Receptors for complement of leukocytes.

Sheep red blood cells sensitized by 7S, but not by 19S rabbit anti-Forssman antibodies, adhere and form rosettes on mouse macrophages and on a few monocytes and polymorphonuclear cells (PMN). When, however, C' factors from mouse serum are added to the antigen-19S antibody complex (EAC'), rosettes are formed on most mouse peritoneal macrophages and PMN and on a few monocytes. In addition EAC' also adheres to 10-25% of lymph node lymphocytes but not to thymus lymphocytes. EAC' prepared with 7S anti-Forssman antibodies has identical properties. The adherence of red cells induces an increase in the membrane activity of the leukocytes and causes injury to the red cells which rapidly become deformed and fragmented. Adherence of EAC' occurs at 37 degrees C and is minimal at 4 degrees C. Probably only the first four C' components are involved in this phenomenon as mouse serum deficient in C'5 or rabbit serum, deficient in C'6 can be used as a source of C' components. Treatment of EAC' with EDTA does not modify its leukocyte-adherence properties. The adherence of EAC' to the leukocytes is not inhibited in the presence of serum. The receptors for C' on macrophages, PMN, and monocytes differ from those found on lymphocytes. Rosette formation by EAC' on macrophages, PMN, and monocytes depends on divalent cations (Mg(++)) and can be reversed by Na(3)H EDTA, while adherence to lymphocytes is independent of these ions and occurs in the presence of 0.01 M Na(3)H EDTA. Both types of receptors for C' components are destroyed by trypsin treatment of the leukocytes, in contrast with the receptors for 7S antibodies on the same cells which persist after enzyme treatment.

The role of C4-binding protein and beta 1H in proteolysis of C4b and C3b.

Two forms of C4-binding protein (C4-bp) (C4-bp low, C4-bp high), which differ slightly in net charge and apparent molecular weight, as determined by SDS- PAGE, were separated by ion-exchange chromatography and contaminants removed with specific antisera. Both forms of C4-bp served as cofactors for the cleavage of C4b in solution by C3b inactivator, and the resulting fragments of the a'-chain of C4b had identical molecular weights. In addition, similarly to beta1H, C4-bp low or high served as cofactors for the cleavage of fluid phase C3b by C3bINA. However, important quantitative differences between the activities of C4-bp and beta1H were observed. With regard to C3b in solution, the cofactor activity of beta1H was {approximately equal to}20 times greater than that of C4-bp on a weight basis. In relation to cell-bound C3b, the differences in activity were even more marked. Whereas beta1H enhanced the effects of C3bINA on the erythrocyte intermediate EC3b, inhibiting the assembly of EC3bBb, C4-bp was without effect even at concentrations {approximately equal to}300 times greater than beta1H. Therefore, under physiological conditions, it is likely that beta1H is the key protein which controls the function of C3b, and that C4-bp activity is directed mainly toward the cleavage of C4b. We also examined the relation between C4-bp and the C3b-C4bINA cofactor described by Stroud and collaborators (3, 4). By functional, physico-chemical and immunological criteria, they are the same protein.

Complement receptor is an inhibitor of the complement cascade.

A glycoprotein from the membrane of human erythrocytes has been identified as a receptor for C3b (CR1). It promotes the dissociation of the alternative pathway C3 convertase C3b,Bb and the cleavage of C3b by C3b/C4b inactivator. We find that CR1 also inactivates the C3 and C5 convertases of the classical pathway. CR1 inhibits the consumption of C3 by C3 convertase EAC142 and enhances the decay of C4b,2a sites. On a weight basis, CR1 is approximately 5-10 times more active than C4 binding protein, a serum inhibitor of C4b,2a. The binding of 125I-CR1 to EAC14 cells is inhibited by C2. Therefore, it is likely that CR1 and C2 compete for a site on C4b. CR1 inhibited C5 convertase even more effectively, but had no effect on the assembly of the late complement components. At high concentrations, CR1 alone has no irreversible effects on cell-bound C4b. In the fluid phase, CR1 can function as a cofactor for the cleavage of the alpha' chain of C4b by C3b/C4b inactivator. A well-known function of CR1 is to promote adherence of microbes or immune complexes bearing C3b and C4b to cells. This interaction could result in a microenvironment damaging to the plasma membrane of the responding cell because the extrinsic C3b and C4b fragments can serve as additional sites of assembly of enzymes of the cascade. We therefore wish to propose that CR1 on the surface of cells supplies an increased local concentration of a strong inhibitor of the amplifying enzymes of the complement system and provides cells with a mechanism for circumventing damage when they bind C3b- and C4b-bearing substrates.

The role of membrane receptors for C3b and C3d in phagocytosis.

In this paper we re-examine the roles of particle-bound IgG and C3 in phagocytosis of sheep erythrocytes (E) by monolayers of purified human monocytes and polymorphonuclear leukocytes (PMN). We conclude that two fragments of the C3 molecule, that is, C3b and C3d, can function as opsonins if the phagocyte has the appropriate membrane receptors. Monocytes, that bind both C3b and C3d, respond to both as opsonins. PMN, which do not bind C3d, respond only to particles opsonized with C3b. C3 and IgG have separate roles in phagocytosis. IgG, through its Fc fragment, directly stimulates particle ingestion, but is relatively inefficient at inducing particle binding. On the other hand, C3 primarily mediates the binding of the particle via complement receptors. A marked synergy exists between C3 and IgG in inducing phagocytosis. Thus, opsonization of the particle with C3 can be a necessary condition for particle ingestion, although by itself C3 does not trigger phagocytosis. The opsonic effect of C3 can be mimicked by a variety of nonimmunologic agents which enhance binding of the particle to the phagocyte without directly stimulating ingestion. The contact-inducing agents used include centrifugation of particle and phagocyte, high molecular weight dextran, protamine, and treatment of E with neuraminidase. These results suggest that the role of C3 in opsonization is mainly or exclusively one of establishing contact between particle and phagocyte.

Control of the function of substrate-bound C4b-C3b by the complement receptor Cr1.

The complement fragments C3b and C4b are the main ligands for the membrane receptor CR1. We showed elsewhere that CR1 functions as an essential cofactor for the factor I-mediated enzymatic breakdown of membrane-bound C3b (*C3b) into C3c and * C3dg . One of the main findings of the present paper is that CR1 also promotes the degradation of bound C4b (*C4b) into C4c and *C4d. On a weight basis, the cofactor activity of CR1 in the cleavage of *C4b present on the cell intermediate EAC14 is 10(3)-fold greater than that of the serum cofactor C4-binding protein ( C4bp ). An additional finding is that the effect of CR1 on either *C3b or *C4b is modulated by the presence of the other ligand in its vicinity; that is, *C4b degradation by CR1 plus I is enhanced by neighboring *C3b and vice versa. For example, upon uptake of optimal amounts of *C3b onto EAC142 and the assembly of the C3-convertase EAC1423 , the activity of CR1 in generating C4c is enhanced 5-10 times further. Conversely, when the number of *C3b molecules on EAC1423 is relatively small (or when EAC1423 has been converted by I plus H into EAC1423i ), the presence of neighboring *C4b enhances the conversion of *C3b (or *iC3b) into C3c plus * C3dg . The enhancing effect of *C3b on the cleavage of *C4b by I is observed only if the cofactor of this reaction is CR1. Indeed, the activity of I or I plus C4bp on *C4b is significantly inhibited when *C3b is fixed and the main product of the reaction is * iC4b . Taken together, these findings suggest that degradation of *C4b will be more effective when enough C3b molecules are fixed nearby, thus facilitating the interaction of *C4b*3b clusters with CR1-bearing cells, and that under physiological conditions, *C4b activity can be efficiently controlled by CR1.

Studies on the mechanism of solubilization of immune precipitates by serum.

Antigen (Ag)-antibody (Ab) aggregates prepared with several different antigens are solubilized by fresh serum at 37 degrees C (complex-release activity of serum or CRA). The rate of solubilization varies in different systems and is strongly influenced by the affinity of Ab for the Ag in the immune precipitate. With a given Ag-Ab precipitate, the maximum amount of complex that can be solubilized by individual sera is independent of the initial concentration of complexes and cannot be increased by prolonged incubation. CRA occurs in the absence of C2 and C4, but not in the absence of C3 and factor B of the properdin pathway. Addition of C2 to C2-deficient serum or C4 to C4-deficient serum enhances CRA. Solubilization does not involve extensive degradation of the complexed antibody, as might be detected by acrylamide gel electrophoresis of released antibody after reduction and alkylation to separate H and L chains. Immune precipitates can also be solubilized by incubation with monovalent fragments (Fab or Fab') of antibodies against determinants of the Ab molecules in the immune precipitate. In contrast, F(ab')2 fragments decrease the solubility of the immune precipitates. In view of these findings, we propose that CRA is mediated by the binding of functionally monovalent C fragments (C3 and C4) onto Ab molecules in the precipitates.

Genetic linkage between serum levels of the third component of complement and the H-2 complex.

AKR/J (H-2kk) mice have higher serum C3 levels than DBA/2J (H-2dd). The F1 hybrids have intermediate levels. Analysis of the progeny of backcrosses at 21 days of age shows that C3 levels in mice of H-2dk type are significantly higher than those with H-2dd type and lower than those with H-2kk type. In addition, mice of H-2kk, H-2dk, and H-2dd types have C3 levels not significantly different from those of AKR/J, AKD2F1, and DBA/2J respectively. These findings demonstrate linkage between a gene controlling C3 levels and the H-2 complex.

Antihapten antibody specificity and L chain type.

Two types of light polypeptide chain (kappa and lambda) are present in guinea pig immunoglobulins. The ratios of K/L molecules in anti-hapten antibodies differ sometimes markedly from that found in normal serum. Anti-DNP antibodies and anti-pipsyl antibodies have, respectively, a higher and a lower than normal K/L ratio. The possibility that the L chain type affects the range of configurations which the antibody-combining site may assume was investigated. Fractional precipitation of anti-DNP antibodies from serum of guinea pigs immunized with DNP(44)-BSA was performed using limiting amounts of antigen. Antibody fractions were purified from each precipitate, their affinities for epsilon-DNP-L-lysine measured by fluorescence quenching (K(0)) and the K/L ratio estimated by precipitation with specific antisera. Increasing concentrations of L molecules were found in fractions with decreasing K(0). In other experiments, fractional precipitation and purification of antibodies which cross-react with DNP was performed in serum of animals immunized with pipsyl-BGG. The K/L ratio in antibodies isolated from these fractions was much higher than in fractions which do not cross-react with DNP. These results show that K molecules are better adapted to react with the DNP hapten than L molecules. The K/L ratio in anti-DNP antibody was shown to increase in the course of immunization, at the same time that the K(0) is increasing. This rise in K(0) is markedly delayed when larger doses of antigen are employed.

Quantitative variations in L chain types in guinea pig antihapten antibodies.

In guinea pig purified antihapten antibodies, the proportion of molecules bearing the kappa- or lambda-type of L chains (K or L molecules) may diverge markedly from that found in normal gamma(2)-globulins. This has been evaluated by precipitation of I(131)-labeled antibody preparations using a specific anti-lambda-chain antiserum. Anti-DNP antibodies isolated 3 wk after immunization of guinea pigs with DNP(65)-BGG antibodies, contain less than 1% of L molecules, while in pipsyl antibodies, isolated from the sera of animals immunized with pipsyl-BGG, the proportion of L molecules is significantly greater than in normal gamma(2)-globulins. Anti-DNP antibodies produced against conjugates of this hapten with carriers other than BGG (BSA, OVA, or poly-L-lysine) or with BGG with a small number of DNP groups (DNP(10)-BGG) contained a greater proportion of lambda-chain bearing molecules than anti-DNP antibodies isolated from late sera of guinea pigs immunized with highly conjugated DNP(65)-BGG. An increased percentage of L molecules was detected in preparations of anti-DNP(65)-BGG antibodies isolated early (10 days), when compared to those isolated later during the course of immunization. However, the level of L molecules in all these anti-DNP antibody preparations was always considerably below that present in normal gamma(2)-globulin. The relative amounts of L molecules in distinct immunoglobulin families (gamma(1) and gamma(2)) in antibody preparations isolated from individual animals was remarkably similar.

Amastigotes of Trypanosoma cruzi escape destruction by the terminal complement components.

We studied the effect of complement on two life cycle stages of the protozoan parasite Trypanosoma cruzi: epimastigotes, found in the insect vector, and amastigotes, found in the mammalian host. We found that while both stages activate vigorously the alternative pathway, only epimastigotes are destroyed. The amounts of C3 and C5b-7 deposited on the amastigotes were similar to those bound to the much larger epimastigotes. Binding of C9 to amastigotes was four to six times less than binding to epimastigotes, resulting in a lower C9/C5b-7 ratio. Although a fairly large amount of C9 bound stably to amastigotes, no functional channels were formed as measured by release of incorporated 86Rb. The bound C9 had the characteristic properties of poly-C9, that is, it expressed a neo-antigen unique to poly-C9, and migrated in SDS-PAGE with an apparent Mr greater than 10(5). The poly-C9 was removed from the surface of amastigotes by treatment with trypsin, indicating that it was not inserted in the lipid bilayer. Modification of amastigote surface by pronase treatment rendered the parasites susceptible to complement attack. These results suggest that amastigotes have a surface protein that binds to the C5b-9 complex and inhibits membrane insertion, thus protecting the parasites from complement-mediated lysis.

Inhibition of complement activation on the surface of cells after incorporation of decay-accelerating factor (DAF) into their membranes.

Decay-accelerating factor (DAF), extracted from the stroma of human erythrocytes, was purified to homogeneity and incorporated into the membrane of sheep red cell complement intermediates, where its functional properties were analyzed. Incorporation of DAF into the cell membranes was temperature dependent, took place on pronase- or trypsin-treated erythrocytes, and did not depend on prior deposition of antibody, C1 or C4. Serum lipoproteins (high and low density) effectively inhibited DAF incorporation, but had no effect on the activity of DAF after its association with the cell membrane. The incorporated DAF could not be removed from the red cell surface by repeated washings in the presence of high salt concentration but was solubilized when the stroma were extracted with 0.1% Nonidet P-40. The presence of DAF in the membrane of EA did not affect the deposition of C1 and C4, but as few as 10(2) DAF molecules per cell profoundly inhibited the assembly of C3 and C5 convertases of both the classical and alternative pathways. The DAF inhibitory effect on EAC14 or EAC43 was not overcome by supplying an excess of C2 or factor B, but the alternative pathway C3 convertase could be assembled in the presence of Ni++, or nonphysiological concentrations of Mg++, which enhances the binding affinity of factor B for C3b. The DAF effect on EAC14 or EAC143 was entirely reversed by treating the cells with specific anti-DAF antibodies, showing that DAF did not alter the structure of C4b or C3b. Taken together, the experimental evidence suggests that DAF interacts directly with membrane-bound C3b or C4b and prevents subsequent uptake of C2 and factor B. DAF can function only within the cell membrane. Indeed, the decay dissociation of the C4b2a enzyme on DAF-containing sheep intermediates was not changed by varying the cell concentration. DAF-treated EA had no influence on the decay of nontreated EAC142 present in the same mixture. Moreover, the inhibitory activity of intact human erythrocytes on C4b2a was not blocked by antibodies to DAF, but was abolished by antibodies to the C3b/C4b receptor (CR1). When incorporated into the membrane of rabbit erythrocytes, human DAF inhibited their lysis by human complement. In conclusion, on the basis of these and previous results, it appears that DAF plays a central role in preventing the amplification of the complement cascade on host cell surfaces.(ABSTRACT TRUNCATED AT 400 WORDS)

Trypanosoma cruzi trans-sialidase and neuraminidase activities can be mediated by the same enzymes.

Trans-sialidase and neuraminidase activities have been detected on the surface membrane of trypomastigotes of Trypanosoma cruzi, and both have been implicated in the parasite's invasion of host cells. We show here that these enzymes are structurally related. They are recognized by two independently derived monoclonal antibodies, are anchored to the membrane by glycosylphosphatidylinositol, copurify by ion exchange, molecular sieving, and hydrophobic chromatography, have maximal activities between pH 6.5 and 7.5, and are inactivated by heating at 56 degrees C. Furthermore, the neuraminidase and trans-sialidase reactions are coupled. An increase of the concentration of acceptors of the transfer reaction decreases the amount of free sialic acid released through the neuraminidase reaction. We conclude that a single enzyme can catalyze the transfer or the hydrolysis of macromolecular-bound sialic acid. The predominant direction of the reaction will depend on the availability of appropriate oligosaccharide acceptors of sialic acid.

The main lytic factor of Trypanosoma brucei brucei in normal human serum is not high density lipoprotein.

Natural immunity of humans to the cattle pathogen Trypanosoma brucei brucei has been attributed to the presence in normal human serum (NHS) of lytic factors for the parasites. We and others have shown that NHS contains two trypanolytic factors (herein termed TLF1 and TLF2) that can be separated by gel filtration. TLF1 copurifies with a subclass of high density lipoprotein (HDL), whereas TLF2 has a much higher molecular weight and does not appear to be a lipoprotein. We find that the trypanolytic activity of purified TLF1 is totally inhibited by exogenous haptoglobin (Hp) at concentrations (0.1 mg/ml) lower than those present in NHS (0.2-2 mg/ml). In contrast, exogenous Hp (up to 2.5 mg/ml) has no effect on the lytic activity of either NHS or isolated TLF2. Hp-depleted sera from patients with intravascular hemolysis is severalfold more trypanolytic than NHS. These sera contain only TLF1, and their lytic activity is totally abolished upon the addition of Hp (0.1 mg/ml). When NHS containing different Hp allotypes is fractionated by gel filtration, TLF1 activity is either revealed or remains masked, depending on whether it coelutes with Hp. Masked TLF1 activity in the column fractions is revealed if Hp is removed by density gradient ultracentrifugation. We conclude that endogenous Hp inhibits TLF1 activity, and that TLF2 is the main trypanolytic factor in NHS.

Human C4-binding protein. I. Isolation and characterization.

C4-binding protein (C4-bp), a new component of the complement system, was isolated from human plasma by precipitation with polyethyleneglycol, followed by chromatography on ion exchangers. C4-bp was identified on sodium dodecyl- sulfate polyacrylamide gel electrophoresis (SDS-PAGE) by two independent criteria: its ability to bind to C4b, and immunoprecipitation with a monospecificantiserum. Purified C4-bp is a 10.7 s glycoprotein. It consists of several disulfide bonded subunits of mol wt 70,000 daltons. Under nonreducing conditions, its mol wt has been estimated on SDS-PAGE as 540- 590,000 daltons. C4-bp moves as a slow B-globulin at pH 8.6 in the absence of free divalent cations, but when the buffers contain Ca(++)-lactate, C4-bp is a gamma globulin. Purified C4-bp binds to purified C4b. The reaction proceeds in the presence or absence of divalent cations and is not inhibited by diisopropylfluorophosphate. The C4b/C4-bp complexes have sedimentation coefficients between 15 and 17 s on sucrose gradient ultracentrifugation, and can be readily identified by crossed immunoelectrophoresis (CIE). The complexes move faster toward the anode than either protein. C4-bp is multivalent. Saturation is reached at molecular ratios of C4b/C4- bp of between 4 and 5. The interaction between C4b and C4-bp may complicate the electrophoretic patterns of these proteins in normal human serum, if the complement system is activated before or during the run. However, in EDTA-plasma, native C4 and C4-bp do not form stable complexes and can be identified in separate peaks after CIE.

Human C4-binding protein. II. Role in proteolysis of C4b by C3b-inactivator.

We recently described the isolation from human serum of a high molecular weight protein with specific binding affinity for fluid-phase activated C4. We show here that the C4-binding protein (C4-Bp) functions as an essential cofactor in the proteolysis of C4b in the presence of C3b-inactivator (C3bINA). C4-bp, together with C3bINA, cleave the alpha'-chain of C4b into three fragments called alpha2, alpha3, and alpha4, with mol wt of 47,000, 25,000, and 17,000 daltons, respectively. The alpha2 fragment was dissociated from C4b without reduction, whereas the alpha3 and alpha4 fragments were disulfide bonded the other chains of C4b. The reaction did not occur when either C4-bp or C3bINA were omitted, nor in the presence of either protein in combination with beta1H. Native C4 was not affected by C3bINA aand C4-bp. C4b was not cleaved when incubated in serum of a patient with genetic deficiency of C3bINA. However, when purified C3bINA was added, the alpha'-chain of C4b was cleaved and fragments with the same molecular weight as alpha2, alpha3, and alpha4 were generated.

Structural and functional differences between the H-2 controlled Ss and Slp proteins.

Based on functional and structural data, it is concluded that the Ss protein in the mouse expresses the activity of the fourth component of complement. Removal of the Ss, but not of Slp, antigen correlates with a high degree of significance (P less than 0.001) with decrease of C4 hemolytic activity. In phenotypically Slp negative mice the plasma/serum levels of Ss correlate with the C4 activity (P less than 0.001). Structurally, Ss is a 209,000-mol wt protein, consisting of three covalently linked polypeptide chains (alpha,beta,gamma). Treatment of Ss with C1 cleaves a 7,000-8,000-mol wt fragment from the alpha-chain. Slp is also a three chain covalently linked protein of 209,000 daltons, however its three chains differ in size from those of the Ss protein. Slp does not express hemolytic activity and its alpha-chain is not cleaved by C1.

Purificaiton and characterization of mouse serum protein with specific binding affinity for C4 (Ss protein).

A new component of the complement (C).system, with a specific binding affinity for the activated Ss-protein (C4) has been identified in mouse serum. This protein, named Ss- (or C4)-binding protein (Ss-bp), was purified about 200 times from mouse plasma. Ss-bp is a heat stable (56 degrees C, 60 rain) beta-globulin with a sedimentation coefficient in sucrose density ultracentrifugation of 10s. Its concentration in serum of adult male and female mice is 160 and 60 mug/ml, respectively. In EDTA-plasma, Ss and Ss-bp are not associated and can be separated by chromatography in Sephadex G-200. However, in serum Ss-bp binds tightly to Ss. The bonds between these proteins cannot be reversed by chelation of divalent cations. As a consequence of the formation of Ss/Ss-bp complexes, the properties of Ss-bp appear to be quite different in serum of mice with high (Ss-H) or low (Ss-L) levels of Ss-protein. In Ss-H serum, all of Ss- bp is bound to Ss. In Ss-L serum, Ss-bp is mostly free. Because the electrophoretic mobilities of free and complexed Ss-bp are quite different, Ss-bp appears to be polymorphic in serum (but not in EDTA- plasma). The strict dependency of the apparent electrophoretic mobility of Ss-bp on the levels of Ss in serum was demonstrated in a series of congenic mice and among the progeny of a cross between Ss-H and Ss-L strains of mice. Without exception, the slow and fast varieties of Ss-bp were associated with the Ss-L and Ss-H traits. Ss-bp of the slow variety can be transformed into the fast variety by addition of pure human C4, or C4-sufficient guinea pig serum, to Ss-L serum. In both instances Ss-bp formed stable complexes with C4 or a C4- derived peptide. These findings highlight the binding specificity of Ss- bp for the fourth component of the complement system, and in addition they demonstrate a functional homology between the Ss-protein and C4 from two different species.

Requirements for the solubilization of immune aggregates by complement: assembly of a factor B-dependent C3-convertase on the immune complexes.

During the solubilization of immune precipitates BSA-rabbit antibodies to BSA by human complement, at least three stages can be distinguished. (A) Generation of alternative pathway C3-convertase sites associated with the immune complexes. During the first minutes of interaction between the immune aggregates and serum, before any solubilization has taken place, properdin (P), factor B, and C3 moieties are incorporated into the lattice. The washed precipitates have C3-convertase activity, which can be completely inhibited by antibodies to factor B, but not to C2. The assembly of the convertase is temperature-dependent, and does not take place in the absence of Mg++. The immune complex-associated C3-convertase activity decays rapidly at 37 degrees C, but it can be restored by addition of purified factor B and properdin. (B) Amplification. When the aggregates bearing C3-convertase are incubated with purified C3, solubilization takes place. It appears that solubilization is caused by the accumulation of a large number of C3 fragments on the Ag-Ab lattice. In solubilized complexes, the molar ratios of Ab/C3 are close to one. (C) Spontaneous release. The final step in the solubilization process is a secondary reaction, during which some rearrangement of the lattice takes place. It occurs in medium devoid of serum and does not require divalent cations.

Identification of the membrane receptor for the complement fragment C3d by means of a monoclonal antibody.

The B2 antigen characterized by means of a monoclonal antibody (14) is a 140,000 Mr protein expressed only in certain stages of the differentiation of lymphocytes of the B lineage. Here we examine the relationship between B2 and the membrane complement receptor type 2 (CR2) for the complement fragment C3d (11, 12), which is also associated only with B cells. Both phenotypic markers are distributed in a similar manner among B cell malignancies and, as shown here, among established cell lines. A polypeptide with binding affinity for C3d was isolated from the membrane of B2-positive cells, i.e., tonsil lymphocytes and Raji cells. We found that this C3d-binding protein not only had the same Mr and isoelectric point (pI) as the B2 antigen, but that it was recognized by the monoclonal antibody to B2. However, anti-B2 does not mask the ligand-binding site of CR2 since it does not prevent the interaction of the purified 140,000 Mr polypeptide with immobilized C3d. Rosette formation between tonsil lymphocytes and erythrocyte intermediates bearing C3d was specifically inhibited by anti-B2. In the case of Raji cells, rosette formation was strongly inhibited only when the lymphocytes were sequentially treated with anti-B2 and with a polyclonal antibody against mouse Ig. In short, B2 and CR2 have a similar distribution among normal and malignant cells, have the same Mr and pI under denaturing conditions, and react with a single monoclonal antibody. We conclude that B2 is identical to CR2.

Release of decay-accelerating factor (DAF) from the cell membrane by phosphatidylinositol-specific phospholipase C (PIPLC). Selective modification of a complement regulatory protein.

Decay-accelerating factor (DAF) is a 70,000 Mr membrane protein that inhibits amplification of the complement cascade on the cell surface, and protects cells from damage. Purified DAF can be reincorporated into the membrane of red cells and is functional. DAF is deficient in paroxysmal nocturnal hemoglobinuria (PNH), a disease characterized by increased sensitivity of erythrocytes to complement lysis. We show here that DAF is part of a newly described family of membrane proteins anchored to the lipid bilayer by means of phosphatidylinositol (PI). Treatment with PI-specific phospholipase C (PIPLC) releases 70-80, 60, and 10% of cell surface DAF from mononuclear cells, neutrophils, and erythrocytes, respectively. The PIPLC-released DAF (DAF-S) is slightly smaller (67,000 Mr) than the membrane form. DAF and DAF-S cannot be distinguished antigenically. Furthermore, DAF-S has lost its ability to significantly inhibit the C3-convertase, as well as its ability to incorporate into cell membranes. Since DAF can only inhibit C3-convertase endogenously, i.e., within the membrane of the same cell, it is likely that the loss of activity of DAF-S is causally related to its inability to reincorporate in the lipid bilayer. As shown by others, the complement-sensitive red cells from PNH patients lack acetylcholinesterase, which is also anchored to the membrane by PI (9). Thus it is possible that the molecular defect in PNH lies in the biosynthetic pathways leading to the attachment of PI to the polypeptide chains, in the transport of these proteins to the surface, or in their release by the action of endogenous phospholipases. From a practical standpoint the specific release of DAF by PIPLC could facilitate killing of tumor cells by amplifying the effects of the complement cascade on the surface of antibody-sensitized cells.