through _L2.

View larger version (39K): [in this window] [in a new window]   Fig. 6.   Role of the actin cytoskeleton in rolling interactions. Cells expressing the wild type or open L2 or isolated I domain were treated with 1 µM latrunculin A or an equivalent volume of Me2SO for 20 min at RT. Cells were resuspended in buffer A containing 2 mM Mg2+, infused into the flow chamber, and allowed to accumulate for 30 s at 0.3 dyn/cm2 over ICAM-1-Fc/protein A. Shear stress was increased every 10 s, and the number of rollingly adherent cells (white bars) and firmly adherent cells (gray bars) was determined. Bars represent the average ±standard deviation. *, p < 0.05 versus Me2SO treatment.

A Small Molecule Allosteric Antagonist of _L2 Inhibits Interactions in Shear Flow-- As an independent method of examining the effect of I domain conformation on adhesive interactions in shear flow, we took advantage of a small molecule antagonist of _L2. LFA703 is a statin-like analogue that is 10-36 times more potent than lovastatin, a previously described _L2 inhibitor (21, 35). Lovastatin was previously shown to inhibit adhesion under static conditions to ICAM-1 through _L2 stimulated by Mn²⁺ or CBR LFA-1/2 (an activating antibody to the ₂ subunit). However, lovastatin did not inhibit adhesion through locked open _L2, confirming that the mode of action of lovastatin is to stabilize the I domain in the closed conformation (13). We first examined cells that were allowed to adhere to ICAM-1 in stasis and then subjected to increasing wall shear stress (Fig. 7). Adhesion through wild type _L2 was activated with either Mg²⁺/EGTA or Mn²⁺. Under these conditions essentially all of the cells were firmly adherent, i.e. there was no rolling. Resistance to detachment in shear was marked in the presence of Mg²⁺/EGTA and even greater in Mn²⁺ (Fig. 7A). The shear resistance of Mn²⁺-activated wild type _L2 was dramatically decreased by incubation with 10 µM LFA703, and no adhesion whatsoever was demonstrable for Mg²⁺/EGTA-activated _L2 in the presence of LFA703 (Fig. 7A). Under the same divalent cation conditions, adhesion through locked open _L2 was markedly resistant to shear, with greater shear resistance in Mn²⁺ than in Mg²⁺/EGTA (Fig. 6B). However, in contrast to firm adhesion through wild type _L2, firm adhesion mediated by open _L2 was not susceptible to inhibition by LFA703 as shown by the lack of effect on resistance to detachment (Fig. 7B, closed squares and circles).

View larger version (16K): [in this window] [in a new window]   Fig. 7.   Effect of the small molecule antagonist LFA703 on interactions of wild type or locked open L2 with ICAM-1. Cells were incubated at room temperature for 15 min with 10 µM LFA703 or an equivalent amount of Me2SO, infused in the flow chamber in medium containing the indicated divalent cations, and allowed to settle on the ICAM-1-Fc/protein A substrate for 2 min. Shear flow was then initiated, and the wall shear stress was increased every 10 s. At the end of each shear stress interval the number of cells that remained bound to the substrate was counted and expressed as a percentage of the cells present during the 2 min incubation at stasis.

We next examined the effect of LFA703 in shear flow under conditions where cells roll on ICAM-1, i.e. with cells expressing wild type _L2 in Ca²⁺ + Mg²⁺ or with cells expressing the wild type I domain (Fig. 8). Use of a range of concentrations of LFA703 with wild type _L2 transfectants demonstrated a dose-dependent decrease in the number of rolling cells at each shear stress (Fig. 8A). The IC₅₀ was shear-dependent, with an IC₅₀ of about 3 µM at 0.8 dyn/cm², about 1 µM at 1.6 and 3.2 dyn/cm², and about 0.5 µM at 6 dyn/cm² (Fig. 8A). LFA703 also inhibited rolling mediated by the wild type isolated _L I domain (Fig. 8B). The IC₅₀ was consistently higher for the isolated _L I domain than _L2 and again was shear-dependent. The IC₅₀ was about 200 µM at 0.8 dyn/cm², about 75 µM at 1.6 dyn/cm², about 30 µM at 3.2 dyn/cm², and about 20 µM at 6 dyn/cm² (Fig. 8B). The 50-100-fold lower IC₅₀ for _L2 than the _L I domain is likely to reflect the finding that the C-terminal -helix under which LFA703 binds has marked segmental mobility in isolated I domains (38), whereas when this helix is connected to the -propeller domain in intact _L2, it is likely to be much more ordered and provide a higher affinity binding pocket. The more intimate association of the C-terminal -helix with the side of the I domain in _L2 is corroborated by the activating effect of mutations in this helix in _L2 but not isolated _L I domains (17).

View larger version (32K): [in this window] [in a new window]   Fig. 8.   Effect of the small molecule antagonist LFA703 on rolling on ICAM-1 mediated by wild type L2 or isolated I domain. Cells expressing the wild type L2 or isolated L I domain were incubated at room temperature with the indicated concentrations of LFA703 or Me2SO for 15 min. Cells were infused into the flow chamber in medium containing 1 mM Ca2+ + 1 mM Mg2+ and allowed to accumulate for 30 s at 0.3 dyn/cm2. Further accumulation or detachment occurred as wall shear stress was increased every 10 s.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Recent studies (8-10) in vivo and in vitro have shown that _L2 can contribute to tethering and rolling interactions in shear flow, although less robustly than in supporting firm adhesion and cell migration. A question of major biological interest is how the conformation of _L2, which is known to be regulated by signals within the cell in inside-out signaling, affects its adhesiveness in shear flow, and in particular, the critical transition from rolling adhesion to firm adhesion. Here we show for the first time that conformational change in an adhesion receptor can alter adhesive behavior in shear flow. Although the wild type I domain and resting wild type _L2 mediate rolling, the locked open I domain and locked open _L2 mediate firm adhesion. High affinity for ICAM-1 resulted in firm adhesion that was highly resistant to detachment by increasing shear. The conversion from rolling adhesion to firm adhesion effected by the change in conformation of the I domain was mirrored by activation with Mg²⁺/EGTA or Mn²⁺ of wild type _L2. Interestingly, reduction of the disulfide bond constraining the I domain in the open conformation fully restored the ability of the isolated I domain to roll and abolished its ability to mediate firm adhesion, in agreement with abolition of adhesion in static assays and abolition of high monomeric affinity (14). Therefore, conformational change with an accompanying increase in affinity is sufficient to convert _L2 from a receptor that mediates rolling adhesion to a receptor that mediates firm adhesion. Although "avidity regulation" has been suggested for _L2, results interpreted in support of avidity regulation could also be explained by an intermediate increase in the affinity of _L2 (14). Our results with latrunculin A-treated _L2 transfectants show that association with the actin cytoskeleton does not regulate the transition from rolling to firm adhesion. Actin cytoskeleton disruption did not change the ratio of rolling versus firmly adherent cells for wild type or open _L2 transfectants, or wild type or open I domain transfectants. Isolated I domains were expressed using heterologous transmembrane and cytoplasmic domains, and only five residues were present in the artificial cytoplasmic domain. This further supports the conclusion that neither inside-out signaling nor avidity changes via clustering are required for regulating the transition between the rolling and firm adhesion states.

Three properties of a receptor-ligand bond are important for its ability to mediate rolling: on-rate, off-rate, and the mechanical property (i.e. the susceptibility of off-rate to increase by force). So far, the mechanical property is not known for the ICAM-1_L I domain bond, although it has been measured for selectin and ₄ integrin bonds (7). It is intriguing that the conformation-induced change in the off-rate of the ICAM-1_L I domain bond appears sufficient to explain the change in state from rolling to firm adhesion. The off-rate of the wild type I domain of 5 s¹ is in the range of 0.5-10 s¹ measured for rolling interactions through selectins and the integrin ₄₇ in Ca²⁺ (reviewed in Refs. 7 and 24). This correlates with the ability of the wild type I domain to support rolling. By contrast, the off-rate of the locked open I domain of 0.025 s¹ (14) is well outside this range, correlating with its ability to support firm adhesion and not rolling adhesion. Longer bond lifetimes are theoretically more conducive to firm adhesion. Furthermore, the off-rate of the ₄₇MAdCAM bond in Mg²⁺ of 0.046 s¹ is in the same range as the locked open _L I domain and supports firm adhesion not rolling adhesion (7). Thus, the transition of the _L I domain to the open conformation triggers a change in off-rate that appears to be perfectly tailored biologically to trigger a transition from rolling adhesion to firm adhesion.

We believe that the surprising effectiveness in mediating rolling of the isolated I domain is a consequence of the downward force exerted on the C-terminal -helix by tethering in shear flow that stabilizes the open, high affinity conformation of the I domain. When the I domain on the cell surface binds to ICAM-1 on the substrate, the cell becomes tethered through the I domain. The hydrodynamic force exerted on the tethered cell is balanced by a force exerted on its connection to the substrate through ICAM-1 and the I domain. If we analyze how this force is transmitted through the I domain, it is clear that a force exerted on the I domain interface with ICAM-1 at the top face of the I domain bearing the metal ion adhesion site is balanced by an opposing force exerted on the C-terminal -helix of the I domain, which is connected through a linker to the cell membrane. The direction of this force is roughly parallel to the axis of the C-terminal -helix, such that it will exert a downward pull on it. Downward movement of this helix stabilizes the open, high affinity conformation of the _L I domain (17). Therefore, whereas the wild type isolated I domain exists predominantly in the closed conformation, after binding to ICAM-1 the tether force will shift the conformational equilibrium toward the open conformation and increase the effective affinity for ICAM-1. This explains a number of our observations. 1) The wild type _L I domain was markedly more active than the locked closed I domain in mediating cell accumulation on ICAM-1 substrates in shear flow, and after adhesion was initiated under static conditions, in mediating cell rolling and resistance to detachment with increasing shear. Rolling can be observed for the closed I domain but requires high ICAM-1 densities and low shear stresses. In the locked closed I domain the disulfide bridge prevents the force exerted on the ₇-helix from pulling it downwards and reshaping the critical ₆-₇ loop. 2) In static assays, adhesion is not detectable with the wild type I domains, whereas it is readily detectable and indeed robust in shear flow assays. 3) LFA703 inhibits rolling through the wild type I domain. Because this compound stabilizes the closed conformation of the I domain this provides direct evidence that the wild type I domain shifts to the open conformation during rolling and that this is crucial to stabilize rolling.

Rolling adhesion through the wild type isolated _L I domain is far more efficient that through wild type _L2. This may in part reflect the connection of the I domain through both its N and C termini in _L2 and through only its C terminus in the isolated I domain, which would alter the effect of applied force on the equilibrium between the open and closed conformations. However, another important difference is that _L2 appears to assume the same bent conformation as _V3 in the resting state (26, 27). This places the I domain close to the cell membrane, in a less favorable orientation for interaction with ICAM-1 than in the isolated I domain.

Recently developed small molecule antagonists of _L2, including the statin analogue tested here, LFA703, stabilize the _L I domain in the closed conformation (13, 21, 22, 35). This has been confirmed here by resistance of locked open _L2 to inhibition by LFA703 in shear flow assays. Although these allosteric antagonists have previously been shown to bind to I domains in crystal and NMR studies, they have not previously been assessed for effect on soluble ligand binding or adhesive activity by isolated I domains. We show here that rolling of cells expressing both the isolated wild type I domain and wild type _L2 was dramatically impaired in a dose-dependent manner by the allosteric inhibitor LFA703. Our findings show that allosteric inhibitors inhibit both the rolling and firm states of adhesion mediated by _L2. These findings have important implications for mode of action of this class of antagonists in vivo.

	FOOTNOTES

^* This work was supported by NIH Grant CA31798.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence should be addressed. Tel.: 617-278-3200; Fax: 617-278-3232; E-mail: springeroffice@cbr.med.harvard.edu.

Published, JBC Papers in Press, October 3, 2002, DOI 10.1074/jbc.M209822200

	ABBREVIATIONS

The abbreviations used are: LFA, lymphocyte function-associated antigen-1; ICAM, intercellular adhesion molecule; PDGFR, platelet-derived growth factor receptor; TNF-, tumor necrosis factor ; DTT, dithiothreitol; dyn, dynes; HUVEC, human umbilical vein endothelial cells; NMR, nuclear magnetic resonance.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES