Structural characterization of insulin receptors. I. Hydrodynamic properties of receptors from turkey erythrocytes.

Insulin receptors from turkey erythrocyte plasma membranes were solubilized in nondenaturing detergents (Triton X-100 and sodium deoxycholate). Their hydrodynamic properties were determined by sedimentation analyses in H2O and D2O, and gel filtration on Sepharose 4B. Two specific insulin-binding species are observed after velocity sedimentation in linear sucrose density gradients: peaks I and II. In Triton X-100, the sedimentation coefficient (s20,w), partial specific volume (Vc), and Stokes radius (a) for peaks I and II are, respectively, 10.2 +/- 0.5 S and 6.6 +/- 0.5 S, 0.75 +/- 0.02 ml/g, and 0.76 +/- 0.02 ml/g, and 89 +/- 3 A and 76 +/- 3 A, to yield Mr = 410,000 +/- 75,000 and 235,000 +/- 55,000, respectively, for the protein-Triton X-100 complex. The corresponding values in deoxycholate solution are: 10.7 +/- 0.5 S and 6.9 +/- 0.5 S, 0.71 +/- 0.03 ml/g and 0.70 +/- 0.04 ml/g, and 86 +/- 3 A and 69 +/- 3 A for peaks I and II, respectively, to yield 360,000 +/- 65,000 and 180,000 +/- 45,000, respectively, for the molecular weight of the protein-deoxycholate complex. These data are consistent with a model whereby each receptor species binds to one micelle of the appropriate detergent. In agreement with this model, it was also found that, in both Triton X-100 and deoxycholate, concentrations higher than the critical micellar concentration are required in order to maintain discrete receptor species in solution. At concentrations below the critical micellar concentration, the receptors aggregate to a broad band that sediments faster than 11.3 S. This is typical of membrane proteins that are stabilized in solution by insertion into detergent micelles. Based on these results, the protein molecular weights of peaks I and II are estimated to be 355,000 +/- 65,000 and 180,000 +/- 45,000, respectively. When membranes are treated with the reducing agent dithiothreitol, peak I is converted to peak II. This fact, together with the estimates obtained for the protein molecular weights of the two receptor species, suggests that peak I is a disulfide-linked dimer of peak II. The sedimentation characteristics of insulin receptors in many different cell types appear to be similar. As with turkey erythrocytes, detergent extracts of membranes from rat liver contained two native receptor species whose sedimentation coefficients were similar to peaks I and II. However, in all the other cell types examined, including rat adipocytes, rat heart muscle, 3T3-L1 adipocytes, 3T3-C2 fibroblasts, and FAO hepatoma cells, peak I (the native dimer) was the predominant species observed.

Structural characterization of insulin receptors. II. Subunit composition of receptors from turkey erythrocytes.

In the preceding paper (Aiyer, R. A. (1983) J. Biol. Chem. 258, 14992-14999), the hydrodynamic properties of insulin receptors from turkey erythrocyte plasma membranes solubilized in nondenaturing detergents (Triton X-100 and sodium deoxycholate) were characterized. Two specific insulin-binding species are observed after velocity sedimentation in linear sucrose density gradients: peak II whose protein molecular weight (Mp) is 180,000 +/- 45,000 and its disulfide-linked dimer, peak I (Mp, 355,000 +/- 65,000). This paper describes the subunit composition of these species determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Insulin receptors were covalently attached to [125I]iodoinsulin with disuccinimidyl suberate. After solubilization in Triton X-100 or deoxycholate, peaks I and II were separated by sedimentation and subjected to SDS-PAGE; the constituent polypeptides were then identified by autoradiography. Under reducing conditions, both peaks I and II yield a major band of apparent molecular weight (Mapp) of 135,000; this band most likely represents the insulin-binding subunit (alpha). Minor bands of lower molecular weight are also seen whose significance is not entirely obvious. Under nonreducing conditions, peak I yields bands at Mapp = 230,000 and at greater than 240,000, while peak II yields bands at Mapp = 120,000 and 200,000. When these bands were cut out of the gel and subjected to SDS-PAGE following reduction with 10% beta-mercaptoethanol, all of them produced a single band that migrated with Mapp = 135,000. These results indicate that the alpha subunit is linked by disulfide bonds to at least one more subunit (beta). It is also apparent that the alpha subunit travels with higher mobility (Mapp = 120,000) under nonreducing conditions, suggesting the presence of intrachain disulfide bonds. Thus, peak II has a minimum subunit composition of alpha beta, where alpha is the insulin-binding subunit with a minimum Mr = 120,000-135,000 and beta has a minimum Mr = 80,000-90,000. And peak I, the disulfide-linked dimer of peak II, has a minimum subunit composition of alpha 2 beta 2. These results were further confirmed by cross-linking of protein subunits with glutaraldehyde, an (alpha, omega)-dialdehyde that reacts with amino groups. Within the limits of error, these molecular weights are in agreement with those estimated from the hydrodynamic properties of the detergent-solubilized, native receptor species reported in the preceding paper.

Identification of the alpha beta monomer of the adipocyte insulin receptor by insulin binding and autophosphorylation.

The insulin receptor consists of an insulin-binding subunit (alpha) of 135,000 daltons. More recently, it has been documented that the receptor undergoes insulin-stimulated autophosphorylation that predominantly labels a 95,000-dalton (beta) subunit. We solubilized rat adipocyte insulin receptors in Triton X-100 and partially purified the protein on a wheat germ agglutinin-Sepharose affinity column. Subsequently, we labeled the two subunits of the receptor independently by using 125I-labeled insulin for the 135,000-dalton alpha-subunit and 32P for the 95,000-dalton beta-subunit. Sucrose density gradient sedimentation and NaDodSO4/PAGE were used to characterize the native, oligomeric structure of the receptor. In 0.1% Triton X-100, the receptor sedimented as a single peak of s20,w = 10.2 S as detected by both 125I and 32P. NaDodSO4/PAGE under nonreducing conditions revealed a large species that appeared to be alpha 2 beta 2 and, to a lesser extent, alpha beta. Treatment of the solubilized, partially purified receptor with 10 mM dithiothreitol led to the partial conversion of the 10.2S species to a smaller one sedimenting at 6.6 S. The composition of this species was determined to be alpha beta by nonreducing NaDodSO4/PAGE. Our results suggest that detergent-solubilized insulin receptors can exist as dimers and monomers. The oligomeric structure of receptors functional in the cell membrane cannot be immediately deduced from these results due to the possibility of artifacts arising from membrane disruption and extraction procedures. However, the ability to label the two subunits of the receptor separately should facilitate a detailed study of its oligomeric structure both in solution and in the membrane.

Characterization of receptors for recombinant human tumor necrosis factor-alpha from human placental membranes.

High affinity receptors for recombinant human tumor necrosis factor-alpha (rhTNF-alpha) were identified on membranes prepared from full term human placenta. Highly purified rhTNF-alpha iodinated by the iodogen method was found to bind placental membranes in a displaceable manner with an approximate dissociation constant (KD) of 1.9 nM. The membrane bound TNF-alpha receptor could be solubilized by several detergents with optimum extraction being obtained with 1% Triton X-100. The binding of 125I-rhTNF-alpha to the solubilized receptor was found to be time and temperature dependent, yielding maximum binding within 1 h, 24 h and 48 h at 37 degrees C, 24 degrees C and 4 degrees C, respectively. However, the maximum binding obtainable at 4 degrees C was only 40% of that at 37 degrees C. The binding 125I-rhTNF-alpha to solubilized placental membrane extracts was displaceable by unlabeled rhTNF-alpha, but not by a related protein recombinant human tumor necrosis factor-beta (rhTNF-beta; previously called lymphotoxin). This is similar to the behavior of TNF-alpha receptors derived from detergent-solubilized cell extracts, although on intact cells, both rhTNF-alpha and rhTNF-beta bind with equal affinity to TNF receptors. The Scatchard analysis of the binding data of the solubilized receptor revealed high affinity binding sites with a KD of approximately 0.5 nM and a receptor concentration of about 1 pmole/mg protein. Gel filtration of the solubilized receptor-ligand complexes on Sephacryl S-300 revealed two different peaks of radioactivity at approximate molecular masses of 50,000 Da and 400,000 Da. The 400,000 dalton peak corresponded to the receptor-ligand complex. Overall, our results suggest that high affinity receptors for TNF-alpha are present on human placental membranes and provide evidence that these receptors may be different from that of rhTNF-beta.

Interferon-gamma binds to high and low affinity receptor components on murine macrophages.

Interferon-gamma (IFN-gamma), a glycoprotein secreted by antigen-or mitogen-activated lymphocytes, modulates the activities of lymphocytes and macrophages. For mouse macrophages, murine IFN-gamma (MuIFN-gamma) stimulates several functions, including phagocytosis, tumoricidal activity, and the increased expression of Ia and H-2 antigens of the major histocompatibility complex. Recent reports have suggested that IFN-gamma specifically binds to cell surface receptors corresponding to a single class of binding sites with a Kd of 10(-8) to 10(-10) M. We present evidence that, on the murine macrophage cell line WEHI-3, MuIFN-gamma specifically binds to two classes of binding sites with Kd of 9.1 X 10(-11) M (500 sites/cell) and 2.7 X 10(-9) M (4400 sites/cell). The higher affinity sites most likely represent the physiologically relevant receptors that mediate at least some of the actions of MuIFN-gamma.

Human tumor necrosis factor-alpha receptor. Purification by immunoaffinity chromatography and initial characterization.

The receptor for human tumor necrosis factor-alpha (TNF-alpha) was isolated from a subclone of the human histiocytic lymphoma cell line U937. These cells exhibit a single class of high affinity receptors (Kd = 0.51 +/- 0.25 nM) with an average density of 55,000 +/- 5,000 binding sites/cell. After solubilization with detergent, the receptor retained its ability to bind free TNF-alpha but failed to bind to TNF-alpha immobilized on various solid supports. For receptor purification, 125I-TNF-alpha was covalently attached to the receptor on intact cells by the bifunctional cross-linking reagents ethylene glycolbis(succinimidylsuccinate) or 3,3-dithiobis(sulfosuccinimidylpropionate). The cells were then solubilized with the nonionic detergent Triton X-100, and the supernatants, clarified by centrifugation, were passed over an IgG-Sepharose column prepared from TNF-alpha antiserum. The receptor-rich fraction from the antibody column was further purified by preparative sodium dodecyl sulfate-polyacrylamide gel electrophoresis. These two steps together provided approximately 165,000-fold purification of the TNF-alpha receptor. The TNF-alpha receptor-ligand complex obtained by this method had a subunit molecular weight of 100,000 +/- 5,000 when examined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis but on gel filtration the complex migrated with an apparent molecular weight of 480,000 +/- 32,000. However, the receptor showed a molecular weight of 65,000 +/- 32,000 when gel filtration was performed in the absence of ligand. Additional characteristics of the receptor are discussed.

Human tumour necrosis factors: structure and receptor interactions.

Activation of lymphoid and myeloid cells causes the production of factors cytotoxic to various tumour cell types in vitro and in vivo. We have investigated the biochemistry, molecular biology and mechanism of action of two such factors. The factor derived from a myeloid cell line was named TNF-alpha (previously referred to as TNF) and that derived from lymphoid cells named TNF-beta (previously called lymphotoxin). Both proteins were purified from the conditioned media of the human cell lines and sequenced. Structural information revealed that TNF-alpha is 157 amino acid residues long and contains one disulphide bond. TNF-beta is a glycoprotein of 171 amino acids that contains no cysteine residues. Protein sequence information was used to isolate and characterize cDNAs for TNF-alpha and TNF-beta by recombinant DNA methods. The expression of the cDNAs in Escherichia coli made available large quantities of these proteins for biological studies. The two proteins are 31% identical and 52% homologous to each other. The genes for both cytokines are approximately three kilobases in size and are closely linked on human chromosome six. TNF-alpha and TNF-beta both bind to various cell types via a single class of high affinity receptors. On most cells the same receptor is recognized by both cytokines. The receptors for TNF-alpha can be up-regulated by both interferons and lectins. Up-regulation of receptors by interferons is accompanied by synergistic enhancement of the biological response whereas up-regulation by lectins results in an antagonistic response. Besides antiproliferative effects, both cytokines exhibit direct antiviral effects on infection by both DNA and RNA viruses.

Detection of hepatitis B virus in plasma using flow cytometric analyses of polymerase chain reaction-amplified DNA incorporating digoxigenin-11-dUTP.

Blood donations are routinely screened by multiple serologic assays for antigens/antibodies associated with infection by blood-borne viruses, including hepatitis B virus (HBV), hepatitis C virus (HCV), human immunodeficiency viruses (HIV-1 and HIV-2), and human T-cell lymphotropic virus (HTLV-I and HTLV-II). A direct detection of these viruses would be more effective for the prevention of transfusion-transmitted infections than the indirect measurement of the variable host immune response to these agents. Because the polymerase chain reaction (PCR) for viral gene amplification offers the most sensitive and direct means of detecting viruses in blood, we have developed a nonisotopic PCR procedure for the detection of HBV, chosen as a prototype. The problems, common to previously described PCR methods, of nucleic acid extraction and inhibition of the PCR by plasma proteins were overcome by isolation of HBV from plasma by means of 450-microns polystyrene beads covalently coated with monoclonal antibody to the Pre-S1 region of the viral envelope protein. Detergent lysis and proteinase K digestion of the immunocaptured virions isolated from plasma released the HBV DNA. A modified PCR-amplification protocol, incorporating digoxigenin-labeled dUTP in the amplified gene products followed by hybridization with a specific biotinylated oligonucleotide probe bound to streptavidin-coated 2.8-microns magnetic beads, allowed flow cytometric analyses of HBV-specific PCR products by means of antibodies to digoxigenin labeled with fluorescein isothiocyanate. The endpoint serial dilutions of pedigreed human plasma samples containing chimpanzee infectious dose (CID50) of 10(7) for adw and CID50 of 10(7.5) for the ayw subtypes were compared in repeated testing of PCR products by our immunoreactive bead (PCR-IRB) assay. HBV DNA was consistently detected in a 5 x 10(-10) dilution of each sample. In testing 20 coded specimens of blood donors, with or without serologic markers of HBV infection, the PCR-IRB was specific and more sensitive than the PCR analyses by slot blot hybridization with radioactive probe. The PCR-IRB assay can be adapted for simultaneous detection of multiple blood-borne viruses by an automated flow cytometric analysis system.

Tumor necrosis factor-alpha as a proliferative signal for an IL-2-dependent T cell line: strict species specificity of action.

The ability of TNF-alpha to stimulate T cell proliferation was examined. We demonstrate that murine rTNF-alpha induces the proliferation of CT6, a murine T cell line previously thought to be responsive only to IL-2. This activity appears to be the result of the direct action of murine rTNF-alpha on the CT6 cells because neither 1) murine IL-2 or murine IL-4, lymphokines also capable of inducing CT6 proliferation, were detected in culture supernatants from murine rTNF-alpha-treated CT6 cells nor 2) did antibodies specific for IL-2 or IL-4 inhibit murine rTNF-alpha-induced CT6 proliferation. Unlike many of the activities displayed by TNF-alpha, its ability to induce CT6 cell proliferation shows strict species specificity as indicated by the failure of human rTNF-alpha to stimulate these cells. Flow cytometric analysis and binding of radiolabeled TNF-alpha have indicated that receptors for TNF-alpha on these cells are specific for murine TNF-alpha. The ability of murine rTNF-alpha to induce the proliferation of certain T cell lines further indicates that this molecule plays an important role in regulation of T cell-mediated immune responses.