L-plastin Peptide Activation of αvβ3-mediated Adhesion Requires Integrin Conformational Change and Actin Filament Disassembly*

L-plastin (LPL) is a leukocyte actin binding protein previously implicated in the activation of the integrin αMβ2 on polymorphonuclear neutrophils. To determine the role for LPL in integrin activation, K562 cell adhesion to vitronectin via αvβ3, a well-studied model for activable integrins, was examined. Cell permeant versions of peptides based on the N-terminal sequence of LPL and the LPL headpiece domain both activated αvβ3-mediated adhesion. In contrast to adhesion induced by treatment with phorbol 12-myristate 13-acetate (PMA), LPL peptide-activated adhesion was independent of integrin β3 cytoplasmic domain tyrosines and was not inhibited by cytochalasin D. Also in contrast to PMA, LPL peptides synergized with RGD ligand or Mn2+ for generation of a conformational change in αvβ3 associated with the high affinity state of the integrin, as determined by binding of a ligand-induced binding site antibody. Although LPL and ligand showed synergy for ligand-induced binding site expression when actin depolymerization was inhibited by jasplakinolide, LPL peptide-induced adhesion was inhibited. Thus, both actin depolymerization and ligand-induced integrin conformational change are required for LPL peptide-induced adhesion. We hypothesize that the critical steps of increased integrin diffusion and affinity enhancement may be linked via modulation of the function of the actin binding protein l-plastin.

A fundamental property of leukocyte integrins is the ability to modulate their adhesive functions. As cells circulate through blood and lymph, adhesion is minimal. In response to inflammatory signals, integrin-mediated adhesion is markedly augmented. This property of the integrins is physiologically critical, because it is required for appropriate migration out of the vasculature into sites of inflammation and at the same time limits the potentially host-damaging inflammatory response to those sites alone. Because this activity of integrins is so important to their role on leukocytes, much attention has been devoted to the molecular mechanisms by which integrin adhesion is regulated. Two distinct alterations in integrins are likely to be involved in enhancement of adhesion. A rapid response to cell activation is an increase in integrin diffusion, due to loss of cytoskeletal constraint of integrin mobility. Increased diffusion may lead in turn to integrin clustering at sites of cell interaction with ligand. A second response to activation is a conformational change in the integrins, which often reflects increased affinity for ligand. This conformational change may require initial interaction with ligand and can be reflected in the generation of new epitopes recognized by monoclonal antibodies, the so-called ligand-induced binding sites (LIBS 1 ). Although a number of signaling molecules, such as PKC and phosphatidylinositol 3-kinase, have been implicated in this process (1)(2)(3), these enzymes may be many steps upstream from the actual change in integrin behavior. GTPases of the Ras and Rho families also have been implicated in regulation of integrin function (4 -6), but their downstream targets for this function have not been identified. Therefore, much remains to be learned about the mechanisms involved in regulation of integrin avidity.
Our own studies have implicated the actin-binding protein L-plastin (LPL) in regulation of integrin function (7,35). Recently, we have shown that cell-permeant peptides from the N terminus of LPL can rapidly activate ␣ M ␤ 2 (Mac-1)-mediated adhesion in polymorphonuclear neutrophils (PMN). When the peptides introduced into PMN cytosol contained phosphoserine at position 5, which is the major if not exclusive site of phosphorylation in LPL (7,35), integrin activation was not inhibited by blockade of PKC or phosphatidylinositol 3-kinase, suggesting that LPL phosphorylation might be a mechanism by which these enzymes signal changes in integrin function. Thus, LPL is likely a downstream effector of multiple signaling pathways leading to integrin activation, and LPL peptide induction of adhesion is a useful experimental system in which to begin to understand regulation of integrin avidity.
To pursue these mechanistic questions, we turned to a well- 1 The abbreviations used are: LIBS, ligand-induced binding site; LPL, L-plastin; Headtat, a recombinant protein containing amino acids 1-105 of L-plastin fused at the C terminus to 12 amino acids of the HIV tat sequence; HSA, human serum albumin; K␣ v ␤ 3 , K562 cells transfected with ␣ v ␤ 3 integrin; LPLtat, a synthetic peptide based on amino acids 2-19 of L-plastin followed by 12 amino acids of the HIV tat sequence; PKC, protein kinase C; PMA, phorbol 12-myristate 13-acetate; PMN, polymorphonuclear neutrophils; RAD, Arg-Ala-Asp peptide; RGD, Arg-Gly-Asp peptide; SCRtat, a synthetic peptide in which amino acids 2-19 of L-plastin have been scrambled, followed by 12 amino acids of the HIV tat sequence; tatHead, a recombinant protein containing amino acids 1-105 of L-plastin fused at the N terminus to 12 amino acids of the HIV tat sequence; tatLPL, a synthetic peptide based on amino acids 2-19 of L-plastin fused at the N terminus to 12 amino acids of the HIV tat sequence; TPLtat, a synthetic peptide based on amino acids 2-19 of T-plastin followed by 12 amino acids of the HIV tat sequence; Vn, vitronectin; FACS, fluorescence flow cytometry; PBS, phosphate-buffered saline; mAb, monoclonal antibody; HBSS, Hanks' balanced salt solution; RT, room temperature; HIV, human immunodeficiency virus. characterized model for studying integrin activation. ␣ v ␤ 3 is not normally expressed in undifferentiated K562 cells, but when expressed through transfection of ␣ v and ␤ 3 cDNAs, ␣ v ␤ 3 -mediated adhesion in K562 cells is dependent on cell activation (8). Moreover, adhesion in response to PMA or thrombin is dependent on phosphorylation of Tyr-747 in the ␤ 3 cytoplasmic tail, an event that is itself dependent on the presence of the ␣ v cytoplasmic tail (8,9). Using these cells, we have examined the requirements for LPL peptide-induced adhesion. We have found that a cell permeant version of the entire headpiece domain of LPL, as well as synthetic LPL peptides, can induce adhesion. In contrast to PMA, LPL peptide-induced adhesion does not require tyrosine phosphorylation of the ␤ 3 cytoplasmic tail, suggesting that LPL is downstream of the tyrosine signal in the activation cascade. Nonetheless, LPL peptide-induced adhesion does require actin depolymerization, presumably to induce integrin release from cytoskeletal constraint (10 -12), and LPL peptide does cooperate with Arg-Gly-Asp (RGD) ligand to induce LIBS epitope expression. These data suggest that the steps of increased integrin diffusion and conformational change may be linked via modulation of the function of the actin binding protein L-plastin.
Tat Fusion Protein Purification-The cDNA containing the LPL Nterminal Ca 2ϩ -binding domain (headpiece fragment, amino acids 1-105) was cloned into pBluescript KSϩ as an XhoI/BamHI fragment. An myc tag was fused in-frame with the C terminus of the headpiece sequence (Headmyc). To obtain the headpiece fragment with the tat sequence at the C terminus (Headtat), the myc sequence was replaced by an oligonucleotide encoding the 12-amino acid tat peptide sequence. The Headmyc fragment also was subcloned into the pTAT vector (a gift of Dr. Steven Dowdy, Washington University, St. Louis, MO) (16) to obtain a headpiece with the tat sequence at the N terminus (tatHead). The constructs were transformed into BL-21 LysS cells (Novagen). Bacteria from 1-liter cultures were resuspended in 50 ml of PBS ϩ 50 g/ml lysozyme. The bacteria were lysed with three freeze-thaw cycles and sonication, and lysates were clarified by centrifugation. The supernatant was run through an immunoaffinity column of LPL4A.1 mAb coupled to Sepharose 4B and eluted with PBS containing 10 mM EDTA as LPL4A.1 binds to LPL in a Ca 2ϩ -dependent manner. EDTA was removed by dialysis.
Adhesion Assay-K␣ v ␤ 3 cells or JY cells (1 ϫ 10 7 /ml) were washed once with HBSS (1ϫ Hanks' buffered salt solution with 20 mM Hepes and 8.9 mM sodium bicarbonate) and incubated with 2 g/ml calcein in HBSS for 30 min at RT. The cells were washed three times and resuspended in HBSS 2ϩ (HBSS with 1.0 mM Mg 2ϩ and 1.0 mM Ca 2ϩ ) at 2 ϫ 10 6 /ml. Cells were subjected to various treatments as detailed in respective figure legends. 1 ϫ 10 5 cells/well were added to HSA-, Vn (10 g/ml or as indicated in the figures)-, or osteopontin (10 g/ml)-coated 96-well Immulon 2 plates. Cells were allowed to settle for 10 min before L-plastin peptides (50 M), PMA (50 ng/ml), or MnCl 2 (1 mM) were added to the wells. Cells were incubated at 37°C for 1 h. After the incubation period, the fluorescence (485-nm excitation and 530-nm emission wavelengths) was measured using an fMax fluorescence plate reader (Molecular Devices, Sunnyvale, CA) before and after washing four times with 180 l of PBS. Percent adhesion was calculated by dividing the fluorescence after washing by the fluorescence before washing. The assays were performed in triplicate, and data are presented as mean Ϯ S.D. All experiments were performed at least three times with similar results. In preliminary experiments, fluorescence was shown to be linearly related to cell number. F-actin Content-F-actin content of K␣ v ␤ 3 cells was determined as previously described (41) with minor modifications. Briefly, cells were incubated with buffer, 50 M LPLtat, 1 M jasplakinolide, or both for 15 min at 37°C and then solubilized by incubation with an equal volume of Tris-buffered saline, pH 7.4, containing 4% Triton X-100, 4 mM EDTA, and 5 g/ml leupeptin. After centrifugation at 38,000 rpm for 30 min in a TL100.2 rotor (Beckman Optima TL Ultracentrifuge, Beckman Instruments, Fullerton, CA), the supernatant was carefully removed, and the gelatinous pellet was dissolved in SDS-polyacrylamide gel electrophoresis sample buffer. After separation by SDS-polyacrylamide gel electrophoresis, actin content of the Triton-insoluble pellet was analyzed by Western blot with rabbit anti-actin (Sigma) and quantitated by densitometry.

LPL N-terminal Peptide Induces
We previously demonstrated that a cell permeant peptide containing LPL amino acids 2-19 fused at the C terminus with the HIV tat sequence (LPLtat) induces ␣ M ␤ 2 -mediated adhesion in PMN (7). LPLtat also causes ␣ M ␤ 2 -dependent adhesion to C3bi, an ␣ M ␤ 2 -specific ligand, in Jurkat cells transfected with ␣ M ␤ 2 integrin. 2 These data suggest that LPL has a role in ␣ M ␤ 2 activation in leukocytes. We next asked whether LPL was specifically involved in ␣ M ␤ 2 activation or if it was involved in the regulation of other integrins. The K562 cell, transfected with exogenous integrins, is a well-established model system for studying integrin functions in a myeloid background (8,9,14,17), and our laboratory has shown previously that cell activation is required for adhesion mediated by ␣ v ␤ 3 expressed in these cells (8). Therefore, we tested the effects of LPLtat on ␣ v ␤ 3 -mediated adhesion to Vn in stable K562 transfectants (K␣ v ␤ 3 ). LPLtat induced K␣ v ␤ 3 adhesion to Vn but not to HSA (Fig. 1A). Although LPLtat did not induce cell adhesion in the absence of ligand, its effect was apparent even at a Vn coating concentration of 0.1 g/ml (Fig. 1B). This effect was specific to LPLtat, because control peptides, including tatLPL, in which the tat sequence mediating membrane permeability of the peptide was N-terminal to the LPL sequence, SCRtat, in which the LPL sequence was scrambled, and TPLtat, which contained amino acids 2-19 of the LPL homologue T-plastin, did not increase adhesion (Fig. 1C). LPLtat did not induce adhesion of the parental K562 cells, which do not express ␣ v ␤ 3 , to Vn (K562ϩLPLtat in Fig. 1C). LPLtatinduced adhesion of K␣ v ␤ 3 was inhibited by the anti-␣ v antibody L230 and the anti-␤ 3 antibody 1A2.1 (Fig. 1D), further demonstrating the requirement for the ␣ v ␤ 3 integrin. LPLtat, but not tatLPL, SCRtat, or TPLtat, also induced K␣ v ␤ 3 adhesion to osteopontin (data not shown). Like adhesion to Vn, K␣ v ␤ 3 adhesion to osteopontin was inhibited by an anti-␤ 3 mAb but not by antibodies recognizing ␣ 5 , ␤ 1 , or HLA. Thus, LPLtat activates ␣ v ␤ 3 -as well as ␣ M ␤ 2 -mediated adhesion. In contrast, ␣ 5 ␤ 1 -mediated adhesion to fibronectin does not require activation in K562 cells (14), and LPLtat had no significant effect on adhesion of untransfected K562 to fibronectin (data not shown). Thus, LPLtat affects integrin activation for adhesion, rather than adhesion by already active integrins.
To determine whether LPLtat could activate ␣ v ␤ 3 in a cell constitutively expressing this integrin, we examined adhesion of the JY B lymphoma line, which is known to have activable ␣ v ␤ 3 (23,37). These cells also adhered to Vn when treated with LPLtat ( Fig. 2A), in a manner dependent on Vn-coating concentration (Fig. 2B) and ␤ 3 integrin (Fig. 2C).
The LPL Headpiece Induces ␣ v ␤ 3 -mediated Adhesion in K␣ v ␤ 3 Cells-In addition to the site for phosphorylation, the N-terminal domain, or headpiece, of LPL contains two EFhand-type Ca 2ϩ binding motifs (18), and Ca 2ϩ markedly affects the conformation of this domain. To determine whether LPLtat induction of adhesion could be recapitulated with the entire LPL N-terminal domain, we expressed an LPL truncation mutant that contained the entire headpiece domain fused to the tat peptide sequence either at the C terminus (Headtat) or the N terminus (tatHead) (Fig. 3A). The purified proteins were recognized by specific monoclonal antibodies directed toward the tat-(49 -58) epitope (19) or toward the LPL N terminus (Fig.  3B). As in the case of LPL peptides, Headtat, but not tatHead induced K␣ v ␤ 3 adhesion to Vn. Headtat-induced adhesion was completely inhibited by both anti-␣ v and anti-␤ 3 antibodies (Fig. 3C). These data demonstrate that, in addition to short peptides, the intact headpiece domain of LPL can activate LPLtat-induced Adhesion Does Not Require the ␤ 3 Cytoplasmic Tail Tyr-747 or Tyr-759 -PMA-and thrombin-activated K␣ v ␤ 3 adhesion to Vn requires Tyr-747 of the ␤ 3 cytoplasmic tail, because Y747F mutants fail to adhere in response to these stimuli (8). To determine whether LPLtat-induced adhesion was also dependent on Tyr-747, we examined the effect of LPLtat on adhesion to Vn in K562 cells expressing ␣ v ␤ 3 with single amino acid mutations Y747F, Y759F, and the double mutant Y747F/Y759F. As previously observed (8), the Y747F but not the Y759F mutation abolished PMA-stimulated adhesion to Vn (Fig. 4). However, K␣ v ␤ 3 Y747F, K␣ v ␤ 3 Y759F, and K␣ v ␤ 3 Y747F/Y759F adhered equally well as cells expressing wild-type ␣ v ␤ 3 when treated with LPLtat (Fig. 4). Thus, unlike PMA, LPLtat-activated adhesion does not require ␤ 3 Tyr-747 or its phosphorylation. This demonstrates that the LPLtat effect on ␣ v ␤ 3 is either downstream or independent of the Tyr-747dependent signaling cascade (20,21) involved in PMA-induced adhesion.
LPLtat and RGD Act Synergistically to Increase Ligandinduced Binding Site (LIBS) Expression-Leukocyte integrin activation can be achieved by several mechanisms. For many integrins, binding of ligand or exposure to the divalent cation Mn 2ϩ induces a conformational change that results in increased adhesion because of enhanced affinity for ligand (22). In the case of ␣ v ␤ 3 , RGD peptide or Mn 2ϩ binding can induce this conformational change, which can be identified by the binding of monoclonal antibodies (mAb) that recognize LIBS epitope (8,23,24). This conformational change is associated with increased affinity for Vn, apparently due to a decreased off-rate of the receptor-ligand interaction (25). As shown in Fig.  5A, the anti-␤ 3 mAb 7G2 (13) recognizes an LIBS epitope as its binding to ␣ v ␤ 3 increased ϳ10-fold in response to RGD peptide. Maximal 7G2 reactivity was achieved with 25 M RGD peptide (Fig. 5A), and increasing the RGD peptide concentration to as high as 2 mM did not induce further 7G2 reactivity (data not shown). Control RAD peptide did not enhance 7G2 binding (Fig. 5B). Treatment of K␣ v ␤ 3 with LPLtat alone only minimally increased 7G2 binding (Fig. 5, A and B). However, even at concentrations of RGD peptide that caused maximum 7G2 binding, 7G2 reactivity was further increased when cells were treated with RGD in combination with LPLtat, demonstrating marked synergy between the cytoplasmic peptide and the ligand (Fig. 5A). Treatment of K␣ v ␤ 3 with RGD together with control peptides such as LPL (LPL aa 2-19 without associated tat), or other peptides bearing the tat sequence, TPLtat, tatLPL, and SCRtat, did not cause 7G2 reactivity to increase over the level induced by RGD alone (Fig. 5B). Total ␣ v ␤ 3 expression on the surface of the transfected K562 cells was not altered by LPLtat treatment as assessed by the (non-LIBS) anti-␤ 3 antibody AP3 (data not shown). LPLtat synergized with  RGD to induce LIBS expression in ␣ v ␤ 3 with the Y747F and Y759F mutations (Fig. 5C), consistent with the finding that LPLtat-induced adhesion does not require these ␤ 3 cytoplasmic tail tyrosines. LPLtat also synergized with Mn 2ϩ to increase LIBS expression in wild type and mutant receptors ( Fig. 5D and data not shown). In contrast, PMA treatment did not induce 7G2 binding either with or without LPLtat (Fig. 5D). These data indicate that LPLtat induces ␣ v ␤ 3 -mediated adhesion by cooperating with ligand to induce expression of the high affinity conformation of the integrin, independent of Tyr-747 or Tyr-759.

Inhibition of LPLtat-induced Adhesion Requires a Higher Concentration of RGD Peptide than PMA-or Mn 2ϩ -induced
Adhesion-As has been shown repeatedly, RGD peptide competes with Vn for the ligand binding site in ␣ v ␤ 3 and inhibits adhesion to this protein. Because LPLtat cooperated with RGD to induce an LIBS epitope and presumably the high affinity state of the integrin, we asked whether adhesion induced by PMA, Mn 2ϩ , and LPLtat had similar sensitivity to RGD peptide inhibition. If LPLtat increased the number of high affinity receptors, it should shift the ID 50 for RGD to a higher concentration. As this hypothesis predicted, LPLtat-activated adhesion was much more resistant to RGD peptide inhibition than either PMA or Mn 2ϩ stimulation (Fig. 6A). Mn 2ϩ -or PMAinduced adhesion was completely inhibited by 50 M RGD peptides, with an ID 50 at about 20 M (Fig. 6, B and C). LPLtat-induced adhesion, however, required 100-fold more RGD peptide for equivalent inhibition, with an ID 50 of ϳ2.5 mM RGD (Fig. 6A). This result is consistent with the higher LIBS expression induced by RGD and LPLtat than by RGD and Mn 2ϩ or RGD and PMA (Fig. 5D).
LPLtat-induced Adhesion Is Inhibited by Jasplakinolide, but Not Cytochalasin D-It is generally thought that integrins require attachment to actin filaments for firm adhesion. However, adhesion mediated by high affinity integrin receptors can be unaffected by cytochalasin D (26), implying that new actin microfilament formation is not required for this mechanism of attachment. Therefore, the sensitivity of LPLtat-induced adhesion to cytochalasin D was tested. As expected, PMA-induced adhesion, which requires the actin cytoskeleton in post-receptor events, was completely inhibited by cytochalasin D at 1 g/ml. In contrast, cytochalasin D had no inhibitory effect on either LPLtat or Mn 2ϩ -induced ␣ v ␤ 3 -mediated adhesion even at 10 g/ml (Fig. 7A). These results demonstrate that LPLtat induces adhesion mediated by high affinity receptors, independent of actin microfilaments.
It is recognized that actin microfilaments can also negatively regulate integrin activation by restricting integrin lateral mobility in resting cells (10). As a result, low doses of cytochalasin D actually induces LFA-1-and Mac-1-mediated adhesion by releasing integrins from cytoskeletal constrains leading to integrin activation (10,(27)(28)(29). In this study, low concentration of cytochalasin D (0.1 g/ml) activated ␣ v ␤ 3 -mediated adhesion as well (Fig. 7B). Apparently, as in the case of LFA-1 and Mac-1, release of unactivated integrins from cytoskeletal constraint can lead to ␣ v ␤ 3 -mediated adhesion.
Because actin depolymerization can induce ␣ v ␤ 3 -mediated adhesion, the requirement for actin depolymerization in ␣ v ␤ 3mediated adhesion was examined. Jasplakinolide stabilizes pre-existing F-actin, prevents actin depolymerization, and induces a net increase in actin polymerization (30). It was previously shown to inhibit LFA-1-dependent adhesion (12). Similarly, jasplakinolide inhibited ␣ v ␤ 3 -mediated adhesion induced by LPLtat, Mn 2ϩ , or PMA in K␣ v ␤ 3 (Fig. 8A) and in JY cells (data not shown). Taken together, the cytochalasin D and jasplakinolide data suggest that, although actin polymerization is not required for LPLtat-initiated ␣ v ␤ 3 -mediated adhesion, actin depolymerization is. F-actin content of cells treated with LPLtat and jasplakinolide was determined (Fig. 8B). Although LPLtat caused about a 50% decrease in total F-actin, jasplakinolide treatment more than doubled cellular F actin. Jas- plakinolide also prevented the LPLtat-induced decrease in F-actin.
Because both actin depolymerization and increase in receptor affinity are involved in LPLtat-induced adhesion, we asked whether actin depolymerization was required for the generation of high affinity ␣ v ␤ 3 . Surprisingly, jasplakinolide had little effect on synergistic expression of the 7G2 epitope by RGD and LPLtat (Fig. 8C). In addition, cytochalasin D had no effect on the ability of RGD, LPLtat, or their combination to induce the LIBS epitope recognized by 7G2 (data not shown). Thus, although actin depolymerization is required for ␣ v ␤ 3 -mediated adhesion, it is neither necessary nor sufficient for achieving a high affinity conformation of the receptor.

DISCUSSION
The term "integrin activation" is sometimes used to mean the modulation of affinity that can be induced in many integrins by Mn 2ϩ , activating antibodies, ligand, and sometimes other physiologic or pharmacologic stimuli. If used in this way, integrin activation is thought be related to a conformational change in the extracytoplasmic domain of the integrins and can occur in a wide variety of cell types. However, ␤ 2 and ␤ 3 integrins on bone marrow-derived cells, including leukocytes and platelets, are different from integrins on other cell types, because without physiologic or pharmacologic activation, they do not mediate adhesion. Thus, for these integrins, cell activation is an absolute requirement for function. Activation of adhesion likely results from integrin clustering within the membrane as well as changes in affinity of individual integrins (10 -12). Moreover, leukocyte integrins can be distinguished from platelet integrin ␣ IIb ␤ 3 because, although ␣ IIb ␤ 3 activation for ligand binding is irreversible, activation of leukocyte integrins is a reversible process. Indeed, reversible activation of integrinmediated adhesion is thought to be important in a variety of critical leukocyte functions, including migration, phagocytosis, and cell-mediated cytotoxicity.
Despite the importance of this reversible activation of leukocyte integrin-mediated adhesion, the mechanisms involved in regulation of leukocyte integrin function remain obscure. Although a number of signaling molecules have been identified that can be involved in the signal transduction pathways of inside-out signaling, the effector mechanisms leading to changes in integrin function have not been elucidated. In general, two not mutually exclusive models have been invoked, one involving receptor rearrangement on the plasma membrane and another involving conformational change in the integrin itself (31,32). Our previous work implicates LPL phosphorylation in a final common pathway influencing integrin function after cell activation with G-protein coupled, tyrosine kinasemediated, and PKC-dependent stimuli (7,(33)(34)(35). The fact that peptides that mimic the phosphorylation site in the N terminus of LPL rapidly stimulate integrin activation when introduced into PMN or monocyte cytoplasm (7) presents an excellent opportunity to gain insight into the effector mechanisms of inside-out signaling. Therefore, in this work we have investigated the mechanism by which the LPL N terminus activates integrin-mediated adhesion, using the well-characterized model of ␣ v ␤ 3 -mediated adhesion in K562 cells.
A cell-permeant version of the LPL N-terminal domain (headpiece) activated ␣ v ␤ 3 -mediated adhesion as well as the 18-amino acid LPL peptides previously shown to activate ␣ M ␤ 2mediated adhesion. As previously discussed (7), integrin activation by the LPL N-terminal peptides requires their entry into the cytoplasm, because peptides of identical sequence without the tat addition have no effect on integrin function. Placing the tat sequence at the N-terminal end of either the peptide or the headpiece domain abrogates function even though this placement does not decrease entry into the cytoplasm. 3 This suggests the possibility that a free LPL N terminus is essential for its integrin regulatory function and that addition of the 12amino acid tat peptide blocks some essential function of this region of the protein. However, this domain binds neither actin nor vimentin (36), and there are no known cytoplasmic proteins that interact with the LPL headpiece. Although the headpiece does bind Ca 2ϩ through its tandem EF-hand domains, the active peptides do not, and modulation of intracytoplasmic Ca 2ϩ by itself is not known to affect integrin function. Thus, the molecular interactions by which the LPL N terminus activates integrin avidity are unknown. Nonetheless, both the headpiece domain and the N-terminal peptides induce ␣ v ␤ 3 -mediated adhesion, and the peptides clearly synergize with RGD ligand to activate a conformational change in ␣ v ␤ 3 associated with increased ligand affinity. We hypothesize that this occurs as the result of an interaction between the N-terminal domain of LPL and an as yet unknown target, which could be the integrin itself.
The requirements in the ␤ 3 cytoplasmic domain for LPLtat induction of adhesion are distinct from PMA or thrombin, because the ␤ 3 Y747F mutant, which fails to support adhesion stimulated by PMA, supports normal adhesion induced by the LPL peptides. Because the Y747F mutant also supports adhe-3 S. L. Jones and E. J. Brown, unpublished data. sion in cells with constitutively active integrins (8), it is clear that phosphorylation of this tyrosine is not required for ␣ v ␤ 3mediated adhesion. Rather, Tyr-747 phosphorylation is likely required in a signal transduction pathway involving the integrin, perhaps by recruiting an SH2-or PTB-containing protein, that leads to a final common pathway of integrin adhesion that does not itself require Tyr-747. If this is the case, then LPL functions downstream of Tyr-747 phosphorylation.
LPLtat acts synergistically with RGD ligand to induce the expression of an LIBS epitope on ␤ 3 recognized by mAb 7G2. The RGD peptide can induce 7G2 binding in a dose-dependent manner, as is expected for a LIBS mAb; in contrast, at optimal concentration, LPLtat induces little 7G2 binding. Thus, LPLtat does not act like an activating antibody or certain ␣ IIb ␤ 3 chimeras, for which LIBS expression becomes independent of ligand. Instead, LPLtat appears to augment the conformational change induced by RGD. In K562, as in other cells, only a minority of integrins expresses the LIBS epitope in the presence of ligand. Virtually nothing is known about what distinguishes receptors undergoing conformational change from those that do not. LPLtat seems to increase the fraction of receptors capable of achieving this conformational change when exposed to ligand. In studying purified ␣ v ␤ 3 , Orlando and Cheresh (25) noted that prior exposure to RGD markedly decreased the off-rate of its binding to Vn and suggested that this was the mechanism by which the ligand-induced conformational change in the integrin induced stable cell adhesion. It is likely that, in the context of an intact cell, LPLtat exaggerates this normal response, leading to marked strengthening of adhesion. It is noteworthy that LPLtat-mediated adhesion is insensitive to cytoskeleton disruption with cytochalasin D, suggesting that the LPL effect on integrin interaction with ligand dominates any potential indirect effect of adhesion strengthening through modulation of actin-integrin interactions.
Although the change in integrin conformation is essential for LPLtat-induced adhesion, it is not sufficient, because jasplakinolide blocks adhesion without significant inhibition of the conformational change. It is now clear that integrin release from cytoskeletal constraint is an early and essential aspect of activation of adhesion (10 -12). We hypothesize that, in addition to recruiting new ␣ v ␤ 3 to respond to a ligand-induced conformational change, LPLtat induces integrin release from cytoskeletal constraint. Consistent with this, treatment of cells with LPLtat diminishes total F-actin content and simultaneous treatment with jasplakinolide inhibits this effect of the peptide. Effects on both cytoskeleton and integrin conformation are required for stable adhesion induced by LPLtat, even though cytoskeletal release is not required for the peptide effects on integrin conformation. Moreover, cytoskeletal release occurs in the absence of ligand (10), so conformational change is not a prerequisite for release. In platelets, alterations in integrincytoskeleton interaction are important for inside-out signaling (38), and an increase in intracytoplasmic Ca 2ϩ can activate ␣ IIb ␤ 3 by releasing it from cytoskeletal constraints (39). Release of integrins from cytoskeletal constraint may require phosphorylation of the myristoylated alanine-rich protein kinase C substrate (40) as well as LPL. Thus, LPL peptide appears to induce two necessary but potentially independent events in control of integrin function. It is intriguing to speculate that the peptide may interact directly with integrin cytoplasmic tails to achieve these effects. Because phosphorylation of LPL may activate its release from actin filaments, 2 it is possible that signal transduction leading to LPL phosphoryla-tion frees the N terminus to interact with integrins, leading in turn to their release from cytoskeleton and diffusion to sites of ligand concentration, where a conformational change leading to high affinity interaction occurs upon ligand binding.