A high affinity conformational state on VLA integrin heterodimers induced by an anti-beta 1 chain monoclonal antibody.

The VLA integrin subfamily includes receptors for extracellular matrix proteins as well as receptors involved in cell-cell adhesive interactions. We have previously described the up-regulation of VLA integrin-mediated cell attachment to different ligands by the anti-beta 1 TS2/16 monoclonal antibody (mAb) (Arroyo, A. G., Sánchez-Mateos, P., Campanero, M. R., Martín-Padura, I., Dejana, E., and Sánchez-Madrid, F. (1992) J. Cell Biol. 117, 659-670). In this report, we have investigated the mechanism involved in this regulatory effect. The anti-beta 1-mediated regulatory effect on cell adhesion did not require "de novo" protein synthesis, since it was not affected by pretreatment with either cycloheximide or actinomycin D. To quantitate the effect of the regulatory anti-beta 1 TS2/16 mAb on the affinity of VLA-5 for its ligand, an RGD-containing fragment of fibronectin (FN80), we performed binding studies of radiolabeled soluble FN80 to U-937 cells. The affinity of VLA-5 for FN80 was enhanced about 4-fold in the presence of TS2/16 mAb (Kd = 0.98 +/- 0.07 microM) compared to the functionally irrelevant anti-beta 1 Alex 1/4 mAb (Kd = 4.23 +/- 0.92 microM), whereas no alteration in the number of binding sites was observed. Indeed, the anti-beta 1 TS2/16 mAb is inducing this high affinity state on VLA heterodimers by a direct change on the conformation of these receptors as demonstrated by affinity chromatography analysis using extracellular matrix proteins covalently bound to Sepharose. The yield of VLA-5 fibronectin receptor bound to FN80-Sepharose columns was strongly increased upon treatment of U-937 cell lysates with mAb TS2/16. Moreover, higher concentrations of EDTA were required for eluting the VLA-5 integrin from this matrix. This up-regulatory effect was also observed with F(ab')2 and Fab fragments of the anti-beta 1 TS2/16 mAb, and was also exerted on the purified VLA-5 receptor. Similarly, the yield of VLA-2 retained on a collagen I-Sepharose column was dramatically increased by pretreatment of A375 melanoma cell lysates with the mAb TS2/16. Altogether, these results indicate that the interaction of VLA beta 1 heterodimers with their ligands can be regulated by switching between differently active conformations inherent to the alpha beta 1 receptors.

J. Cell Biol. 117,[659][660][661][662][663][664][665][666][667][668][669][670]. In this report, we have investigated the mechanism involved in this regulatory effect. The anti-B1-mediated regulatory effect on cell adhesion did not require "de novo" protein synthesis, since it was not affected by pretreatment with either cycloheximide or actinomycin D. To  Indeed, the anti-Bl TS2/16 mAb is inducing this high affinity state on VLA heterodimers by a direct change on the conformation of these receptors as demonstrated by affinity chromatography analysis using extracellular matrix proteins covalently bound to Sepharose. The yield of VLA-5 fibronectin receptor bound to FNSO-Sepharose columns was strongly increased upon treatment of U-937 cell lysates with mAb TS2/16. Moreover, higher concentrations of EDTA were required for eluting the VLA-5 integrin from this matrix. This up-regulatory effect was also observed with F(ab')z and Fab fragments of the anti-B1 TS2/16 mAb, and was also exerted on the purified VLA-5 receptor. Similarly, the yield of VLA-2 retained on a collagen I-Sepharose column was dramatically increased by pretreatment of A375 melanoma cell lysates with the mAb TS2/16. Altogether, these results indicate that the interaction of VLA B1 heterodimers with their ligands can be regulated by switching between differently active conformations inherent to the crB1 receptors. 'This work was supported in part by Grant FIS 91/0259 from Instituto Nacional de la Salud (INSALUD) and a grant from Fun-daci6n Ram6n Areces (to F. s. M.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. J Supported by Grant SAL91/0785 from the Comisi6n Interministerial de Ciencia y Tecnologia, Madrid.
7l To whom correspondence should be addressed Secci6n de Immunologia, Hospital de la Princesa, Universidad Aut6noma de Madrid, C/Diego de Le6n n 62, 28006 Madrid, Spain.
Cells communicate with environment through several classes of cell surface receptors. One of the most important families of adhesion receptors are integrins, which include receptors for extracellular matrix proteins as well as receptors involved in cell-cell adhesive interactions (1,2). All integrins are a8 heterodimers in which the a subunits are each noncovalently associated with one p subunit (3). Classically, integrins are grouped on three subfamilies depending on the common p subunit: (VLAs),' PZ (LeuCAMs), and P 3 (CYtoadhesins) subfamilies (4). Recently, novel associations of a and alternative p subunits have been found which extend the functional versatility of integrins (1). The VLA-integrin subfamily is composed of at least eight heterodimers that behave essentially as receptors for several ECM proteins. Thus, the integrins VLA-2, VLA-5, and VLA-6 are prototype receptors for collagen, fibronectin, and laminin, respectively (3). Recently, it has been described from the ability of VLA-2 heterodimer to bind laminin in some cell types (5). Other VLA heterodimers such as VLA-1 and VLA-3 can bind more than one ECM component (3,6,7). Given the wide variety of integrins, individual cells can vary their adhesive properties by the selective expression of integrin repertoire. Further versatility is introduced by the ability of cells to modulate the binding properties of integrins. Thus, the affinity of integrins for their ligands is not always constant and can be dynamically regulated. It is well known that activation by different stimuli of platelets, neutrophils, and lymphocytes, up-regulates the avidity for their ligands of P3, P2, and P,-integrin subfamilies, respectively (8)(9)(10). Possible mechanisms accounting for these effects are either triggering of intracellular signals that activate the integrin receptor or induction of a conformational change on the heterodimer that enables or enhances the interaction with ligand(s). This conformational change has been demonstrated on soluble GP IIb-IIIa (p3 subfamily) receptor whose affinity can be modulated by mAb as well as by physiologic ligands (11,12); the affinity of purified CR3 or Mo-1 (Pz subfamily) for its ligands is also enhanced when treated with a lipid termed IMF-1 (13). For the ( 3, subfamily, only one report has described up-regulation of affinity of soluble fibronectin receptor after treatment of lysates with MnZ+ cations, and the effect was probably occurring at the level of the divalent cation-binding domains of the LY subunit (14).
We have previously described the functional regulatory effects of the anti-& TS2/16 mAb that enhances the VLAmediated cell avidity to different substrates (15). To further investigate the mechanism of action of this mAb, we have The abbreviations used are: VLA, very late activation antigen; BSA, bovine serum albumin; COL, collagen; ECM, extracellular matrix; FN, fibronectin; PAGE, polyacrylamide gel electrophoresis; PBS, phosphate-buffered saline; mAb, monoclonal antibody; CAPS, 3-(cyclohexylamino)propanesulfonic acid GP, glycoprotein.

9863
performed measurements of affinity of VLA integrins for their ligands by Scatchard analysis. In addition, the possibility of the induction of a conformational change on the heterodimer has been also explored by using affinity chromatography assays with TS2/16 mAb-pretreated cell lysates. In this report, we show that this anti-& mAb is enhancing not only the avidity but the affinity of VLA heterodimers for their ligands and is exerting this functional effect by inducing a conformational change on different VLA a& heterodimers when present either on intact cells or in detergent soluble conditions.
MAbs, F(ab'), and Fab Fragments-The anti-& Alex 1/4 and TS2/ 16 mAb have been previously described (18,19). Quantitative studies of binding of the two anti-P1 Alex 1/4 and TS2/16 mAb to U-937 cells by cytofluorometry analysis demonstrated that binding titers obtained were equivalent for both mAbs in a wide range of doses from 50 ng/ml to 50 pg/ml. The anti-as SAM-1 was purchased from Central Laboratory of Blood Transfusion Service (Netherland Red Cross, Amsterdam, The Netherlands).
Anti+% mAbs were purified using affinity chromatography on protein A-Sepharose columns while F(ab'), and Fab fragments were prepared as previously reported (20).
Cell Attachment Assays-The adhesion assays were performed essentially as previously described (15). Briefly, 96-well flat-bottomed plates (EIA, Costar) were coated overnight at 4 "C with different substrates diluted in phosphate-buffered saline (PBS) with divalent cations. Thereafter, the plate was washed once with PBS and then saturated with 1% BSA for 1 h at 37 "C. The plate was washed twice with PBS and cells (preincubated or not with different reagents for 30 min at 37 "C) were added in serum-free medium and incubated for 30 min at 37 "C. To quantify cell attachment, the plate was washed twice with PBS with divalent cations, cells were fixed with a mixture of acetone/methanol, 1:1, and dyed with violet crystal, 0.5%. Absorbance at 540 nm was measured in a spectrophotometer (LP400, Kallestad, Chaska, MN) and optical density was found to be practically a linear function of number of cells by a calibration curve (optical density vs number of cells) made for each cell type used in OUT assays. To calculate percentage of attachment, basal adherence to BSA was subtracted from the attachment values obtained. The assays were performed in triplicate. Total cellular input was calculated by spinning wells with the original number of cell aliquots, then fixing, staining, and measuring optical density.
Binding of "='I-Labeled 80-kDa FN Fragment to U-937 Cells-The proteolytic 80-kDa fragment of fibronectin was radioiodinated using chloroglycoluril (IODO-GEN, Pierce Chemicals Co.), and integrity of radiolabeled fragment was assessed by SDS-PAGE and autoradiography. U-937 cells were washed once with RPMI, and resuspended in RPMI containing 25 mM Hepes and 1.5% BSA (binding medium) at 10' cells/ml. Aliquots of lo7 cells were pretreated with either 1 pg/ml anti-fl, Alex 1/4 or TS2/16 mAb for 15 min at 4 "C. Different amounts of 1261-labeled FNSO (specific activity of 2.5 X lo6 cpm/pg) were added alone or with a 10-fold excess of unlabeled fragment. Then, the suspension was incubated at room temperature for 30 min on a rotatory shaker. A t the end of the incubation, cells were washed three times in binding medium. The final pellet was counted in a y-counter. Total and specific binding, as well as equilibrium constants, were calculated by Scatchard analysis.
Affinity Chromatography-FN8O and COL I were covalently coupled at 2 mg/ml CNBr-activated Sepharose CL-4B (Pharmacia Fine Chemicals, Uppsala, Sweden) according to the manufacturer's instructions. U-937 myelomonocytic or A375 melanoma cells were labeled with sulfo-N-hydroxysuccinimido-biotin (Pierce Chemical Co.) at 0.4 mg/ml in PBS. Then, cells were lysed with 100 mM noctyl glucoside, 50 mM Tris-HC1, pH 7.5, 150 mM NaCl, 1 mM CaC12, 1 mM MgC12, and 3 mM phenylmethylsulfonyl fluoride for 30 min on ice. Lysates were centrifuged for 30 min and the supernatant was collected. Lysates were precleared with glycine-Sepharose, and either untreated or pretreated with 50 pg/ml purified Ig of either Alex 1/4 or TS2/16 mAb unless otherwise indicated. Then, lysates were applied to either the 80-kDa or COL I-Sepharose matrices previously saturated with 1% BSA and equilibrated with column buffer (25 mM noctyl glucoside, 50 mM Tris-HC1, pH 7.5, 150 mM NaC1, 1 mM CaC12, and 1 mM MgC12). Lysates were incubated for 30 min at room temperature with the affinity matrices. After this, the matrix was washed with 6 volumes of column buffer, and bound receptors were eluted with column buffer without divalent cations, and different doses of either EDTA or RGES and RGDS synthetic peptides. Finally, the columns were washed with 6 M urea. Fractions of 0.5 ml were collected from columns. In some experiments (indicated in legends to figures) incubation with the matrix was performed in microcolumns (Eppendorf tubes) and elution of the heterodimer was achieved by boiling Sepharose to recover all the receptor bound to the matrix. Volume-equivalent samples of fractions were analyzed by SDS-7% PAGE under nonreducing or reducing conditions, and electrotransferred onto nitrocellulose membrane (Bio-Rad) in 10 mM CAPS, pH 11, as transfer buffer, for 12 h at 0.2 A, 50 V at 4 "C. Protein bands were visualized by incubation with horseradish peroxidase-streptavidin complex (Amersham, United Kingdom) and detection with the enhancer chemiluminescence Western blotting protocol developed by Amersham (Amersham). Filters were immediately exposed to film for 1 min to 1 h. Quantitative estimation of bands were performed by densitometric analysis with Image 1.3 software.
Immunoprecipitation-Lysates were immunoprecipitated with monoclonal antibodies as previously described (21) and samples subjected to SDS-7% PAGE under nonreducing conditions, and then electrotransferred and developed as described.

PI Up-regulated Cell Attachment to Different ECM Proteins
Is Independent of "de Novo" Protein Synthesis-The ability of the anti-@, TS2/16 mAb in up-regulating the cell attachment to different ECM ligands has been previously described (15). The functional effects exerted by this anti-& mAb on cellular attachment did not involve any associated change in cell surface expression of the different VLA heterodimers and followed a very rapid kinetics (15). T o ascertain whether 0,mediated enhanced adhesion was independent of de novo protein synthesis, we tested the effect of cycloheximide and actinomycin D, inhibitors of protein and RNA synthesis, respectively, on P1-mediated enhanced attachment of either U-937 cells to FN80 or A375 melanoma cells to COL I. As shown in Fig. 1, cell treatment with either cycloheximide or actinomycin D did not affect the anti-P1 enhanced binding of U-937 cells to FN80 or the binding of A375 cells to COL I.
The lack of effect of these inhibitors on the anti-Pl mediated attachment of U-937 cells to FN80 was observed in a wide range of substrate concentrations (from 1 to 25 Kg/ml, data not shown). Cycloheximide and actinomycin D were active inhibitors of protein and RNA synthesis on these assays as demonstrated by inhibition of [35S]methionine (>85%) and [3H]thymidine (>94%) incorporation to U-937 cells, respectively. These results demonstrate that de nouo protein synthesis is not required for P1-mediated regulatory effects on cell adhesion to ECM proteins.
Anti-& TS2/16 MAb Enhances the Affinity of Surface Bound VLA-5 for Soluble '251-Labeled FNBO Fragment-To quantitatively estimate the effect of anti$, TS2/16 mAb on VLA-ligand interactions, we performed binding studies using soluble lz5I-1abeled FN80, a fragment which contains the RGD binding sequence for the VLA-5 integrin (17) and U-937 cells. As shown in Fig. 2 A , specific binding of FN80 to U-937 cells U-937 myelomonocytic and A375 melanoma cells were untreated or treated with either 10 pg/ml cycloheximide or 5 pg/ml actinomycin D for 30 min. Then, cells were preincubated with 1 pg/ml purified anti-& mAb, and assayed for attachment to plates coated with either 5 pg/ml FN80 or 10 pg/ml COL I for U-937 and A375 cells, respectively. Control without antibody is also included. A representative experiment out of three independent ones is shown. S.D. within triplicates was less than 10%.
was enhanced when cells were preincubated with the activatory TS2/16 mAb compared to the functionally irrelevant anti-P1 Alex 1/4 mAb used as control. Scatchard analysis of these data (Fig. 2B) (15). These previous results, together with the independence of protein synthesis reported here (Fig.  l), suggested that TS2/16 mAb could be enhancing the affinity of VLA heterodimers for its ligands (Fig. 2) by causing a conformational change at the extracellular part of VLA apl heterodimers. To assess this possibility, we performed affinity chromatography experiments using U-937 cell lysates and Sepharose conjugated to FN80. As shown in Fig. 3A, the yield of VLA-5 eluted from the FN80-Sepharose matrix was strongly increased by pretreatment of the cell lysates with the TS2/16 mAb (lanes 3-6) when compared to the yield of VLA-5 obtained from cell lysates which had been either untreated or treated with the functionally irrelevant anti& Alex 1/4 mAb. Densitometric analysis of bands corresponding to a6 and P1 chains from the different specifically eluted fractions is graphically represented (Fig. 3C). As shown, about %fold greater amounts of VLA-5 heterodimer can be eluted when lysates were pretreated with TS2/16 mAb compared to Alex 1/4 mAb. Furthermore, after treatment of lysates with TS2/ 16, higher doses of EDTA were required to completely elute the a5P1 heterodimer from the column. These results clearly indicate that the TS2/16 mAb was enhancing the affinity of fibronectin receptor for its ligand by inducing a conformational change on the heterodimer.
TO ascertain the specificity of this effect, a similar assay was performed using increasing doses of a synthetic peptide containing the RGD sequence as eluent (Fig. 3B). The yield of VLA-5 heterodimer was approximately 2-fold greater when shown. E , Scatchard analysis of the specific binding. The data were fit by linear regression analysis using the least square method. The linear correlation coefficient for the analysis is 0.97 for the TS2/16 mAb line (0) and 0.6 for Alex 1/4 mAb line (A). U-937 cell lysates were pretreated with TS2/16 mAb compared to treatment with Alex 1/4 mAb, as determined by densitometric analysis (Fig. 3 0 ) . These and the previous results from Scatchard analysis strongly suggest that the mechanism by which TS2/16 mAb is inducing the high affinity conformation on VLA-5 for FN80 is by enhancing the affinity of pre-existing binding sites; however, the possibility of induction of alternative binding sites for VLA-5 cannot be completely ruled out at this time.
We next studied whether the anti+$ enhanced interaction of VLA-5 with FN80-Sepharose could also be induced on solubilized receptor using both divalent F(ab'), and monovalent Fab fragments of the TS2/16 mAb. Fig. 4A shows that, the yield of VLA-5 was higher (about 2-fold) when the lysates were pretreated with purified Ig, F(ab')2, or Fab fragments of TS2/16 as compared with lysates pretreated with purified Ig from Alex 1/4 mAb (Fig. 4A). These results support the view that TS2/16 mAb is inducing a conformational change at the extracellular domain of the receptor which facilitates ligand binding without requiring receptor cross-linking.
To investigate whether the effect of TS2/16 mAb could also be induced on the purified a5PI heterodimer in the absence of other protein or lipid cellular components, VLA-5 was purified Immunoprecipitates from U-937 cell lysates with either an irrelevant mAb or a n t i a s SAM-1 are also shown (lanes u and b, respectively). c and D, densitometric analysis of bands from specific EDTA elution (A, lanes 3-6) of FN80-Sepharose chromatography with U-937 cell lysates pretreated with either Alex 1/4 or TS2/16 mAb (C), and from specific RGD elution ( B , lanes [4][5][6][7] under the same conditions ( D ) . The mean of two independent measurements is shown. Data are represented in arbitrary density units.
from U-937 cell lysates by affinity chromatography on FN80-Sepharose as described above. Purified soluble receptor was incubated with TS2/16 mAb and re-applied to the FN80-Sepharose affinity matrix. As shown in Fig. 4B (lanes 2 and  3 ) , the yield of fibronectin receptor was higher (about 2.5fold) by using the regulatory anti-P1 TS2/16 as compared with the anti-& Alex 1/4 treatment as control. This result further demonstrates that TS2/16 mAb is exerting its regulatory effect on cell adhesion by directly acting on the integrin and it rules out an indirect effect on other associated cellular components.

The Anti-& TS2/16 MAb Reinforces Weak Interactions of Integrins with Their Ligands: Effect on Binding of A375 Melanoma Cell Lysates to COL I-Sepharose
Columns-We next examined the effect of TS2/16 mAb on other VLA integrinligand interactions. For this purpose, we performed affinity chromatography using A375 melanoma cell lysates and COL I-Sepharose. Collagen-binding integrin heterodimers were eluted from matrices incubated with lysates pretreated with TS2/16 mAb (Fig. 5 A ) . This clearly contrasted with eluates from cell lysates pretreated with Alex 1 / 4 where almost no integrin molecules were detected (Fig. 5 A ) .
As A375 melanoma cells express two integrins known to function as receptors for COL I, VLA-2 and VLA-3, we assessed which of the two heterodimers was involved in the &-mediated interaction. Analysis by SDS-7% PAGE under reducing conditions of the eluted fractions from the COL I-Sepharose matrix showed that the main heterodimer eluted from A375 cell lysates pretreated with TS2/16 mAb had a molecular weight pattern consistent with the VLA-2 integrin (Fig. 5B).

DISCUSSION
In this report, we demonstrate the regulation of the affinity of VLA PI heterodimers for their ligands by induction of a  (lanes 3 and 4, respectively). Then, lysates were incubated with FN80-Sepharose. After washing three times with column buffer, the FN8O-Sepharose matrix was boiled in sample buffer, and analyzed by SDS-7% PAGE under nonreducing conditions. B, TS2/16 mAb enhances binding of purified VLA-5 fibronectin receptor to FN80-Sepharose. VLA-5 receptor was purified from U-937 cell lysates incubated with FN80-Sepharose matrix in microcolumns by elution with 50 mM EDTA (lane 1 shows a small aliquot of this eluate analyzed by SDS-7% PAGE under nonreducing conditions). This purified receptor was then pretreated with 50 pg/ml of either Alex 1/ 4 (lane 2) or TS2/16 (lane 3 ) mAb, and re-applied to FN80-Sepharose microcolumn. After washing three times with column buffer, Sepharose matrix was boiled in sample buffer, and analyzed by SDS-7% PAGE under nonreducing conditions (lanes 2 and 3 ) . Immunoprecipitates from U-937 cell lysates with either an irrelevant mAb or antia 6 SAM-1 are also shown (lanes a and b, respectively). high affinity conformational state through the anti-P1 TS2/ 16 mAb. This effect was observed both on intact cells as determined by binding studies with soluble ligand, and most importantly, on detergent-solubilized receptor. Hence, we have found that anti-P1 TS2/16 mAb strongly increases the binding interaction of VLA-5 fibronectin receptor present in U-937 cell lysates to FN80-Sepharose. The changes in affinity caused by the anti-& TS2/16 mAb were even more evident when the interaction of other VLA heterodimers such as VLA-2 present in A375 melanoma cell lysates was assayed using COL I-Sepharose affinity matrices. The fact that the upregulated interaction can be achieved not only on intact cells but also on detergent-solubilized receptor points out to a conformational change as the mechanism accounting for the adhesive promoted effects of the anti-P1 mAb. In a previous report (15), cellular adhesion appeared to be blocked by metabolism and cytoskeletal (cytochalasin B) inhibitors. We have recently observed that integrity of cytoskeleton is required for cellular spreading, but not for TS2/ 16-induced attachment, by using a more specific inhibitor of cytoskeleton, cytochalasin D, and milder washing conditions during cell adhesion assays.' Our results on the quantitative estimation of the apparent dissociation constant of VLA-5/FN-80 interaction upon different anti-& mAb treatment revealed Kd values in the range of previous reports (17,26,27). Scatchard analysis of the data clearly demonstrate that the anti& TS2/16 mAb is increasing the affinity (about 4-fold) of VLA-5/FN80 interaction compared to Alex 1/4 mAb as control. This effect appeared to occur without modification of the number of single-class binding sites on the heterodimer. Thus, these results point out to a direct enhancement of the affinity of pre-existing binding sites on the VLA-5 receptor as the mechanism involved in TS2/16 mAb-mediated effects.

Different Affinity
The existence of different affinity conformational states for * A. G. Arroyo and F. Sanchez-Madrid, unpublished observations. P3-and &integrin subfamilies has been previously reported. Conformational changes on aIIbp3 receptor can be induced by agonist-induced activation (22), engagement with anti-aIIbP3 mAbs (23), and physiologic ligands such as small synthetic peptides or fibrinogen (12, 24, 25). These conformational changes can also be observed on the purified receptor (12) and are responsible for the enhanced fibrinogen-binding functions. However, there is not a previous demonstration that the affinity of P1-integrin heterodimers for their ligands can be regulated through induction of a conformational change. Our present results showing that the TS2/16 mAb can directly enhance the interaction of solubilized &-integrin receptors with ECM proteins conjugated to Sepharose, therefore represents the first evidence for such regulation.
The fact that the high affinity conformational state can be induced not only with whole purified Ig of TS2/16 mAb but also with F(ab'), and Fab fragments indicates that the upregulating effect is independent of cross-linking or clustering of solubilized receptors. It is conceivable that these smaller Fab fragments could be mimicking the functional effects of still undiscovered soluble physiologic ligands for P,-integrins that acting on the "non-activated state" of aP1 heterodimer would change the conformation to a fully active form able to bind ligands with high affinity. This mechanism could play an important role in the regulation of cell interactions with endothelium and ECM proteins in inflammatory and metastatic foci. Alternatively, binding of VLA receptors to certain sequences on different ligands could resemble the regulatory effect of anti-& TS2/16 mAb and thus, trigger secondary conformational changes. These changes might lead to the expression of additional binding sites for other synergistic sequences in the ligand (as), that may be required for complete post-occupancy effects such as transmembrane intracellular signaling (29), induction of gene expression (30), and spreading and cytoskeletal reorganization (31, 32).
We have herewith provided evidence for the actual mechanism of the previously observed effects of TS2/16, and possibly other anti-P1 mAb, on cell adhesion (15, 33-35). The increased affinity state of PI heterodimers observed has been demonstrated for the interaction of at least two different integrins, VLA-5 and VLA-2 with FN80 and COL I-affinity columns, respectively. However, we have found differences in the level of enhanced binding of different VLA receptors. It is worth mentioning the fact that anti$, TS2/16 mAb not only increased well described strong interactions such as those of VLA-5 with FN80, but that most importantly, has been proven very useful in revealing and reinforcing weak interactions such as binding of VLA-2 to COL I. Thus, the anti-/& TS2/16 mAb could be a useful tool to detect novel low affinity integrin-ligand specific interactions.
Binding of ligands to integrins requires an associated cup heterodimer, presumably to form a functional binding pocket. The TS2/16 mAb could be acting in the proximity to this ligand binding domain and thus, could alter the conformation of the molecule allowing the exposure of a high affinity ligand binding site as it has been described for G P IIb-IIIa (36). The mechanism by which this mAb is inducing the conformational change is still unknown. Although the sequence or domains contributing to the epitope defined by TS2/16 mAb remain undetermined, the epitope recognized by this anti-& mAb is dependent on disulfide bonding as this mAb does not recog-nize the reduced p r~t e i n .~ T h i s is similar to what has been described for the anti-GP IIIa-D3GP3 mAb which enhances binding of this integrin to fibrinogen (23). The ligand binding site for P3-integrins has been localized (37, 38). It would be interesting to determine the region recognized by TS2/16 mAb on p1 and its possible relationship to the ligand-binding pocket.
In summary, we have provided evidence for the existence of different active conformations on P1-integrins which appear to be an intrinsic property of the VLA receptors as these conformations can be observed on soluble and purified receptor. Switching between the distinct active conformations could be achieved by regulation through the PI chain.