Integrin associated proteins.

Integrin cytoplasmic domains may interact directly with serveral different cytoskeletal proteins and intracellular signaling molecules. Also, integrins interact directly with other transmembrane structures, including transmembrane-4 superfamily (TM4SF) proteins. New evidence suggests that TM4SF proteins may act as linkers between extracellular integrin alpha chain domains and intracellular signaling molecules, such as phosphatidylinositol 4-kinase and protein kinase C.

Interchangeable alpha chain cytoplasmic domains play a positive role in control of cell adhesion mediated by VLA-4, a beta 1 integrin.

Integrins can exist in a range of functional states, depending on the cell type and its state of activation. Although the mechanism that controls activity is unknown, it has been suggested that for some integrins, alpha chain cytoplasmic domains may exert either a negative effect or no effect on adhesion function. To address this issue for VLA-4 (an alpha 4 beta 1 heterodimer), we constructed an alpha 4 cytoplasmic deletion mutant and chimeric alpha chains composed of the extracellular domains of alpha 4 and the cytoplasmic domains of alpha 2, alpha 4, or alpha 5. Upon stable transfection of wild-type alpha 4, VLA-4 heterodimer was obtained that mediated (a) poor adhesion to CS1 peptide, fibronectin, or vascular cell adhesion molecule 1 (VCAM-1) (in K562 cells); (b) poor adhesion to CS1 peptide but moderate adhesion to VCAM-1 (in MIP101 cells); and (c) moderate adhesion to both CS1 peptide and VCAM-1 (in PMWK cells). Chimeric alpha 4 constructs and wild-type alpha 4 yielded similar results in these cell lines. In contrast, truncation of the alpha 4 cytoplasmic domain (after the conserved GFFKR motif) caused an almost complete loss of adhesive activity in all three cell lines. Thus, several interchangeable alpha chain cytoplasmic domains play a fundamentally positive role in determining the state of constitutive activity for VLA-4. The alpha chain cytoplasmic domain is also required for agonist-stimulated adhesion, since phorbol ester stimulated the cell adhesion mediated by wild-type and chimeric alpha chains, but not by the cytoplasmic deletion mutant. The inactivity of both wild-type VLA-4 (in K562 cells), and truncated VLA-4 (in all three cell lines) was overcome by the addition of a stimulatory anti-beta 1 monoclonal antibody. Thus, the alpha cytoplasmic domain-dependent cellular mechanism controlling both constitutive and agonist-stimulated VLA-4 activity could be bypassed by external manipulation of the integrin.

Genetic and biochemical characterization of human lymphocyte cell surface antigens. The A-1A5 and A-3A4 determinants.

The genes that code for the human lymphocyte cell surface determinants defined by monoclonal antibodies A- 1A5 and A- 3A4 have been genetically mapped. All human chromosomes, except Y, were included in a series of human less than mouse lymphocyte hybrid populations that retained expression of lymphocyte-specific surface markers. Expression of the A- 1A5 and A- 3A4 antigens was quantitated by indirect immunofluorescence and fluorescence-activated cell sorter (FACS) analysis. Hybrid populations heterogeneous for antigen expression were sorted to yield antigenically homogeneous subpopulations. Isozyme analysis indicated concordant segregation of the A- 1A5 determinant with chromosome 10, and the A- 3A4 determinant with chromosome 4. In contrast to the unhybridized human parent cell line (MOLT-4), from which A- 1A5 immunoprecipitated two proteins (160,000 and 125,000 Mr), A- 1A5 only immunoprecipitated a single band (125,000 Mr) from an A- 1A5 -expressing human less than mouse hybrid. The genetic disassociation of these two proteins from the A- 1A5 -reactive complex suggests that the appearance of the 160,000 Mr protein requires a gene locus that is unlinked to the locus for the 125,000 Mr protein on chromosome 10. A third component of the A- 1A5 -reactive protein complex (210,000 Mr), which is recognized by the monoclonal antibody TS2/7, was not expressed on the parent MOLT-4 cells, but was weakly expressed on MOLT-4 less than mouse BW5147 hybrids. This allowed preliminary mapping of that determinant to either chromosome 10 or 15. The A- 3A4 antigen (approximately 45,000 Mr) is a novel cell surface structure expressed on all hematopoietic cell lines tested, and represents the first cell surface marker mapped to chromosome 4.

Mutational evidence for control of cell adhesion through integrin diffusion/clustering, independent of ligand binding.

Previous studies have shown that integrin alpha chain tails make strong positive contributions to integrin-mediated cell adhesion. We now show here that integrin alpha4 tail deletion markedly impairs static cell adhesion by a mechanism that does not involve altered binding of soluble vascular cell adhesion molecule 1 ligand. Instead, truncation of the alpha4 cytoplasmic domain caused a severe deficiency in integrin accumulation into cell surface clusters, as induced by ligand and/ or antibodies. Furthermore, alpha4 tail deletion also significantly decreased the membrane diffusivity of alpha4beta1, as determined by a single particle tracking technique. Notably, low doses of cytochalasin D partially restored the deficiency in cell adhesion seen upon alpha4 tail deletion. Together, these results suggest that alpha4 tail deletion exposes the beta1 cytoplasmic domain, leading to cytoskeletal associations that apparently restrict integrin lateral diffusion and accumulation into clusters, thus causing reduced static cell adhesion. Our demonstration of integrin adhesive activity regulated through receptor diffusion/clustering (rather than through altered ligand binding affinity) may be highly relevant towards the understanding of inside-out signaling mechanisms for beta1 integrins.

Retooling of the beta 4 integrin in tumor cells--ligands lost and kinase gained.

In carcinoma cells, the beta 4 integrin functions in a ligand-independent manner to promote proliferation, migration, and invasion. An interesting new paper describes a mechanism whereby the beta 4 integrin cytoplasmic tail becomes an integrin ligand-independent adaptor protein for the Met receptor tyrosine kinase, thereby enhancing the mitogenic, morphogenic, and motogenic properties of Met.

Specific tetraspanin functions.

Relatively little attention has been given to the large family of abundantly expressed transmembrane proteins known as tetraspanins. Now, the importance of tetraspanins is strongly supported by emerging genetic evidence, coupled with new insights into the biochemistry and functions of tetraspanin protein complexes.

Mutation of putative divalent cation sites in the alpha 4 subunit of the integrin VLA-4: distinct effects on adhesion to CS1/fibronectin, VCAM-1, and invasin.

To investigate the functional significance of putative integrin divalent cation binding sites, several mutated alpha 4 subunit cDNAs were constructed. Mutants contained the conservative substitution of Glu for Asp or Asn at the third position in each of three putative divalent cation sites. Transfection of wild-type or mutated alpha 4 into K562 cells yielded comparable expression levels and immunoprecipitation profiles. However, for all three alpha 4 mutants, adhesion to CS1/fibronectin was greatly diminished in either the presence or absence of the stimulatory anti-beta 1 mAb TS2/16. Constitutive adhesion to vascular cell adhesion molecule (VCAM) 1 was also diminished but, unlike CS1 adhesion, was restored upon TS2/16 stimulation. In contrast, adhesion to the bacterial protein invasin was minimally affected by any of the three mutations. For each of the mutants, the order of preference for divalent cations was unchanged compared to wild-type alpha 4, on CS1/fibronectin (Mn2+ > Mg2+ > Ca2+), on VCAM-1 (Mn2+ > Mg2+ = Ca2+) and on invasin (Mg2+ = Ca2+). However for the three mutants, the efficiency of divalent cation utilization was decreased. On VCAM-1, 68-108 microM Mn2+ was required to support half-maximal adhesion for the mutants compared with 14-18 microM for wild-type alpha 4. These results indicate (a) that three different ligands for VLA-4 show widely differing sensitivities to mutations within putative divalent cation sites, and (b) each of the three putative divalent cation sites in alpha 4 have comparable functional importance with respect to both divalent cation usage and cell adhesion.

Contrasting roles for integrin beta 1 and beta 5 cytoplasmic domains in subcellular localization, cell proliferation, and cell migration.

To carry out a detailed comparison of the roles of integrin beta 1 and beta 5 cytoplasmic domains, we expressed both wild type beta 1 and chimeric beta 1/5 constructs in CHO cells. In the latter, the cytoplasmic domain of beta 1 was replaced with that of beta 5. The human beta 1 and beta 1/5 constructs appeared at similar levels at the cell surface (mostly as alpha 5 beta 1 heterodimers) and contributed equally to CHO cell adhesion to fibronectin. However, beta 1 but not beta 1/5 localized to focal adhesion-like structures when CHO cells were spread on fibronectin. Furthermore, only the beta 1-CHO cells showed increased proliferation in response to fibronectin plus an integrin-activating anti-beta 1 antibody, and showed increased appearance of 32P-labeled protein (p90) that correlated with proliferation. In sharp contrast, the beta 1/5-CHO cells were notably more migratory than beta 1-CHO cells in a transwell haptotactic migration assay. These results indicate that the beta 1 and beta 5 integrin subunit cytoplasmic domains can translate similar adhesive information into highly contrasting subsequent events. Thus, we have established that "inside-out" and "outside-in" integrin signaling pathways are regulated by fundamentally distinct mechanisms. In addition, we suggest that the same properties of the beta 1 cytoplasmic domain that promote recruitment to visible focal adhesion-like structures may also be conductive to cell proliferation. Conversely, the properties of the beta 5 tail that make it less likely to localize into focal adhesion-like structures may contribute to enhanced cell migration.

The primary structure of the VLA-2/collagen receptor alpha 2 subunit (platelet GPIa): homology to other integrins and the presence of a possible collagen-binding domain.

VLA-2 (also called gpIa/IIa on platelets) is a collagen receptor with a unique alpha subunit and a beta subunit common to other adhesion receptors in the VLA/integrin family. Multiple cDNA clones for the human VLA-2 alpha 2 subunit have been selected from a lambda gtll library by specific antibody screening. The 5,374-bp nucleotide sequence encoded for 1,181 amino acids, including a signal peptide of 29 amino acids followed by a long extracellular domain (1,103 amino acids), a transmembrane domain, and a short cytoplasmic segment (22 amino acids). Direct sequencing of purified alpha 2 protein confirmed the identity of the 15 NH2-terminal amino acids. Overall, the alpha 2 amino acid sequence was 18-25% similar to the sequences known for other integrin alpha subunits. In particular, the alpha 2 sequence matched other integrin alpha chains in (a) the positions of 17 of its 20 cysteine residues; (b) the presence of three metal-binding domains of the general structure DXDXDGXXD; and (c) the transmembrane domain sequence. In addition, the alpha 2 sequence has a 191-amino acid insert (called the I-domain), previously found only in leukocyte integrins of the beta 2 integrin family. The alpha 2 I-domain was 23-41% similar to domains in cartilage matrix protein and von Willebrand factor, which are perhaps associated with collagen binding. The NH2-terminal sequence reported here for alpha 2 does not match the previously reported alpha 2 NH2-terminal sequence (Takada, Y., J. L. Strominger, and M. E. Hemler. 1987. Proc. Natl. Acad. Sci. USA. 84:3239-3243). Resolution of this discrepancy suggests that there may be another VLA heterodimer that resembles VLA-2 in size but has a different amino acid sequence.

Role of transmembrane 4 superfamily (TM4SF) proteins CD9 and CD81 in muscle cell fusion and myotube maintenance.

The role of transmembrane 4 superfamily (TM4SF) proteins during muscle cell fusion has not been investigated previously. Here we show that the appearance of TM4SF protein, CD9, and the formation of CD9-beta1 integrin complexes were both regulated in coordination with murine C2C12 myoblast cell differentiation. Also, anti-CD9 and anti-CD81 monoclonal antibodies substantially inhibited and delayed conversion of C2C12 cells to elongated myotubes, without affecting muscle-specific protein expression. Studies of the human myoblast-derived RD sarcoma cell line further demonstrated that TM4SF proteins have a role during muscle cell fusion. Ectopic expression of CD9 caused a four- to eightfold increase in RD cell syncytia formation, whereas anti-CD9 and anti-CD81 antibodies markedly delayed RD syncytia formation. Finally, anti-CD9 and anti-CD81 monoclonal antibodies triggered apoptotic degeneration of C2C12 cell myotubes after they were formed. In summary, TM4SF proteins such as CD9 and CD81 appear to promote muscle cell fusion and support myotube maintenance.

Multiple functional forms of the integrin VLA-2 can be derived from a single alpha 2 cDNA clone: interconversion of forms induced by an anti-beta 1 antibody.

The integrin VLA-2 was previously found to bind to either collagen alone, or collagen plus laminin, but the mechanism for this cell-specific functional difference was unknown. Here we transfected VLA-2 alpha 2 subunit cDNA into K562 cells and obtained VLA-2 (called Form-O) which bound to neither collagen nor laminin. We then used a Matrigel selection procedure to enrich for a minor subpopulation of K562 cells stably expressing a form of VLA-2 (Form-C) that bound to collagen but not laminin. In contrast, the same alpha 2 cDNA transfected into RD cells yielded VLA-2 (Form-CL) which bound to both collagen and laminin. These Form-O, -C, and -CL activities were stably expressed during extended cell culture, and could not be qualitatively altered by adding phorbol esters or by exchaning the resident divalent cations. However, addition of stimulatory anti-beta 1 antibodies (TS2/16, A-1A5) rapidly converted VLA-2 Form-O and Form-C into Form-CL. Anti-beta 1 antibody stimulation of VLA-2 activity was observed not only on whole cells, but also with solubilized receptors. These results suggest (a) that the ligand binding specificity of VLA-2 can be determined by its cellular environment, rather than by variations in the primary sequence of the alpha 2 subunit, (b) that stably inactive or partly active VLA-2 can be rapidly converted to a fully active form through conformational changes initiated at a nonligand binding site on the beta 1 subunit, and (c) that the mechanisms for VLA-2 stimulation by phorbol ester and by antibody are quite distinct, because the latter does not require an intact cell.

Receptor functions for the integrin VLA-3: fibronectin, collagen, and laminin binding are differentially influenced by Arg-Gly-Asp peptide and by divalent cations.

The capability of the integrin VLA-3 to function as a receptor for collagen (Coll), laminin (Lm), and fibronectin (Fn) was addressed using both whole cell adhesion assays and ligand affinity columns. Analysis of VLA-3-mediated cell adhesion was facilitated by the use of a small cell lung carcinoma line (NCI-H69), which expresses VLA-3 but few other integrins. While VLA-3 interaction with Fn was often low or undetectable in cells having both VLA-3 and VLA-5, NCI-H69 cells readily attached to Fn in a VLA-3-dependent manner. Both Arg-Gly-Asp (RGD) peptide inhibition studies, and Fn fragment affinity columns suggested that VLA-3, like VLA-5, may bind to the RGD site in human Fn. However, unlike Fn, both Coll and Lm supported VLA-3-mediated adhesion that was not inhibited by RGD peptide, and was totally unaffected by the presence of VLA-5. In addition, VLA-3-mediated binding to Fn was low in the presence of Ca++, but was increased 6.6-fold with Mg++, and 30-fold in the presence of Mn++. In contrast, binding to Coll was increased only 1.2-fold with Mg++, and 1.7-fold in Mn++, as compared to the level seen with Ca++. Together, these experiments indicate that VLA-3 can bind Coll, Lm, and Fn, and also show that (a) VLA-3 can recognize both RGD-dependent and RGD-independent ligands, and (b) different VLA-3 ligands have distinctly dissimilar divalent cation sensitivities.

Molecular cloning and expression of the cDNA for alpha 3 subunit of human alpha 3 beta 1 (VLA-3), an integrin receptor for fibronectin, laminin, and collagen.

alpha 3 beta 1 (VLA-3), a member of the integrin family of cell adhesion receptors, may function as a receptor for fibronectin, laminin, and collagen. A partial cDNA clone (2.4 kb) for the human alpha 3 subunit was selected from an endothelial cell lambda gt11 cDNA library by specific antibody screening. Several overlapping cDNA clones were subsequently obtained, of a total length of 4.6 kb from various cDNA libraries. The reconstructed alpha 3 cDNA was expressed on the surface of chinese hamster ovary cells as detected by an alpha 3-specific mAb after transfection, suggesting that the cDNA is authentic. Within this sequence was an open reading frame, encoding for 1,051 amino acids, including a signal peptide of 32 residues, a long extracellular domain (959 residues), a transmembrane domain (28 residues), and a short cytoplasmic segment (32 residues). Overall, the alpha 3 amino acid sequence was 25-37% similar to the other integrin alpha subunits that are cleaved, with most similarity to the alpha 6 sequence (37%), and less similarity to those alpha subunits that have I domains (15-20%, excluding the I domain sequence itself). Features most like those in other alpha subunits are (a) the positions of 18/19 cysteine residues, (b) three potential metal binding domains of the general structure DX(D/N)X(D/N)GXXD, and (c) the predicted transmembrane domain. The mass of alpha 3 calculated from its amino acid sequence is 113,505. The human alpha 3 sequence was 89% identical to hamster galactoprotein b3, and 70% similar to the chicken CSAT antigen band 2 protein partial sequence, suggesting that these two polypeptides are homologues of human alpha 3.

Cloning and expression of cDNAs for the alpha subunit of the murine lymphocyte-Peyer's patch adhesion molecule.

cDNA clones encoding the alpha chain of the murine lymphocyte-Peyer's patch adhesion molecule (LPAM), which is associated with lymphocyte homing, have been isolated by screening with the human VLA-4 (alpha 4h) probe. Several alpha 4 antigenic determinants were identified on COS-7 cells after transfection. From overlapping clones, approximately 5 kb of contiguous nucleotide sequence have been determined, encoding a protein sequence of 1039 amino acids for the LPAM alpha chain (alpha 4m). LPAM is a member of the integrin family of cell-surface heterodimers, and alpha 4m is the murine homologue of the human alpha 4 h chain. The two proteins have a total sequence similarity of 84%, with an almost perfect conservation (31/32 amino acids) in the cytoplasmic domain. Like alpha 4h, alpha 4m is distinct from other integrin alpha chains because it has neither an I-domain nor a COOH-terminal cleavage site. The positions of the characteristic Cysteine residues are conserved, and a putative protease cleavage site is located near the middle of the protein sequence. The NH2-terminal part of the protein contains seven homologous repeats, and three of them include putative divalent cation-binding sites. These sites are among the most conserved between the alpha 4m sequence and other alpha chains, and may therefore be involved in the binding of integrin alpha and beta chains. An additional cDNA clone was isolated which shares a sequence of perfect homology with the alpha 4m encoding cDNAs, but has a unique 3' poly-A end. This observation correlates with the fact that three discrete murine RNA bands are observed in Northern blot experiments using alpha 4m as a probe, whereas only two human RNA species are described for alpha 4h, indicating a higher complexity for murine than for human sequences.

The integrin VLA-4 supports tethering and rolling in flow on VCAM-1.

Selectins have previously been shown to tether a flowing leukocyte to a vessel wall and mediate rolling. Here, we report that an intergrin, VLA-4, can also support tethering and rolling. Blood T lymphocytes and alpha 4 integrin-transfected cells can tether in shear flow, and then roll, through binding of the intergrin VLA-4 to purified VCAM-1 on the wall of a flow chamber. VLA-4 transfectants showed similar tethering and rolling on TNF-stimulated endothelium. Tethering efficiency, rolling velocity, and resistance to detachment are related to VCAM-1 density. Tethering and rolling did not occur on ICAM-1, fibronectin, or fibronectin fragments, and tethering did not require integrin activation or the presence of an alpha 4 cytoplasmic domain. Arrest of rolling cells on VCAM-1 occurred spontaneously, and/or was triggered by integrin activating agents Mn2+, phorbol ester, and mAb TS2/16. These agents, and the alpha 4 cytoplasmic domain, promoted increased resistance to detachment. Together the results show that VLA-4 is a versatile integrin that can mediate tethering, rolling, and firm arrest on VCAM-1.

In vitro and in vivo consequences of VLA-2 expression on rhabdomyosarcoma cells.

Cloned integrin alpha 2 subunit complementary DNA was expressed on human rhabdomyosarcoma (RD) cells to give a functional VLA-2 (alpha 2 beta 1) adhesion receptor. The VLA-2-positive RDA2 cells not only showed increased adhesion to collagen and laminin in vitro, but also formed substantially more metastatic tumor colonies in nude mice after either intravenous or subcutaneous injection. These results show that a specific adhesion receptor (VLA-2) can markedly enhance both experimental and spontaneous metastasis. In contrast to the metastasis results, there was no difference in either the in vitro growth rate or apparent in vivo tumorigenicity of RD and RDA2 cells.

Identification of the integrin VLA-2 as a receptor for echovirus 1.

Cell surface receptors for echovirus, a common human pathogen, were identified with monoclonal antibodies that protected susceptible cells from infection with echovirus 1. These monoclonal antibodies, which prevented virus attachment to specific receptor sites, recognized the alpha and beta subunits of the integrin VLA-2 (alpha 2 beta 1), a receptor for collagen and laminin. RD rhabdomyosarcoma cells expressed little VLA-2, did not bind to 35S-labeled virus, and resisted infection until transfected with complementary DNA encoding the alpha 2 subunit of VLA-2. Thus, integrins, adhesion receptors important in interactions between cells and with the extracellular matrix, can mediate virus attachment and infection.

Are changes in integrin affinity and conformation overemphasized?

The activation of integrin-type adhesion receptors might result in the increased affinity of the receptor for ligand. In addition, the activated receptor might display new epitopes, which are increasingly monitored in clinical settings. Here, we highlight examples of integrin 'activation' that is not accompanied by enhanced ligand binding. Also, we emphasize that the dominant integrin conformational changes occur not with 'activation', but after integrins have already bound ligand.

Integrin-ligand binding. Do integrins use a 'MIDAS touch' to grasp an Asp?

Integrins use divalent cations, held within a novel 'MIDAS' motif, for ligand binding. It is proposed that a missing coordination site for the cation is provided by a critical acidic residue in the ligand.

The integrin VLA-2 binds echovirus 1 and extracellular matrix ligands by different mechanisms.

The integrin VLA-2 mediates cell adhesion to collagen and laminin and also functions as a virus receptor, mediating cell surface attachment and infection by a human pathogen, echovirus 1. To determine whether extracellular matrix proteins and virus interact with VLA-2 in the same manner, we carried out a detailed comparison of these two functions and found that they differed markedly in six different respects. In contrast to the ECM/VLA-2 interaction, echovirus 1 binding did not discriminate between functional forms of VLA-2, showed a different pattern of inhibition by anti-beta1 and -alpha 2 antibodies, was not stimulated by phorbol esters, was not activated by beta 1 antibodies that stimulate ECM binding, was not inhibited by any particular divalent cation, and most notably was not inhibited by EDTA. These striking differences were found both with intact cells expressing VLA-2 and with solubilized VLA-2, suggesting that VLA-2 interacts with these different ligands by markedly different mechanisms, and probably at different functional sites. In addition, alterations in the alpha 2 cytoplasmic domain that had marked effects on cellular responses to collagen and laminin had no effect on virus internalization and cell killing. Thus VLA-2-mediated events that occur after receptor occupancy by extracellular matrix proteins also appear to be distinct from those that occur after receptor interaction with virus.