On the nature of the presumed receptor for IgE on mast cells. V. Enhanced binding of 125I-labeled IgE to cell-free particulate fractions in the presence of protease inhibitors

Cell-free particulate preparations derived from sonicated, purified rat peritoneal mast cells can be radiolabeled with IgE. The IgE concentration-dependence and the kinetics of the binding reaction at pH 6.6 were very similar to those observed with intact mast cells. The number of IgE molecules bound per mast cell equivalent at saturation varied somewhat from one particulate preparation to the next and, on average, was about 50% of the binding capacity of the intact cells. Incubation of intact mast cells or of particulate fractions with low concentrations of IgE in serum-free media resulted in a decrease in the amount of bindable IgE which remained in the supernatants in excess over the amount of IgE which was bound to the cells. Specific binding of IgE to both particles and to intact cells at limiting IgE concentrations was stimulated up to approximately twofold by a variety of antibiotic, synthetic or high molecular weight (protein) protease inhibitors of which soybean trypsin inhibitor and p-nitrophenyl-p'-guanidinobenzoate were the most active. These substances also markedly protected the IgE, leaving more of it bindable to rat basophil leukemic cells in a second incubation. Binding to particulate fractions also occurred at pH 4.8 when bovine serum albumin was added to the incubations. The results are consistent with the view that normal peritoneal mast cells have a membrane-bound protease which has high selectivity for IgE.     

Rat mast cell granules reacting with a mouse monoclonal antibody M6764 which recognizes the same epitope of a mouse monoclonal antibody HNK-1.

Rat mast cell granules and plasma membrane fractions were obtained by homogenization of highly purified rat mast cells and isolation in a Percoll gradient and a sucrose gradient, respectively. Immunostaining of rat mast cells, granules and plasma membrane fractions was performed with mouse monoclonal antibody M6764 which was produced against the crude membrane fractions of the neural tubes. Rat mast cells and granules were immunostained with the monoclonal antibody, but not the plasma membrane fractions. The granules fixed with glutaraldehyde-paraformaldehyde showed ring-like forms. Chloroform-methanol treatment did not effect the staining of rat mast cells and granules with the monoclonal antibody. Western blotting analysis of rat mast cells and granules with the monoclonal antibody showed broad protein bands ranging from 100 to 250 kD.     

Translocation of Protein Kinase C to Subcellular Fractions of Human Neutrophils

The subcellular localization of protein kinase C in unstimulated human neutrophils and neutrophils stimulated by phorbol-myristate-acetate (PMA), 1-oleoyl-2-acetyl-rac-glycerol (OAG), and ionomycin was investigated in subcellular fractions obtained by nitrogen cavitation and Percoll density gradient centrifugation. Protein kinase C was found to be localized mainly in the cytosol in unstimulated cells, whereas significant translocation to fractions containing the plasma membrane was observed after stimulation by PMA, OAG, and ionomycin. At the same time, phospholipid-insensitive protein kinase activity appeared in the cytosol and the plasma membrane fractions. To determine whether binding of protein kinase C occurred to the plasma membrane or to intracellular membranes that had translocated to the plasma membrane, we investigated the ability of isolated azurophil, specific and secretory granules, and plasma membrane vesicles to bind protein kinase C in response to addition of PMA and OAG. Only fractions containing plasma membranes and secretory granules were able to bind protein kinase C. The observation explains the selective activation of plasma membrane structures by protein kinase C.     

Preparation and characterization of mast cell cytoplasts

A new model system for studying biochemical reactions in mast cell plasma membranes was developed. Particles termed cytoplasts consisting of organelle-depleted cytoplasm surrounded by an intact plasma membrane were formed from cytochalasin B-treated mast cells ultracentrifuged through a discontinuous Ficoll gradient. Two cytoplasts were formed per mast cell and 95% were recovered. Mast cell cytoplasts had a mean diameter of 3.2 microns with a median volume of 38 microns 3. Enzyme marker studies indicated that subcellular recoveries in the mast cell cytoplast were: plasma membrane = 16%, cytoplasm = 39%, nucleus = 1.1%, granule = 0.5%. Analysis of IgE receptors indicated that mast cell cytoplasts retained the normal asymmetric orientation of the plasma membrane. Mast cell cytoplasts synthesized ATP, incorporated labeled fatty acids into complex lipids and retained fluorescein after deacylation of diacetylfluorescein. The quantity of cAMP (adenosine 3':5'-cyclic monophosphate) maintained in mast cell cytoplasts was 0.0304 pmol/10(6) original mast cells. Cytoplasts offer the opportunity to study plasma membrane and cytoplasmic biochemical events that occur during stimulation in a relatively physiologic environment.     

Subcellular fractionation of amoebapore and plasma membrane components of Entamoeba histolytica using self-generating Percoll gradients

Separation and initial characterization of subcellular organelles from Entamoeba histolytica have been achieved by the use of self-generating Percoll gradients. Adequate resolution was obtained within one hour of amoeba homogenization. The ion-channel forming activity, amoebapore, was found associated with highly dense, small diameter particles, that were resolved from the numerous digestive vacuoles, and the plasma membrane. Following iodination of intact trophozoites with lactoperoxidase, the label was found to be incorporated into two distinct sedimentable fractions. The major component contained the near totality of the concanavalin A binding glycoproteins and the 5-iodonaphthalene-1-azide labelled intrinsic membrane proteins. It therefore was identified as the plasma membrane. The minor lactoperoxidase-iodinated component was found in the upper section of the gradient, associated with a particulate fraction nearly devoid of concanavalin A-binding glycoproteins. This fraction appears to represent particulate material on the external surface of the amoeba that is distinct from the plasma membrane. The digestive vesicles, which were identified by the presence of acid phosphatase and beta-D-glucosaminidase activities, appeared in the gradient as one main peak with a shoulder in the region of the plasma membrane. Analysis of the polypeptides and their labelling pattern in each fraction of the gradient are presented. Distinctive characteristics are discussed including the identification of a unique soluble protein highly labelled by 5-iodonaphthalene-1-azide.     

A neutrophil-dependent pathway for the generation of a neutral peptide mediator. II. Subcellular localization of the neutrophil protease.

The human neutrophil neutral peptide-generating protease was associated with the plasma membrane marker 5'-nucleotidase on sucrose density gradient centrifugation of sonicates of granule-free fractions following homogenization and velocity sedimentation. The two activities were also associated on sucrose density gradient fractionation of plasma membranes obtained by hypotonic lysis in EDTA containing buffers, a technique which minimizes aggregation. Treatment of fractions containing these enzymatic activities with 1-0 M NaCl separated the neutral peptide-generating proteasein to the eluate while leaving the 5'-nucleotidase in the pellet. Gel filtration of the solubilized neutral peptide-generating protease through Sephadex G-100 in 1-0 M NaCl demonstrated that the protease had an approximate mol. wt of 20,000 while filtration in physiological salt concentrations yielded activity only in the excluded volume. In both cases, there was complete recovery of neutral peptide-generating activity suggesting that the filtration characteristics of the protease were determined by the salt concentration. The solubilized purified protease, the whole cell sonicates, and the intact cells interacted with heat-inactivated plasma to yield the same product, a neutral peptide with a 1000 molecular weight and an isoelectric point of 7-2-7-6. The neutral peptide-generating protease in each instance was inhibited in dose-response fashion by alpha-1-antitrypsin, LBTI, and DFP. Only 30-60% of the protease sites were functional on intact cells as revealed by substrate cleavage or were available to inhibitors. The neutrophil protease which generates neutral peptide is an extrinsic plasma membrane protein with an approximate mol. wt of 20,000 which functions as an ectoenzyme.     

Seasonal variations in in vitro synthesis of rye pollen-specific IgE by human peripheral mononuclear cells. Functional heterogeneity within the IgE-B pool

Peripheral blood mononuclear B cells from 6 rye pollen-allergic patients, with no consistent perennial symptoms, were isolated before (July), during (October) and after (February and May) the pollen season. The T-depleted cells were fractionated on a discontinuous percoll density gradient and the B cell fractions, together with unfractionated B cells, incubated in vitro for quantitation of spontaneous synthesis of rye pollen-specific IgE. Markedly higher levels of IgE were synthesised by the fractions, as opposed to unfractionated B cells. The low-density fraction (B5) contributed most towards synthesis in the pollen season and the denser B6 cells in the pre- and post-pollen season. All low-density B cell fractions (B1-3-B5) and some high-density fractions contained large but variable amounts of preformed specific IgE which was retained, even in the absence of de novo synthesis in vitro, during the post- and pre-pollen season. Since only a fraction of this preformed IgE escaped into culture supernatants the contribution of preformed IgE to in vitro IgE synthesis in general may require reappraisal.     

On the nature of the presumed receptor for IgE on mast cells. IV. Inhibition of the PCA-blocking activity of cell-free particulate preparations and intact rat peritoneal mast cells by inhibitors of proteases.

The incubation of IgE-containing solutions from rat serum with particulate preparations from rat peritoneal mast cells results in the disappearance of some of the PCA-reactive IgE in the solution. This PCA blocking assay was used to measure the 'binding' of IgE to intact rat mast cells or to particulate preparations derived from mast cells. The PCA-blocking activity at pH 4.8 was up to 100-fold greater than that seen at a physiologic pH of 6.6. PCA-blodking activity was inhibited at both these pH conditions by high concentrations of several trypsin inhibitors. The inhibitors were generally more active at the more acid pH. Among the active inhibitors were soybean and limabean trypsin inhibitors, chymostatin, and p-nitrophenyl-p'-guanidinobenzoate. Inhibitors of acid proteases, such as pepstatin and diazaacetylnorleucine methyl ester were inactive. The results support the proposition that under certain conditions IgE degradation by a specific proteolytic enzyme which is located uniquely on the plasma membrane of mast cells can account for a major portion of the PCA-blocking activity of these cells.     

Solubilization of tumour-specific antigen from plasma membrane of an aminoazo-dye-induced rat hepatoma

Plasma membrane fractions isolated from cells of an aminoazo-dye-induced rat hepatoma were solubilized by limited papain digestion. DEAE-cellulose chromatography yielded a discrete component retaining tumour-specific antigen as measured by its capacity to neutralize antibody in tumour-immune sera, which reacts in immunofluorescence tests with the plasma membrane of intact hepatoma cells. Solubilized plasma membrane fractions also elicited a tumour-specific humoral antibody response in syngeneic rats. The relative inefficiency of antigen isolation procedures would seem to preclude the use of solubilized antigen for immunotherapy, but these preparations are important in studying the nature of tumour antigen expression during chemical carcinogenesis and for analysing the involvement of antigen–antibody complexes in tumour immunity.     

Membranes of activated CD4+ T cells expressing T cell receptor (TcR) alpha beta or TcR gamma delta induce IgE synthesis by human B cells in the presence of interleukin-4.

In the present study it is demonstrated that human B cells can be induced to switch to IgE production following a contact-mediated signal provided by activated T cell receptor (TcR) gamma delta+, CD4+ and TcR alpha beta+, CD4+ T cell clones and interleukin (IL)-4. The signal provided by these T cell clones was antigen nonspecific, indicating that the TcR alpha beta/CD3 or TcR gamma delta/CD3 complexes were not involved in these T-B cell interactions. Activated TcR alpha beta+, CD8+, and TcR gamma delta+, CD4-CD8-, or resting CD4+ T cell clones were ineffective. Intact TcR alpha beta+ or TcR gamma delta+, CD4+ T cell clones could be replaced by plasma membrane-enriched fractions isolated from these activated CD4+ T cell clones. In contrast, membranes isolated from resting TcR alpha beta+, CD4+, TcR gamma delta+, CD4+ T cell clones or an Epstein-Barr virus (EBV)-transformed B cell line (EBV-LCL) failed to provide the costimulatory signal that, in addition to IL-4, is required for induction of IgE synthesis. As described for intact CD4+ T cells, CD4+ T cell membranes induced purified surface IgM+ B cells to switch to IgG4- and IgE- but not to IgA-producing cells, excluding the possibility of a preferential outgrowth of IgG4- and IgE-committed B cells. The membrane activity was inhibited by protease or heat treatment. Induction of IgE synthesis by B cells co-cultured with both TcR alpha beta+, CD4+ and TcR gamma delta+, CD4+ T cell clones and membrane preparations of these cells was blocked by anti-class II major histocompatibility complex (MHC) monoclonal antibodies (mAb), whereas various anti-CD4 mAb had differential blocking effects. Murine L cells, or EBV-LCL transfected with CD4 could not replace CD4+ T cell clones. These results indicate that, although CD4 and class II MHC antigens are required for productive CD4+ T cell clone-B cell interactions, an additional signal, provided by a membrane associated (glyco)protein that is induced by activation of both TcR alpha beta and TcR gamma delta, CD4+ T cells, is needed for induction of IgE production in the presence of IL-4.     

Characterization of the low-affinity receptor for IgE on murine B cells by using membrane impermeant cross-linking reagents

The murine B lymphocyte receptor for the Fc portion of IgE (Fc epsilon R) was further characterized by using the membrane impermeable cross-linking reagents 3,3'-dithiobis-(sulfosuccinimidyl) proprionate (DTSSP) and bis-(sulfosuccinimidyl) suberate (BS3). IgE could be cross-linked to the Fc epsilon R on the intact cells with either reagent and, in addition, up to 10% of the B cell surface immunoglobulin (sIg; both IgM and IgD) was also found to cross-link to the IgE/Fc epsilon R complex. Analysis of isolated sIg/IgE/Fc epsilon R complexes indicated that about 60% of the Fc epsilon molecules were becoming cross-linked to sIg. Thus, the data suggest that on the intact murine B cell the Fc epsilon R is frequently in close association with sIg. The murine B cell Fc epsilon R was also examined for the presence of receptor-associated proteins that are buried in the membrane. Advantage was taken of the membrane-impermeant nature of DTSSP and BS3. IgE was cross-linked to the Fc epsilon R on intact cells by using the disulfide-cleavable DTSSP and following solubilization with nonionic detergent; BS3 was used to cross-link possible internal membrane components to the Fc epsilon R. In these experiments, the high-affinity Fc epsilon R on rat basophilic leukemia (RBL) cells could be cross-linked to a nonreducible high molecular weight complex of 100 kilodaltons. However, when intact murine B cells were treated with DTSSP, solubilized and treated with BS3 in the same manner as indicated above, no evidence was found for the presence of membrane-buried receptor-associated proteins with the B cell Fc epsilon R. Thus, these data further support the concept that there may be little relationship between the high-affinity mast cell/basophil Fc epsilon R and the low-affinity lymphocyte Fc epsilon R.     

Interaction of thymopoietin peptides with the specific receptor of facteur thymique serique (FTS)

The capacity of ubiquitin (UB) and thymopoietin II (TP) related peptides (TP 5 and TP 13) to interfere with the specific binding of [3H]FTS on intact 1301 cells or on 1301 plasma membrane preparations was studied. All 3 peptides significantly inhibited the binding of [3H]FTS to its receptors on intact 1301 cells at concentrations 20 to 100 times higher than FTS itself. Conversely, none of the 3 peptides provided a significant inhibition of the specific [3H]FTS binding to plasma membrane preparations of 1301 cells compared to control peptides. These contrasted results suggest that TP-related peptides and, to a lesser degree, UB may share the same target cells with FTS and interfere with FTS effects on T cells, but this interaction probably does not involve direct high affinity binding to the FTS receptors.     

Evidence for the presence of immunoglobulin E antibodies specific to the cell wall phosphomannoproteins of Candida albicans in patients with allergies.

To determine the major antigenic component of Candida albicans against immunoglobulin E (IgE) antibodies in the sera of patients with allergies who were positive for IgE antibodies to C. albicans crude antigen in a CAP system, phosphomannoproteins (CAMP/A or CAMP/B for serotype A or B strain, respectively) and their acid-stable portions (CAMP-S/A or CAMP-S/B) were isolated from beta-mercaptoethanol (2-ME) extracts of C. albicans cells of serotypes A and B, and IgE antibodies against these components were compared with those against protein complex and enolase (CAE) fractions isolated from C. albicans cells. The dot blot test, which was used to detect IgE antibodies to the C. albicans antigens, showed that IgE antibodies to the 2-ME extract and phosphomannoprotein fractions were present in the sera of 98.0% (2-ME extract), 96.8% (CAMP/A), 93.2% (CAMP-S/A), 97.2% (CAMP/B), and 81.5% (CAMP-S/B) of the patients, whereas IgE antibodies to the protein complex and CAE fractions were found in the sera of 73.6 and 48.8% of the patients, respectively. The extent of IgE binding to the 2-ME extract and phosphomannoproteins was well correlated with the fluorescence intensities estimated with the CAP system. Furthermore, the results obtained from the inhibition experiment with the CAP system indicated that the binding of IgE antibodies to Candida antigens is strongly inhibited by the phosphomannoprotein fraction and is an indication that the serum of the patients contained IgE antibodies specific to the cell wall phosphomannoproteins of C. albicans. Finally, an initial chemical analysis indicated that the epitopes for IgE antibodies on the phosphomannoproteins is a carbohydrate portion, since the ability of CAMP/A to inhibit the binding of IgE antibodies to the homologous CAMP/A was destroyed after oxidation by sodium periodate but not after digestion with proteinase K.     

Basophil sensitivity and reactivity to monoclonal anti-human IgE afterin vitro sensitization with human myeloma IgE

Leucocytes from human peripheral blood were acid eluted and sensitizedin vitro with increasing concentrations of radiolabelled myeloma IgE. This sensitization step was performed with or without 30% IgE depleted serum. After the IgE binding, cells were washed and submitted to a challenge with monoclonal anti-IgE for the determination of the cellular sensitivity and reactivity in a histamine release assay. A sample of each of the sensitized cells was analyzed for its radioactivity and the number of basophils quantified, thus allowing the determination of the mean number of IgE molecules per basophil. Raising the IgE concentrations in the sensitization procedure led to an increase of the IgE on the basophil membrane, and to a concommitant elevation of the cell sensitivity. The presence of serum during the binding of IgE onto the cells lowers slightly the binding of IgE to the basophils but decreases strongly the cellular reactivity.     

Identification of orchard grass (Dactylis glomerata) pollen allergens following electrophoretic transfer to nitrocellulose

Orchard grass (cocksfoot) pollen extracts, fractionated by polyacrylamide gradient electrophoresis or SDS gel electrophoresis were electroblotted onto nitrocellulose membranes and probed with sera from orchard grass pollen-allergic patients and 125I-anti-human IgE. The IgE-binding components of the pollen were detected by autoradiography. Elution studies showed that allergens could be extracted immediately and continuously over a 3-hour period. Two fractions of MWs 28,000 and 30,000 could be detected only after 20 min extraction. SDS-PAGE separations gave the better resolution revealing 19 electrophoretically-separate components, 13 of which bound human IgE. All of the IgE-binding components had MWs in the range 14,000 - 70,000. Three of the bands bound IgE from more than 85% of the serum samples. Following gradient gel electrophoresis, IgE binding was exhibited by 10 bands in the range MW 5,000 to greater than 669,000. The technique used allows one to quantitatively examine patients' sera for allergen-specific IgE antibodies and to identify the clinically important allergens. Results revealed numerous allergenic components over a wide MW range while patterns of IgE binding with different patients' sera demonstrated a great diversity of IgE antibody responses. This study demonstrates the suitability of the electroblotting technique combined with autoradiography for the investigation of allergenic components of grass pollen extracts and hence has application to extract standardization and immunotherapy. Such studies can be carried out rapidly, economically and with a high degree of sensitivity.     

Immunoglobulin E (IgE) and IgE-containing cells in human gastrointestinal fluids and tissues.

Human gastric, small intestinal, colonic and rectal mucosae were examined for IgE-containing cells by single- and double-antibody immunofluorescence techniques, and IgE in intesinal fluids was measured by a double-antibody radioimmunoassay. IgE-containing cells were identified in all tissue specimens and comprised about 2% of all immunoglobulin-containing cells. Although less numerous than cells containing IgA, IgM or IgG, they were remarkably numerous in relation to the concentration of IgE in serum (about 0-001% of total immunoglobulin). IgE immunocytes were significantly more numerous in stomach and proximal small bowel than in colon and rectum, and were very numerous at bases of gastric and duodenal peptic ulcers. Measurable IgE was found in seventy-eight of eighty-five (92%) intestinal fluids. Sucrose gradient ultracentrifugation analysis of four of the fluids revealed that the immunologically reactive IgE was largely in fractions corresponding to molecules of lower molecular weight than that of albumin, which suggests that IgE in gut contents is degraded by proteolytic enzymes. The presence of IgE-forming cells in gastrointestinal tissues, and IgE or a fragment of IgE in intestinal fluids, suggests that IgE antibodies are available for participation in local reaginic-type reactions in the gut.     

Repair of Immunoglobulin Response in B Cell Line (JK32.1) Originating from Immunodeficient Patient via Implantation of Functional Plasma Membranes

Human-human B cell hybridoma JK32.1, constructed from B lymphocytes of a common variable immunodeficient patient and nonsecreting cell line, retains the defects of B cell immunodeficiency. Efforts to clarify whether the defect is located within the plasma membranes of this cell line were carried out by implanting them with plasma membrane fraction derived from normal functional cells via intact noninfectious Sendai virus. The implanted cells were activated with various mitogens and their Ig responses and isotype switching were examined. Restoration of IgM secretion was achieved in the implanted JK32.1 cells following stimulation with SAC, PWM, or retinoic acid. Augmented IgM response was also obtained in the implanted cells treated with retinoic acid and lipopolysaccharide (LPS) despite their unresponsiveness to LPS alone. No IgG or IgA response could be detected in the implanted JK32.1 cells. These data suggest that this immunodeficient cell line possesses at least two different malfunctions, one located within the plasma membrane moiety of the cells and the other located within the cytoplasmic and/or nucleic components. The plasma membrane moiety defect can be repaired temporarily by delivering proper signals via the implanted plasma membranes. However, this manipulation of the cells could not overcome the intrinsic defect of the cells which blocks isotype switching and secretion of IgG, IgE, and IgA antibodies.     

Binding of human IgE immunoglobulin to rat mast cells. A study by means of three independent techniques.

The binding of human myeloma IgE immunoglobulin on rat mast cells was studied by three independent techniques. A mixed agglutination reaction with anti-IgE-coated Sephadex granules demonstrated that only human IgE-coated rat mast cells were clearly agglutinated. This binding is strong (50% agglutination) in 3 min and progresses for 30 min (95% agglutination). Autoradiographic studies with 125I-labelled human serum proteins demonstrated the selective formation of grains on mast cells incubated with labelled IgE. Upon action of anti-IgE antiserum on IgE-coated rat mast cells, the mast cells released up to 47.5% of their total histamine content in a fluorometric histamine assay. A relationship was established between sensitizing doses of human IgE and histamine release. These results bring evidence for a binding of human IgE on rat mast cells and imply the existence of receptors for this immunoglobulin on mast cell membrane.     

A study on the regulation of translocation of glucose transporters during hepatocarcinogenesis induced by 3'-Me DAB

The mechanism of glucose transported (GT) expression on the plasma membranes of hepatoma cells in rats induced by 3'-methyl-4-dimethylaminoazobenzene (3'-MeDAB) was studied. Cytochalasin B binding to plasma membrane fractions from control and 3'-MeDAB group in the absence of cold cytochalasin B showed 9,825 +/- 925 and 30,165 +/- 625 dpm/mg membrane protein. Scatchard plot analysis showed that the GTs present on the plasma membrane fractions in control and 3'-Me DAB groups were 5.0 and 16.0 pmol/mg membrane protein and their Kd values were 151 and 157 nM, respectively. These results suggest that the numbers of GTs in plasma membrane were increased in the 3'-Me DAB group compared to the control group. In contrast, the amounts of GTs in low density microsomal (LDM) fractions measured by a photoaffinity labeling technique using [3H]-cytochalasin B were 31,207 and 11,702 dpm/mg protein in the control and 3'-Me DAB group, respectively. These results suggest that GTs were translocated from LDM to plasma membranes during carcinogenesis. To confirm these results by an independent method 10% SDS-polyacrylamide gel electrophoresis was carried out. Gel slice No. 13 corresponding to MW of 45 kDa from plasma membrane fractions showed increased radioactivities in the 3'-Me DAB group compared to the control group. However, LDM fractions of the 3'-Me DAB group showed decreased radioactivities compared to the control group. Western blot analysis using anti-human RBC GT antibody present in the plasma membranes and LDM fractions from control and 3'-Me DAB groups did not show any significant difference, indicating low cross-reactivity between them. These results indicate that increased glucose transport seems to be more likely due to reciprocal redistribution of GTs between plasma membrane and LDM fractions.     

Dengue viral antigens in host cell membranes.

Membranes from type 2 dengue virus (DEN-2) infected BHK cells, when used as immunogen, elicit antibodies detectable by complement fixation, hemagglutination inhibition and plaque reduction neutralization tests with suckling-mouse-brain-derived DEN-antigens. Membrane fractions from infected cells separated by sucrose density gradient centrifugation were employed as antigen in serological tests with anti-DEN-2 infected suckling mouse brain hyperimmune mouse serum and hyperimmune mouse ascitic fluid. Antigen (s) capable of fixing complement were associated with all membrane fractions, including the plasma membrane, whereas antigen capable of reacting with antihemagglutinin was largely confined to the two membrane bands consisting of (1) a mixture of rough and smooth endoplasmic reticulum (2) predominantly rough endoplasmic reticulum.