Abstract. Two conditions were identified that interfered with the complex polymerization process in biosynthesis of von WiUebrand factor (vWf). Treatment of human umbilical vein endothelial cells with tunicamycin inhibited N-linked glycosylation of nascent vWf and the resulting pro-vWf monomers failed to dimerize. The single subunits accumulated in the endoplasmic reticulum and were neither processed further nor secreted. In the presence of a weak base (ammonium chloride or chloroquine), interdimer disulfide bond formation was inhibited in a dose-dependent manner. This process appeared therefore to be pH sensitive and likely to be initiated in the acidic transGolgi apparatus (Anderson, R. G. W., and R. K. Pathak, 1985, Cell, 40:635-643). The weak base had no obvious effect on the other processing steps, i.e. dimerization, complex carbohydrate formation and sulfation, and produced only slight inhibition of prosequence cleavage. On the other hand, the weak base interfered with the targeting of newly synthesized vWf into Weibel-Palade bodies, with all of the vWf being secreted constitutively and none stored in the WeibelPalade bodies. In summary, initial glycosylation of the nascent vWf protein and low pH in the trans-Golg~ apparatus were important conditions for the successful polymerization of human vWf. Genetic defects disrupting any one of these conditions could result in the phenotype of von Willebrand disease.V ON Willebrand factor (vWf) ~ is an adhesive glycoprotein necessary for binding platelets to subendothelium after vascular injury (12,23,30). vWf is synthesized as a large precursor (260,000-mol-wt) that contains high mannose carbohydrate ( 17, 31). After dimer formation and transport to the Golgi apparatus, the carbohydrate is processed to the complex type, the propiece is cleaved, and interdimer disulfide bond formation begins (32). Carbohydrate processing is accompanied by an apparent increase in molecular weight of the precursor subunit to 275,000. These observed molecular weights of precursor subunits underestimate actual size by ~30,000-50,000, as determined by the size of vWf cDNA clones that span the vWf messenger RNA (9).vWf is stored in the endothelial cells within specific organelles called Weibel-Palade bodies (34). Secreted vWf consists of a series of multimers that range in size from 0.5 to over 20 x 106 daltons, depending upon the number of dimers that are disulfide bonded. In type IIA yon Willebrand disease, only the large and intermediate size multimers are lacking (22), but patient plasma does not support efficient attachment of platelets to subendothelium (24), indicating that multimer size is important for vWf biological activity. As yet, the cellular defects that are responsible for incomplete polymerization of vWf in the type IIA von Willebrand disease are not known. Abbreviation used in this paper: vWf, yon Willebrand factor.Although vWf multimerization is inhibited in vitro by the ionophore monensin, the latter affects all of the processing steps of the Golgi apparatus (33). In t...

Platelets contain receptors for vW protein, and monoclonal antibodies against these membrane glycoproteins inhibit the binding of vW protein (9-13). The binding site in the subendothelium has been considered to be collagen, based on vW protein affinity experiments (14, 15) and electron microscopic observations of its proximity to collagen fibers in the subendothelium (16). We have studied the relationship of vW protein and collagen in the extracellular matrix of endothelial cells and of other cells to which vW protein has been added and have found that the presence of vW protein in the matrix is independent of the presence of collagen. Our results strongly indicate that the major binding site for vW protein in the subendothelium is not collagen. MATERIALS AND METHODSCells and Culture Conditions. Endothelial cells were obtained from human umbilical vein by mild proteolytic digestion as described (8). Cells were cultured in McCoy's 5A medium (Flow Laboratories) containing 20% fetal bovine serum. Human foreskin fibroblasts in 16th passage, a gift from M. Menegus (University of Rochester, Rochester, NY), were grown in the same medium. NIL 8 hamster cells were a gift from R. 0. Hynes (Massachusetts Institute of Technology, Cambridge, MA) and were cultured in Dulbecco's modified Eagle's medium (GIBCO) supplemented with 5% fetal bovine serum.Antisera. The preparation and characterization of antisera against human vW protein have been described (8,17). Other antisera were anti-type IV collagen (18), a gift from G. R. Martin (National Institutes of Health, Bethesda, MD), antitype V collagen (19), a gift from H. E. Sage (University of Washington, Seattle, WA), and rhodamine-conjugated antifibronectin (20), supplied by R. 0. Hynes.Immunofluorescence. Cells grown on glass coverslips were fixed in 3.7% HCHO in phosphate-buffered saline (saline) for 20 min and, if desired, permeabilized in 0.5% Triton X-100 in saline for 15 min at room temperature. Incubations with the first antibody and with goat fluorescein isothiocyanate antirabbit immunoglobulin (Miles) were for 30 min at 37°C. AntivW protein antiserum was used at 1:40 dilution; anti-type IV collagen, at 1:10 dilution; and anti-type V collagen, at 1:20 dilution. For double-label fluorescence, the protocol of Hynes and Destree was used (20). After staining with the first antibody and fluorescein-conjugated goat anti-rabbit immunoglobulin, cells were incubated for 30 min with normal rabbit serum (1:10 dilution) and then stained with rhodamine-conjugated anti-fibronectin for 30 min at 37°C. All intermediate washes were in saline. Coverslips were mounted in Gelvatol. The fluorescein and rhodamine stains showed no interference. For example, when anti-vimentin was used as first antibody, the visualized intermediate filament pattern did not coincide with the straighter extracellular fibronectin pattern.Binding of Exogenous vW Protein to Matrices. Cells were either plated in the presence of vW protein-containing media or switched to this medium after they became established. The ...

Interchain disulfide bonds between the subunits in von Willebrand factor (vWf) dimers and in vWf multimers have been studied using some unique features of the cultured human umbilical vein endothelial cell system. Ammonium chloride inhibition of multimerization of vWf allowed selective examination of vWf dimeric molecules, and monoclonal antibody against the vWf propolypeptide was used to separate pro-vWf dimers from mature dimers. After cleavage of dimers and multimers with Staphylococcus aureus V-8 protease, the location of interchain disulfide bonds in amino (N)-terminal or carboxyl (C)-terminal fragments was determined by gel electrophoresis under reduced and nonreduced conditions. The first interchain disulfide bonds formed during dimerization are in the C-terminal region of the subunits, whereas interdimer disulfide bonds are located in the N-terminal portion. These data confirm recent electron microscopic projections of disulfide bond locations and provide support to the hypothetical role of the propolypeptide in the multimerization process.

P-selectin is expressed on the surfaces of activated platelets and endothelium where it mediates binding to leukocytes. P-selectin- deficient mice were shown to exhibit peripheral neutrophilia (Mayadas et al: Cell 74:541, 1993). We now show that this is not caused by changes in bone marrow precursors nor by a lack of neutrophil margination. Both P-selectin-positive and -negative animals displayed similar increases in peripheral blood neutrophil numbers after injection of epinephrine. However, clearance of 51Chromium-labeled neutrophils is delayed in mice deficient for P-selectin, indicating that the neutrophilia is at least in part the result of delayed removal. We detected no obvious alterations in lymphocyte differentiation, distribution, or adhesion to high endothelial venules in peripheral lymph nodes. Through intravital microscopy, we examined the impact of P-selectin deficiency on leukocyte/endothelial interaction beyond the initial stages of inflammation. Four hours after the administration of an inflammatory irritant, leukocyte rolling was observed even in the absence of P-selectin. There were significantly fewer rolling cells relative to wild-type mice, and their velocity was reduced. Moreover, in the peritonitis model, the number of peritoneal macrophages in wild-type mice increased threefold at 48 hours, whereas the macrophages in the mutant mice remained near baseline levels. Thus, whereas P-selectin is known to be involved in early stages of an inflammatory response, our results indicate that it is additionally responsible for leukocyte rolling and macrophage recruitment in more prolonged tissue injury.