Functional comparison of the alpha3A and alpha3B cytoplasmic domain variants of the chicken alpha3 integrin subunit.

Integrin alpha3beta1 can be alternatively spliced to generate alpha3A and alpha3B cytoplasmic domain variants that are conserved among vertebrates. To identify distinct functions of these variants, we transfected cells with intact alpha3 integrins or chimeric receptors. alpha3Abeta1 and alpha3Bbeta1 each localized to focal contacts in keratinocytes on an extracellular matrix rich in laminin-5, to which both are known to bind with high affinity. However, alpha3B accumulated intracellularly in keratinocytes on collagen, suggesting that laminin binding may stabilize alpha3Bbeta1 surface expression. Neither alpha3 cytoplasmic domain affected recruitment of chimeric alpha5 integrins to fibronectin-induced focal contacts, and either substituted for the alpha5 cytoplasmic domain in alpha5beta1-mediated cell migration. However, the alpha5/alpha3B chimera localized to cell-cell borders in MDCK or CHO cells to a lesser extent than did the alpha5/alpha3A chimera. To determine whether the alpha3 cytoplasmic domains conferred distinct localization to a nonintegrin protein, we transfected cells with interleukin-2 receptor (IL-2R) chimeras containing the alpha3 cytoplasmic domains. The IL-2R/alpha3A chimera was expressed efficiently on the cell surface, while the IL-2R/alpha3B chimera accumulated intracellularly. Our findings suggest that the alpha3B cytoplasmic domain harbors a retention signal that is regulated in an intact integrin and can alter cell surface expression and distribution of alpha3beta1.

Overlapping and independent functions of fibronectin receptor integrins in early mesodermal development.

Mouse embryos deficient in fibronectin (FN-null) die at E8.5 with mesodermal defects. Eight integrin heterodimers alpha3beta1, alpha4beta1, alpha5beta1, alpha8beta1, alphavbeta1, alphavbeta3, alphavbeta6, and alphaIIbbeta3 can bind to FN. However, embryos deficient in each of these integrins exhibit less severe defects than do FN-null embryos, raising questions as to which integrin(s) are the key FN receptors for these early FN-dependent processes. alpha5beta1 is believed to be the key receptor and alpha5-null embryos display mesodermal defects similar to, although less severe than, those of FN-null. Here we report that the alpha5-null mutation exhibits a more severe phenotype on a 129Sv (129) than on a C57BL/6 (B6) background, as does the FN-null mutation. While alpha5-null/B6 embryos develop normal headfolds, alpha5-null/129 embryos have headfold defects similar to those of FN-null. The differences between FN-null and alpha5-null embryos, however, cannot be attributed to genetic background. FN-null embryos never form somites, whereas in alpha5-null/129 embryos the somites do condense but fail to epithelialize. Second, we examined double mutants carrying all possible pairwise combinations of null mutations in alpha3, alpha4, and alpha5 integrin genes. There was no evidence for any synergy between paired mutations, suggesting that these integrin genes do not have overlapping functions during early embryonic development. Finally, we examined double-mutant embryos deficient in both alpha5 and alphav integrin genes. These double-mutant embryos have an amniotic defect similar to that of FN-null embryos, but die even earlier with a defect in gastrulation. These studies thus revealed a gradation in the severity of defects in the mutations alpha5(-/-); alphav(-/-) > FN(-/-) (129) > FN(-/-) (B6) > alpha5(-/-) (129) > alpha5(-/-) (B6), and in each step in this series there is a certain degree of phenotypic overlap, suggesting that the defects arising from these mutations may result from disruptions of the same embryonic process.

Talin contains three actin-binding sites each of which is adjacent to a vinculin-binding site.

We have determined the sequence of chicken talin (2,541 amino acids, M(r) 271,881) which is very similar (89% identity) to that of the mouse protein. Alignments with the Caenorhabditis elegans and Dictyostelium discoideum talin sequences show that the N- and C-terminal regions of the protein are conserved whereas the central part of the molecule is more divergent. By expressing overlapping talin polypeptides as fusion proteins, we have identified at least three regions of the protein which can bind F-actin: residues 102-497, 951-1,327 and 2,269-2,541. The N-terminal binding site contains a region with homology to the ERM family of actin-binding proteins, and the C-terminal site is homologous to the yeast actin-binding protein Sla2p. Each of the actin-binding sites is close to, but distinct from a binding site for vinculin, a protein which also binds actin. The Pro1176 to Thr substitution found in talin from Wistar-Furth rats does not destroy the capacity of this region of the protein to bind actin or vinculin. Microinjection studies showed that a fusion protein containing the N-terminal actin-binding site localised weakly to stress fibres, whereas one containing the C-terminal site initially localised predominantly to focal adhesions. The former was readily solubilised, and the latter was resistant to Triton extraction. The N-terminal talin polypeptide eventually disrupted actin stress fibres whereas the C-terminal polypeptide was without effect. However, a larger C-terminal fusion protein also containing a vinculin-binding site did disrupt stress fibres and focal adhesions. The results suggest that, although both the N- and C-terminal regions of talin bind actin, the properties of these two regions of the protein are distinct.

Expression of the alternatively spliced EIIIB segment of fibronectin.

Previous descriptions of the expression and distribution of the alternatively spliced EIIIB segment of fibronectin (FN) relied upon an antibody which, on subsequent testing, was shown not to recognize this segment directly. This raises concerns regarding the reliability of all such previous descriptions. In order to prepare reagents directly reactive with this segment, we raised polyclonal antibodies to two different bacterial fusion proteins containing intact EIIIB segments, and to a synthetic 36 amino acid peptide from the center of the EIIIB segment. Antibodies raised to each of these three immunogens recognized fusion proteins containing the EIIIB segment, but failed to recognize full length EIIIB+ FNs produced by mammalian cells, suggesting that oligosaccharide linked to Asn1359 within the EIIIB segment, or potentially to other residues in FN, might interfere with antibody recognition of this segment. Consistent with this hypothesis, N-deglycosylation of recombinant full and partial length EIIIB+ FNs permitted their specific recognition by the anti-fusion protein (but not anti-peptide) antibodies. Using anti-fusion protein antibodies coupled with deglycosylation procedures, we provide a series of new results relevant to the functions of the EIIIB segment: 1) Endogenously synthesized EIIIB+ FN is incorporated into the extracellular matrix of cultured fibroblasts, where it appears by immunofluorescence microscopy and radioimmunoprecipitation analyses to have a distribution very similar to both EIIIA+ forms and the total pool of FNs. 2) No reproducible changes can be shown to occur in the extent of synthesis or matrix incorporation of EIIIB+ FNs upon cellular transformation. 3) Low levels of EIIIB+ FN are normally present in blood plasma and consequently also in purified preparations of plasma FN. 4) EIIIB+ FN is present in blood platelets, where it constitutes a minor fraction of total platelet FN, yet is greater than 4-fold enriched relative to plasma FN. 5) EIIIB+ FN is synthesized by first passage cultured endothelial cells, suggesting that the endothelium could constitute a source for this FN isoform in the circulating blood. The antibodies and methods used in this study constitute the first direct assays of EIIIB+ FN protein expression and are applicable to a variety of species.

Fibronectin/integrin interaction induces tyrosine phosphorylation of a 120-kDa protein.

We describe a 120-kDa protein (pp120) that is phosphorylated on tyrosine in cells attached to fibronectin-coated surfaces. The protein appears to be located in focal contacts where it codistributes with beta 1 integrins. pp120 is distinct from the beta 1 subunit of integrins and from vinculin and alpha-actinin. pp120 is rapidly dephosphorylated in cells suspended by trypsinization but becomes rapidly phosphorylated in cells attaching and spreading on fibronectin. Attachment of cells to RGD-containing peptides, polylysine, or concanavalin A is not sufficient to induce phosphorylation of pp120. The 120-kDa cell-binding domain of fibronectin can induce some phosphorylation of pp120, but further phosphorylation occurs in the presence also of the heparin-binding domain of fibronectin. Phosphorylation of pp120 precedes, but is correlated with, subsequent cell spreading. Phosphorylation of pp120 can also be triggered by attachment of cells to anti-integrin antibodies, and this requires the cytoplasmic domain of the integrin beta 1 subunit. Thus interaction of beta 1 integrins with extracellular ligands (fibronectin or antibodies) triggers phosphorylation of an intracellular 120-kDa protein, pp120, that may be involved in the responses of cells to attachment.

Topographic map of the HLA-A2 CREG epitopes using human alloantibody probes.

The topographic architecture of the epitopes expressed on the HLA-A2 glycoprotein using murine monoclonal antibody (mAb) probes indicates at least two sterically distinct domains. Previously, we have demonstrated using human HLA alloantibodies (aAb) that multiple determinants are expressed on each HLA antigen: the highly polymorphic private epitopes and the public determinants that are shared within a family of crossreactive groups (CREG). Our objectives now focus on probing the antigenic structure of the HLA-A2-28-9-B17 CREG using highly specific aAb in conjunction with mAb that have previously been used for structural studies. Both mAb-mediated blockage of complement-dependent cytotoxic aAb and reciprocal antibody (Ab) binding inhibition assays with quantitation by fluorescence flow cytometry have been utilized. We have found that xenogeneic mAb directed against A2-69, A2-B17, and A2-28 crossblock aAb of the same serologic specificity, and vice versa, indicating that the epitopes they respectively recognize are at least in close steric proximity. However, additional HLA-A2, A28, and B17 aAb of private specificity and A2-28-9 aAb of public specificity, for which there are no known mAb counterparts, paint an additional complexity not previously known. We conclude that at least four different alloepitopes can be expressed by each serologically defined HLA antigen. Based on the primary sequence data, we have assigned the location and the amino acid substitutions which most likely account for these discrete epitopes. The unique private determinants are located on the alpha 1 domain together with the interlocus A2-B17 epitope while the public epitopes A2-69, A2-28-9 and A2-28 are located on the alpha 2 domain.

Mapping of the functional determinants of the integrin beta 1 cytoplasmic domain by site-directed mutagenesis.

We describe here the expression of deletion mutants of the cytoplasmic domain of the avian integrin beta 1 subunit. These mutants, which contain termination codons at positions 767, 776, 791, and 800, were transfected into mouse 3T3 cells to determine which sequences were essential for localization of integrins into focal contact sites. In all cases, high-level expression of the truncated avian integrins was obtained. Heterodimers were formed between the exogenous truncated avian beta 1 subunits and endogenous mouse alpha subunits, and these heterodimers were efficiently exported to the cell surface. The longest truncated beta 1 subunit tested, which is only four amino acids shorter than the wild type, does localize to focal contacts. In contrast, beta 1 subunits with moderately long truncations of the cytoplasmic domain failed to localize to focal contacts, including one which contains the consensus sequence for tyrosine phosphorylation. Surprisingly, a mutant subunit in which the bulk of the cytoplasmic domain was missing (but the segment nearest the membrane including the dibasic residues (RR) remained) did localize weakly to focal contacts. These results implicate the peptide segment nearest to the transmembrane region in focal contact localization. In addition, mutant subunits that included this segment together with a larger portion of the cytoplasmic domain did not localize as well as the shorter form, suggesting that these cytoplasmic domain segments are defective, presumably because of abnormal folding.

Retroviral expression of alternatively spliced forms of rat fibronectin.

We describe the construction in retroviral vectors and the expression of recombinant rat fibronectin (FN) cDNAs corresponding with the various alternatively spliced forms of FN. In NIH 3T3 cells, the exogenous rat FN subunits are efficiently secreted as heterodimers with endogenous mouse subunits. In contrast, in lymphoid WEHI231 cells, there is no endogenous FN synthesis and the recombinant FNs are secreted and can be purified as homogeneous proteins. We show that the purified recombinant FNs are biochemically and biologically functional. In basic assays for adhesion, spreading, cytoskeletal organization, and migration using various established adherent cell lines, different forms of FNs containing the different alternatively spliced segments show no marked differences in activity. We have used these recombinant FNs to investigate three systems in which earlier results had suggested potential differences between different forms of FN. First, all forms tested appear equally active in restoring normal morphology to a transformed cell line. Second, we detect minor differences in their ability to assemble into preexisting extracellular matrices. Finally, we report that only those forms of FN that contain the V segment will promote the spreading of a lymphoid cell line indicating that this segment confers additional biological functions for some cell types, a result that confirms and extends earlier data. These homogeneous, biologically active recombinant FNs will allow further studies of the role of the alternatively spliced segments of FN.

Expression of normal and mutant avian integrin subunits in rodent cells.

We describe the expression of the beta 1 subunit of avian integrin in rodent cells with the purpose of examining the structure-function relationships of various domains within this subunit. The exogenous subunit is efficiently and stably expressed in 3T3 cells, and it forms hybrid heterodimers with endogenous murine alpha subunits, including alpha 3 and alpha 5. These heterodimers are exported to the cell surface and localize in focal contacts where both extracellular matrix and cytoskeleton associate with the plasma membrane. Hybrid heterodimers consisting of exogenous beta 1 and endogenous alpha subunits bind effectively and specifically to columns of cell-binding fragments of fibronectin. The exogenous avian beta 1 subunit appears to function as well as its endogenous murine equivalent, consistent with the high degree of conservation noted previously for integrins. In contrast, expression of a mutant form of avian integrin beta 1 subunit lacking the cytoplasmic domain produces hybrid heterodimers which, while efficiently exported to the cell surface and still capable of binding fibronectin, do not localize efficiently in focal contacts. This further implicates the cytoplasmic domain of the beta 1 subunit in interactions required for cytoskeletal organization.

Antigenic polymorphism of the T4 differentiation antigen expressed on human T helper/inducer lymphocytes.

The human TH lymphocyte population has been established to express a differentiation antigen (T4) which appears to function in cellular collaboration and T cell recognition of Class II MHC alloantigens. Because we observed altered immunofluorescence staining of the TH cells of some individuals using the OKT4 mAb, a systematic investigation on both the epitopic structure of the T4 glycoprotein molecule and possible polymorphism of these epitopes was undertaken. From competitive blocking assays using eight murine anti-T4 mAbs coupled with quantitative flow cytometry, at least five and possibly seven different epitopes can be recognized on the T4 molecule. Population studies showed some individuals had a reduced phenotypic expression of the OKT4 reactive determinant to one-half that of normal and others completely lacked this epitope. The OKT4 reactive epitope variations are common but have so far been racially restricted to American Blacks and do not appear related to the stage of TH cell differentiation, any identifiable immune abnormality in vitro, or a definable disease process. The OKT4 epitope cannot be unmasked by neuraminidase treatment or T cell stimulation with lectins, soluble antigens, or allogeneic lymphocytes. Coupled with a family study, the alterations in OKT4 phenotype are best explained by autosomal, codominant expression of the T4 gene product. The significance of this polymorphism on TH cell function remains unclear.