Research Article

Regulation of outside-in signaling and affinity by the β2 I domain of integrin αLβ2

See allHide authors and affiliations

PNAS August 29, 2006 103 (35) 13062-13067; https://doi.org/10.1073/pnas.0605666103

  1. Contributed by Timothy A. Springer, July 6, 2006

  2. ↵*J.C. and W.Y. contributed equally to this work.

Abstract

The adhesiveness of integrin αLβ2 is modulated by divalent cations. We mutated three metal ion-binding sites in the β2 I domain. The metal ion-dependent adhesion site (MIDAS) and the ligand-induced metal-binding site are required for ligand binding and sufficient for synergism between Ca2+ and Mg2+. Adjacent to MIDAS (ADMIDAS) mutants are constitutively active but remain bent, with poor exposure of a β2 stalk region epitope. Fluorescence resonance energy transfer between fluorescent protein-fused αL and β2 cytoplasmic domains showed that ADMIDAS mutation abrogated ligand binding-induced spatial separation of cytoplasmic domains. Furthermore, ADMIDAS mutation abolished spreading on ligand-bearing substrates. Thus, β2 I domain metal ion-binding sites regulate αL I domain affinity, and the ADMIDAS is required for outside-in signaling.

Integrins are α/β heterodimeric metalloproteins, the ligand-binding function of which depends on the interplay of Ca2+, Mg2+, and Mn2+ ions. The liganded crystal structures of αIIbβ3 and αVβ3 reveal a striking linear cluster of three divalent cation-binding sites in the β I domain (1, 2). The metal ion-dependent adhesion site (MIDAS) is at the center of the cluster. The outer two sites are known as the ligand-induced metal-binding site (LIMBS) and the adjacent to MIDAS (ADMIDAS) site. Work with integrins that lack α I domains (α4β7 and α5β1) has shown that the LIMBS and ADMIDAS modulate binding of ligand to the MIDAS (3–5).

Integrins adopt a bent, compact conformation in their low-affinity state and an extended conformation in the high-affinity state (1, 2, 6–10). Furthermore, ligand binding induces the extension of the extracellular domain and a movement at the β I domain–hybrid domain interface (1, 9, 10). The β I domain is inserted in the hybrid domain (i.e., it is connected at both its N and C terminus to the hybrid domain). Crystal structures of integrin αIIbβ3 in the liganded, high-affinity conformation reveal that the C-terminal α7-helix of the β I domain moves axially toward the hybrid domain, causing the β hybrid domain to pivot about its other connection and to swing outward by 62°, away from the α-subunit (1). Movement of the α7-helix is linked to that of the adjacent α1-helix and to movements of the preceding β6-α7 and β1-α1 loops, which form important coordinations at the MIDAS and ADMIDAS. Thus, the α7-helix is a key allosteric link between the ligand-binding site in the β I domain and the hybrid domain. Accompanying integrin extension, the cytoplasmic and transmembrane domains separate, which enables bidirectional signal transmission across the plasma membrane (11).

Half of the vertebrate integrin α-subunits contain α I domains, which mediate ligand binding (6, 8, 12). Here, we examine in integrins with α I domains whether the β I domain LIMBS and ADMIDAS modulate ligand binding as they do in integrins where the β I domain directly participates in ligand binding. Furthermore, we find that the ADMIDAS is required to couple ligand binding with αL and β2 cytoplasmic domain separation and outside-in signaling.

Results

Design and Cell Surface Expression of Mutant αLβ2 Integrins.

To investigate the function of the β I domain LIMBS, MIDAS, and ADMIDAS (LMA) sites in regulating the ligand-binding affinity of αLβ2, representative primary coordinating residues of each metal ion-binding site were mutated to Ala (Table 1). The β3 and β2 I domains are highly conserved, with 50.4% amino acid sequence identity. All residues with side chains coordinating to LMA metal ions are identical (Fig. 1 A). Because of the high identity in sequence and absolute identity in coordinating side chains, we assume that the β3 I domain structure is an excellent model for understanding the effect of β2 LMA mutations. Immunofluorescence flow cytometry with TS2/4, a mAb that recognizes the αL-subunit β-propeller domain only when it is associated with β2-subunit I domain (13), demonstrated that the β2 AMIDAS and LIMBS mutants were expressed at WT levels in 293T transfectants and that the β2 MIDAS mutant was expressed at 60% of WT levels (Table 1).

View this table:
Table 1.

Expression of αLβ2 mutants

Fig. 1.
Fig. 1.

The β I domain metal-binding sites. (A) Sequence alignment of human integrin β3 and β2 I domains. Residues with metal-coordinating side-chain oxygen atoms are highlighted, and residues with metal-coordinating backbone carbonyl oxygen atoms are asterisked. Highlights and asterisks are color-coded as follows: orange, MIDAS; green, ADMIDAS; cyan, LIMBS. (B) The linear cluster of β I domain metal-binding sites. The sites are shown in the liganded, open αIIbβ3 structure (cyan) (1) and the unliganded, closed αVβ3 structure in Ca2+ (yellow) (41), superimposed by using the β I domain. Metal ions are shown as large green and orange spheres for open and closed structures, respectively. Metal-coordinating side chains and backbone carbonyl groups are shown, with N and O atoms in blue and red, respectively. Residues are labeled for β3 and are shown in parentheses for β2.

Regulation of αLβ2 Adhesion by Divalent Cations.

The binding in different divalent cations of soluble human intercellular adhesion molecule-1 (ICAM-1)–IgA Fc fusion protein (ICAM-1–Fcα) complexed with FITC-anti-IgA to αLβ2 293T transfectants was measured by immunofluorescence flow cytometry. There was little binding of ICAM-1–Fcα complexes to cells expressing WT αLβ2 in 1 mM Ca2+/1 mM Mg2+. ICAM-1–Fcα binding was greatly enhanced by replacement of 1 mM Ca2+/1 mM Mg2+ with 1 mM Mn2+ (Fig. 2 A). Similar results were obtained in parallel plate shear flow adhesion assays with human ICAM-1–IgG1 Fc fusion protein (ICAM-1–Fcγ) incorporated on the lower wall of the flow chamber (Fig. 2 B). Cells were allowed to accumulate on ICAM-1–Fcγ at a wall shear stress of 0.3 dyne·cm−2 (1 dyne = 10 μN) for 30 s. Then, the wall shear stress was incrementally increased every 10 s, and the velocity of the cells remaining bound at each increment was determined and used to enumerate rolling and firmly adherent cells. Compared with 1 mM Ca2+/1 mM Mg2+, 1 mM Mn2+ greatly increased the number of WT firmly adherent cells (Fig. 2 B).

Fig. 2.
Fig. 2.

Integrin affinity regulation by mutations of the β2 I domain metal ion-binding sites. (A) Binding of ICAM-1–Fcα to 293T cell transient transfectants. Binding was assayed in Hepes/NaCl/glucose/BSA (20 mM Hepes, pH 7.5/140 mM NaCl/2 mg/ml glucose/1% BSA) supplemented with 1 mM CaCl2/1 mM MgCl2, 1 mM MnCl2, or 1 mM MnCl2/1 μM XVA143 as indicated at room temperature. The specific mean fluorescence intensity was normalized by dividing by the ratio of mutant/WT TS2/4 mAb-specific mean fluorescence intensity. (B) The adhesive modality of αLβ2 and resistance to detachment in shear flow. Rolling and firmly adherent WT and mutant αLβ2 293T transient transfectants were enumerated in 1 mM CaCl2/1 mM MgCl2, 1 mM MnCl2, or 1 mM CaCl2/1 mM MgCl2 plus 1 μM XVA134 at incremental wall shear stresses. Results at 1, 2, 4, and 8 dyne·cm−2 were very similar and are averaged together. Error bars show the SD of three independent experiments.

Regulation of Adhesion by β2 I Domain LMA Sites.

The β2 MIDAS mutation, S114A, completely abolished αLβ2 ligand binding under all conditions (Fig. 2). Relative to WT, the two ADMIDAS mutations, D119A and D120A, strikingly enhanced multimeric ICAM-1–Fcα binding in the presence of 1 mM Ca2+/Mg2+ (Fig. 2 A). Similarly, the ADMIDAS mutations greatly enhanced accumulation in shear flow in the presence of 1 mM Ca2+/Mg2+ (Fig. 2 B). In contrast, the two LIMBS mutations, N207A and D209A, abolished adhesion in shear flow in 1 mM Ca2+/Mg2+, although they mediated very weak adhesion in the presence of 1 mM Mn2+ (Fig. 2 B). Adhesion to ICAM-1 in the above two assays is mediated by the αL I domain, as previously shown (14, 15). The above results demonstrate that αL I domain-mediated ligand binding to ICAM-1 is regulated by the β2 I domain LMA sites.

The Effect of αLβ2 Small Molecule Antagonist XVA143.

XVA143 is an αLβ2 allosteric antagonist that binds to the β2 I domain MIDAS and blocks transmission of activation signals to the αL I domain (14, 16, 17). In both Ca2+/Mg2+ and Mn2+, XVA143 (1 μM) completely inhibited multimeric ICAM-1–Fcα binding to WT αLβ2 and the ADMIDAS mutants, D119A and D120A (Fig. 2 A). XVA143 somewhat increased the number of WT cells that accumulated in shear flow and greatly increased the number of rolling cells and decreased the number of firmly adherent cells as described in ref. 14 (Fig. 2 B). The firm adhesion induced by the ADMIDAS mutants in Ca2+/Mg2+ was completely blocked by XVA143 (Fig. 2 B). These results demonstrate that, as expected, and as previously demonstrated for regulation of adhesiveness by Mn2+ (14, 18), signal transmission from the β2 I domain to the αL I domain is required for activation of adhesiveness by the ADMIDAS mutants.

Positive and Negative Regulation by Low and High Ca2+ Concentrations.

For most integrins, Ca2+ has both positive and negative regulatory effects. Low (≈0.1 mM) concentrations of Ca2+ augment adhesion at suboptimal (≈0.1 mM) Mg2+ concentrations, whereas higher (1–10 mM) concentrations of Ca2+ inhibit adhesion at optimal (≈1 mM) Mg2+ concentrations (3, 19–21). The ADMIDAS mutants allowed us to study these effects under conditions in which basal adhesiveness was augmented and the contribution of metal ion binding to the ADMIDAS could be ruled out. At a range of wall shear stresses, little accumulation on ICAM-1 substrates of ADMIDAS mutant D120A was seen in 50 μM Ca2+ or 50 μM Mg2+, whereas the combination of 50 μM Ca2+ and 50 μM Mg2+ synergistically promoted strong adhesiveness (Fig. 3 A). Therefore, the β2 I LIMBS and MIDAS are sufficient for synergism between Ca2+ and Mg2+.

Fig. 3.
Fig. 3.

Positive and negative regulation of αLβ2 by Ca2+. (A) Synergistic effect of low concentrations of Ca2+ and Mg2+ on the adhesion of αLβ2 D120A 293T transfectants on ICAM-1. The number of adherent cells was measured in the indicated divalent cations at different wall shear stress. (B and C) Effect of 10 mM Ca2+ on the adhesion of αLβ2 293T transfectants to ICAM-1. Adherent cells were enumerated in the indicated divalent cations at the wall shear stress of 2 dyne·cm−2. The data show mean ± SD from three independent experiments.

Accumulation in shear flow on ICAM-1 substrates of WT αLβ2 transfectants in 1 mM Mn2+ was greatly decreased by addition of 10 mM Ca2+ (Fig. 3 B). In contrast, 10 mM Ca2+ had no inhibitory effect on the accumulation of the αLβ2 ADMIDAS mutant. Similarly, high concentrations of Ca2+ inhibited the Mg2+-promoted adhesion of WT αLβ2 transfectants but not αLβ2 ADMIDAS transfectants (Fig. 3 C). Indeed, Ca2+ enhanced adhesiveness by the ADMIDAS mutant (Fig. 3 C), consistent with the synergism seen in Fig. 3 A. These results strongly suggest that the β2 I domain ADMIDAS is responsible for inhibition of αLβ2 adhesiveness by high Ca2+ concentrations.

Integrin Conformational Changes Induced by β I Domain LMA Mutations.

To study the effect on integrin conformation by the β2 I domain LMA mutations, the exposure of epitopes recognized by two anti-β2 activation-dependent mAbs was measured. The extended conformation of the β2-subunit is reported by the mAb KIM127, which maps to species-specific residues in the integrin EGF2 domain of integrin β2 subunit that are buried in the headpiece–tailpiece interface in the bent integrin conformation and exposed in the extended conformation (7, 22). The active conformation of the β2 I domain is detected by mAb m24, which recognizes the species-specific residues Arg-122 in the α1-helix and Glu-175 in the specificity-determining loop of the β2 I domain (23–25).

KIM127 bound poorly to WT αLβ2 transfectants in the presence of Ca2+/Mg2+ (Fig. 4 A). Activation by Mn2+ or Mn2+ plus a soluble, monomeric ectodomain fragment of ICAM-1 containing Ig superfamily domains 1–5 (sICAM-1) greatly enhanced the KIM127 binding to WT αLβ2 (Fig. 4 A). Binding of KIM127 was highest in XVA143. The results with WT αLβ2 and the m24 epitope were similar, except that binding in XVA143 was lower than in Mn2+ and Mn2+/sICAM-1 (Fig. 4 A). Neither antibody bound to the S114A MIDAS mutant. Binding of KIM127 and m24 antibodies to the D120A ADMIDAS mutant was abolished in Mn2+ and Mn2+/sICAM-1 and only slightly enhanced (if at all) by XVA143 (Fig. 4 A). In contrast, D119A transfectants showed low binding to KIM127 antibody, with little influence by divalent cations and sICAM-1, and WT level binding in XVA143 (Fig. 4 A). The D119A mutant showed Mn2+-augmented binding to m24 that was below WT levels and augmentation by XVA143 that was near WT level. The LIMBS mutants N207A and D209A showed greatly decreased KIM127 and m24 binding compared with WT in Mn2+ and Mn2+/sICAM-1, and the N207A mutant showed near WT levels of binding in XVA143.

Fig. 4.
Fig. 4.

Effect of β I domain metal ion-binding site mutations and XVA143 on integrin conformation. The epitope exposure of β2 activation-dependent mAbs KIM127 (A) and m24 (B) was measured. WT and mutant 293T transient transfectants in medium containing 1 mM CaCl2/1 mM MgCl2, 1 mM MnCl2, or 1 mM MnCl2/100 μg/ml sICAM-1 were stained with KIM127 or m24 mAbs and subjected to immunofluorescence flow cytometry. Expression of the activation-insensitive mAb TS2/4 was not affected by MnCl2 and sICAM-1. Data show specific mean fluorescence intensity as a percentage of TS2/4-specific mean fluorescence intensity, and error bars represent the SD of three independent experiments.

The above experiments demonstrate that the KIM127 epitope is largely masked in Ca2+/Mg2+, Mn2+, and Mn2+/sICAM-1 in the ADMIDAS and LIMBS mutants. Thus, despite activation of ligand binding by the ADMIDAS mutations, and in contrast to activated WT αLβ2, the ADMIDAS mutants remain in the bent conformation. The complete inhibition by XVA143 of firm adhesion of the ADMIDAS mutants in Fig. 2 B acts as a positive control for the effects of XVA143 on the KIM127 epitope in Fig. 4 A. Thus, we may conclude that ADMIDAS mutant D120A blocks KIM127 epitope exposure induced by Mn2+, Mn2+/sICAM-1, and XVA143 and that the D119A ADMIDAS mutant blocks exposure induced by Mn2+ and Mn2+/sICAM-1 but not exposure induced by XVA143. The results with m24 in the above experiments must be interpreted with caution, because m24 binds to species-specific residues that lie on either side of the LMA sites. Thus, the LMA sites may comprise part of the m24 epitope, and mutations in these sites, the presence of Mn2+, and binding of XVA143 may directly affect the m24 epitope.

The ADMIDAS Is Essential for Coupling Integrin Ligand Binding and Cytoplasmic Domain Separation.

The lack of basal KIM127 exposure in the ADMIDAS mutants suggested that the ADMIDAS may not only regulate ligand binding but also couple ligand binding to global integrin conformation and to signaling by integrin cytoplasmic domains. We investigated this possibility with the ADMIDAS mutant that shows the least basal KIM127 epitope exposure, D120A. Monomeric mutants of cyan fluorescent protein (mCFP) and yellow fluorescent protein (mYFP) were fused to the C termini of integrin αL- and β2- subunits, respectively, as previously described for WT αLβ2 (11). The addition of C-terminal mCFP and mYFP did not affect ligand binding by WT or D120A transfectants (Fig. 5 A).

Fig. 5.
Fig. 5.

Ligand binding to activated αL-mCFP/β2-mYFP causes spatial separation of the cytoplasmic domains of WT αLβ2 but not the ADMIDAS mutant. (A) The C-terminal mCFP and mYFP fusions have no effect on the function of αLβ2. The binding of ICAM-1–Fcα complexes to 293T cell transfectants was measured in the presence of 1 mM CaCl2/1 mM MgCl2 and 1 mM MnCl2. Error bars show the SD of three independent experiments. (B) Intersubunit FRET efficiency was measured in K562 stable transfectants expressing WT and D120A ADMIDAS mutant αL-mCFP/β2-mYFP. Cells were treated in L15 medium supplemented with 2.5% FCS (control), 1 mM MnCl2 plus 100 μg/ml sICAM-1 (Mn/sICAM-1), or 20 μg/ml αLβ2-activating mAb CBR LFA-1/2 (CBR LFA-1/2). Data show mean ± SD for 5–10 cells.

Significant FRET between mCFP and mYFP was seen with WT αL-mCFP/β2-mYFP transfectants in L15 medium, which contains Ca2+ and Mg2+ (Fig. 5 B, control). Activation with 1 mM Mn2+ and 100 μg/ml sICAM-1 greatly decreased FRET (Fig. 5 B), as described in ref. 11. In contrast, 1 mM Mn2+ and 100 μg/ml sICAM-1 did not decrease FRET of the ADMIDAS D120A mutant, despite evidence that it is more active in ligand binding than WT in Mn2+ (Fig. 2). As a positive control, we used CBR LFA-1/2, a mAb that binds to the β2 integrin EGF3 domain, activates ligand binding, and induces integrin extension (7, 22). CBR LFA-1/2 decreased FRET of WT αLβ2, as reported in ref. 11, and decreased FRET of the D120A mutant αLβ2 to a similar level (Fig. 5 B). These results demonstrate that the D120A ADMIDAS mutation disrupts outside-in signaling in αLβ2.

ADMIDAS Mutation Blocks αLβ2-Mediated Cell Spreading.

To further examine the requirement of the β2 I domain ADMIDAS for outside-in signaling, αLβ2-mediated cell spreading was studied. Transfectants expressing WT αLβ2 spread on ICAM-1 substrates (Fig. 6 A). Interference-reflection microscopy (IRM) showed an extensive area of close cell–substrate contact (Fig. 6 A). In contrast, the αLβ2 ADMIDAS mutant transfectants did not spread on ICAM-1 (Fig. 6 A) and showed the same area of projection on the substrate as cells in suspension (Fig. 6 B). IRM confirmed close contact of the ADMIDAS mutant transfectants with the substrate, in agreement with their adhesiveness, but the area of contact was much smaller than for WT transfectants and remained round, in contrast to the irregular shape of the cell–substrate contacts of the WT transfectants. These results demonstrate that the β2 I domain ADMIDAS mutation abolishes αLβ2-mediated cell spreading.

Fig. 6.
Fig. 6.

ADMIDAS mutation blocks αLβ2-mediated cell spreading. (A) Differential interference contrast and interference-reflection microscopy images of αLβ2-expressing 293T transient transfectants after adhering at 37°C to immobilized human ICAM-1–Fcγ (for 2 h) in DMEM without FCS. The localization of the yellow fluorescent protein-fused β2-subunit and actin stained with phalloidin is also shown. The images are representatives from one of three independent experiments. (B) Cell area (projection on the substrate) was measured as described in Materials and Methods.

Discussion

We have examined the role of the LMA sites in the β I domain of the β2 subunit in regulating ligand binding, integrin extension, and outside-in signaling. As previously demonstrated (26), mutation of the β2 I MIDAS abolishes ligand binding by αLβ2. Mutation of αL Glu-310 (27) and the corresponding residue in αM (28) shows that this residue, which follows the α I domain C-terminal helix and is invariant in I domain-containing integrins, is required for I domain activation. Cross-linking of this residue to loops surrounding the MIDAS activates ligand binding by αLβ2 (18) and supports the proposal that, when activated, the β2 I domain MIDAS coordinates Glu-310 of the αL I domain and activates the αL I domain by exerting a downward pull on its α7-helix (18, 28). According to this model, β I domains of integrins that lack or contain α I domains function similarly and bind to extrinsic ligands or an intrinsic ligand, respectively (e.g., Glu-310 in αL).

Mutation of the β2 I domain LIMBS residues almost completely eliminated multivalent soluble ligand binding and adhesion in shear flow. The presence of the β2 LIMBS and MIDAS sites is sufficient for synergy between 50 μM Ca2+ and 50 μM Mg2+ in supporting ligand binding, as shown with an ADMIDAS mutant. In a structure of the αIIbβ3 headpiece determined in the presence of Ca2+ and Mg2+, Ca2+ and Mg2+ ions were assigned at the LIMBS and MIDAS, respectively (1). Assignment was based on the presence of two carbonyl oxygen coordinations at the LIMBS and none at the MIDAS, and the far greater propensity of Ca2+ than Mg2+ for carbonyl oxygen coordination (29). Synergy between LIMBS and MIDAS is readily explicable, based on the high-affinity αIIbβ3 headpiece structure, which shows that the Glu-220 side chain coordinates with one of its oxygens to the LIMBS Ca2+ ion and with its other oxygen to the MIDAS Mg2+ ion (Fig. 1 B). Synergy between Ca2+ and Mg2+ is also seen for binding of ligands to ADMIDAS mutant integrins α4β7 (3) and α5β1 (5). LIMBS mutants abolish ligand binding to α4β7 (3) and have previously been reported to abolish ligand binding to αLβ2 (25). These observations demonstrate the similar function of the LIMBS in regulation of ligand binding by integrins that lack and contain α I domains.

Mutation of the ADMIDAS activated ligand binding but abolished conformational linkage between the αL I domain and the αL and β2 cytoplasmic domains, revealing the crucial role for the ADMIDAS in relaying allostery to cytoplasmic domains in integrins. The ADMIDAS mutations activated binding to soluble, multimeric ICAM-1 and to ICAM-1 substrates in shear flow. We did not see binding at levels above WT of the ADMIDAS mutants to ICAM-1 in static adhesion/washing assays (data not shown), which is in agreement with another report on D120A mutation (25). Assay-dependent differences may be explained by activation of ligand binding but not extension or cellular spreading by the ADMIDAS mutations; the shear flow LFA-1 adhesion assay used here is markedly more sensitive than adhesion assays in which cells are subjected to high and poorly defined shear during the wash step (15). Mutation of the ADMIDAS prevented inhibition of ICAM-1 binding by high 10 mM Ca2+ concentrations and, as mentioned above, did not affect stimulation of adhesion by low Ca2+ concentrations. These observations, together with the negative and positive regulatory effects of LIMBS and ADMIDAS mutations, respectively, suggest that the LIMBS and ADMIDAS are the sites for positive and negative regulation, respectively, of αLβ2 by Ca2+. Similar observations on α4β7 (3) and α5β1 (5) further support similarity in function between β I domains in integrins that lack and contain α I domains. Studies on ADMIDAS mutants in α5β1 used α5β1 headpiece or α5β1 ectodomain fragments fused to Fc (5) and are somewhat difficult to compare with the assays on cell surface αLβ2 presented here, in part because the “WT” α5β1 ectodomain Fc fusion is active, whereas WT cell surface αLβ2 is inactive. The observation that a Mn2+-dependent increase in a hybrid domain activation-dependent epitope was blocked by ADMIDAS mutations in the α5β1 headpiece–Fc fusion was interpreted as suggesting a function for the ADMIDAS in transmitting conformational change from the MIDAS to the hybrid domain (5) and is compatible with our results in full-length αLβ2. However, ADMIDAS mutation in α5β1 inhibits ligand binding (5). This finding contrasts with results in α4β7 and αLβ2 and may suggest a direct role in ligand binding for the ADMIDAS in α5β1, in addition to a regulatory role.

Our most striking finding is that the ADMIDAS is required to link ligand binding to integrin extension, cytoplasmic domain separation, and cellular spreading on adhesive substrates. Despite activation of ligand binding by ADMIDAS mutation, exposure of the KIM127 epitope, which reports integrin extension, was markedly decreased in the D119A mutant and essentially absent in the D120A mutant. Structural studies in αIIbβ3 (1) and mutational studies in other integrins, including αIIbβ3, α5β1, and αLβ2 (30–33), have shown the importance of the α1-helix, β6-α7 loop, and α7-helix in allosteric communication between the β I and hybrid domains. The ADMIDAS comprises residues in the α1-helix and β6-α7 loop (Fig. 1 B). In the high-affinity, liganded integrin conformation, the α1-helix straightens and moves laterally, displacing the β6-α7 loop (Fig. 1 B) and facilitating the downward, axial displacement of the α7-helix that triggers hybrid domain swing-out (1). However, ligand binding can be dissociated from all of these rearrangements except those immediately in the ligand-binding site when hybrid domain swing-out is prevented by crystal lattice contacts (2). Similarly, our results suggest that ADMIDAS mutations facilitate shifts at the MIDAS of α1-helix residues S114 and S116 toward the ligand-binding configuration and, at the same time, release structural constraints on the α1-helix imposed by its Ca2+-binding ADMIDAS residues, D119 and D120 (Fig. 1 B), so that the more C-terminal portion of the α1-helix and the β6-α7 loop and α7-helix can retain conformations similar to those in the low-affinity state.

Our study shows that the ADMIDAS is required for signal transmission all of the way to the integrin α- and β-subunit cytoplasmic domains. FRET studies showed that, despite activation by ADMIDAS mutation of ligand binding, and in agreement with lack of integrin extension reported by the KIM127 antibody, the α- and β-subunit cytoplasmic domains did not separate. Furthermore, high concentrations of soluble ICAM-1, which induced cytoplasmic domain separation by WT αLβ2, failed to do so with the ADMIDAS mutant. Perhaps most remarkably, despite adhesion to ICAM-1 substrates, the ADMIDAS mutant transfectants completely failed to spread, as shown by visualizing cells, αLβ2, the actin cytoskeleton, and regions of close contact with the substrate. Spreading is an active process that requires cooperative interactions between integrin cytoplasmic domains and cytoskeletal components and is one of the key “postligand-binding” or “outside-in signaling” events triggered by integrins. Thus, our results demonstrate that the ADMIDAS in the β2 I domain not only is an important regulator of ligand binding by the αL I domain but also is required to couple ligand binding to αL and β2 cytoplasmic domain separation and downstream signals that trigger cellular spreading on ligand-bearing substrates.

Materials and Methods

cDNA Construction and Expression.

The β2 site-directed mutations were generated with the QuikChange site-directed mutagenesis system (Stratagene, La Jolla, CA). WT human β2 cDNA in vector pcDNA3.1(+) (Invitrogen, Carlsbad, CA) was used as the template. All mutations were confirmed by DNA sequencing. The αLβ2 constructs used for FRET were exactly as described in ref. 11, except for introduction of the D120A mutation.

Cells Lines and Antibodies.

Transient transfection of 293T and stable transfection of K562 cells were performed as described in refs. 11, 34, and 35. Activating mouse anti-human mAb CBR LFA-1/2 specific for β2 is described in refs. 22 and 36. Activation-dependent mAbs KIM127 (37) and m24 (23, 24) were kindly provided by M. Robinson (Celltech, Slough, U.K.) and N. Hogg (Imperial Cancer Research Fund, London, U.K.), respectively. Mouse anti-human αL mAb TS2/4 was used to measure the cell surface expression of αLβ2.

Immunofluorescence Flow Cytometry.

Immunofluorescence flow cytometry was performed as described in ref. 35. mAbs were used as 10 μg/ml purified IgG. Binding of KIM127 to cells was in Hepes saline (20 mM Hepes, pH 7.5/150 mM NaCl) at 37°C, supplemented as indicated and detected by FITC-conjugated anti-mouse IgG (Pharmingen, San Jose, CA). Binding of other mAbs to cells was in Hepes saline at 4°C, supplemented as indicated and detected by FITC-anti-mouse IgG.

Soluble ICAM-1.

Binding of soluble human ICAM-1-IgA Fc fusion protein (ICAM-1–Fcα) (38) complexed with affinity-purified, FITC-conjugated anti-human IgA (Zymed, South San Francisco, CA) or phycoerythrin (PE)-conjugated anti-human IgA (Open Biosystems, Huntsville, AL) was measured by flow cytometry (16). PE-conjugated anti-human IgA was used for αL-mCFP/β2-mYFP transfectants. Soluble ICAM-1 (sICAM-1, residues 1–453) was expressed in CHO Lec 3.2.8.1 cells and purified as described in ref. 39. Before use, sICAM-1 was purified by using Superdex 200 (GE Healthcare, Pittsburgh, PA) permeation chromatography to ensure that it was monomeric.

Flow Chamber Assay.

The flow chamber assay was performed as described in ref. 3. A polystyrene Petri dish to be used as the lower wall of the chamber was coated with a 5-mm-diameter, 20-μl spot of 10 μg/ml purified human ICAM-1/Fcγ (R & D Systems, Minneapolis, MN) in coating buffer (PBS/10 mM NaHCO3, pH 9.0) for 1 h at 37°C, followed by 2% human serum albumin in coating buffer for 1 h at 37°C to block nonspecific binding sites (3, 40). EDTA-prewashed cells were diluted to 1 × 106 per ml in Ca2+- and Mg2+-free Hanks' balanced salt solution/10 mM Hepes/0.5% BSA containing different divalent cations immediately before infusion in the flow chamber. Cells were allowed to accumulate for 30 s at 0.3 dyne·cm−2 and for 10 s at 0.4 dyne·cm−2. Then, shear stress was increased every 10 s from 1 dyne·cm−2 up to 32 dyne·cm−2 in 2-fold increments. The number of cells remaining bound at the end of each 10-s interval was determined. Rolling velocity at each shear stress was calculated from the average distance traveled by rolling cells in 3 s.

FRET.

FRET was measured exactly as described in ref. 11. Cells (106 per ml) were washed and resuspended with 1 ml of L15 medium supplemented with 2.5% FCS and incubated in six-well plates with or without 1 mM MnCl2/100 μg/ml sICAM-1 or 20 μg/ml αLβ2-activating mAb CBR LFA-1/2 at 37°C for 30 min. Then, cells were fixed with 3.7% formaldehyde/PBS for 10 min at room temperature.

Cell Spreading and Microscopy.

Cell spreading assays were carried out in Delta T imaging chambers (Bioptechs, Butler, PA). The chambers were coated for 1 h at 37°C with 30 μg/ml protein A in coating buffer, followed by blocking with 1% BSA for 1 h at 37°C; then, they were coated with 20 μg/ml human ICAM-1–Fcγ (from R & D Systems) in coating buffer for 1 h at 37°C. Transiently transfected 293T cells were detached with trypsin/EDTA, washed once with DMEM containing 0.5 mg/ml soybean trypsin inhibitor (Calbiochem, La Jolla, CA), and washed twice with DMEM. Cells were suspended in DMEM and seeded on ICAM-1–Fcγ-coated chambers. After incubation at 37°C for 2 h, cells were washed three times with PBS at pH 7.4 and fixed with 3.7% formaldehyde in PBS at room temperature for 5 min. The fixed cells were stained with phalloidin conjugated to Alexa Fluor 546 (Molecular Probes, Eugene, OR) after permeabilization with 0.05% Triton X-100 in PBS.

Differential interference contrast, interference-reflection microscopy, and fluorescence imaging was conducted on an Axiovert S200 epifluorescence microscope (Zeiss MicroImaging, Thornwood, NY) by using a ×63 oil objective coupled to an Orca CCD camera (Hamamatsu, Hamamatsu City, Japan). For the quantification of cell spreading, outlines of at least 100 adherent cells and cells in suspension (control) from randomly selected fields in each of three separate experiments were generated, and the number of pixels contained with each of these regions was measured by using OpenLab software (Improvision, Lexington, MA).

Acknowledgments

This work was supported by grants from the National Institutes of Health (HL48675 and CA31798 to T.A.S.) and the Arthritis Foundation (to C.V.C.).

Footnotes

  • §To whom correspondence should be addressed. E-mail: springeroffice{at}cbr.med.harvard.edu

  • ↵ †Present address: Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences, Shanghai 200031, China.

  • ↵ ‡Present address: Division of Surgical Research, Brown University School of Medicine, Rhode Island Hospital, 593 Eddy Street, Middlehouse 207, Providence, RI 02903.

  • Author contributions: J.C., W.Y., and T.A.S. designed research; J.C., W.Y., M.K., and C.V.C. performed research; J.C., M.K., C.V.C., and T.A.S. analyzed data; and J.C., W.Y., and T.A.S. wrote the paper.

  • Conflict of interest statement: No conflicts declared.

  • Abbreviations:
    MIDAS,
    metal ion-dependent adhesion site;
    ADMIDAS,
    adjacent to MIDAS;
    LIMBS,
    ligand-induced metal-binding site;
    LMA,
    LIMBS, MIDAS, and ADMIDAS;
    ICAM-1,
    intercellular adhesion molecule-1;
    mCFP,
    monomeric mutant of cyan fluorescent protein;
    mYFP,
    monomeric mutant of yellow fluorescent protein
Back to top
Article Alerts
Email Article
Citation Tools
Request Permissions
Regulation of outside-in signaling and affinity by the β2 I domain of integrin αLβ2
JianFeng Chen, Wei Yang, Minsoo Kim, Christopher V. Carman, Timothy A. Springer
Proceedings of the National Academy of Sciences Aug 2006, 103 (35) 13062-13067; DOI: 10.1073/pnas.0605666103
del.icio.us logo Digg logo Reddit logo Twitter logo CiteULike logo Facebook logo Google logo Mendeley logo
Proceedings of the National Academy of Sciences: 103 (35)
Table of Contents

Submit

Jump to section

You May Also be Interested in

Similar Articles