Search

The authors studied the appearance and behavior of complement-fixing anti-ROS antibodies in the sera of GrCh rabbits following one or two intracutaneous injections of retinal antigens emulsified in Freund's complete adjuvant (KFA). Fourteen different antigenic preparations from heterologous, homologous, and autologous retinal components incorporated in KFA were tested and compared. Apart from the supernatant of a centrifuged homogenized preparation of homologous retina, all immunogens generally induced anti-ROS antibodies. The antibody titers (reciprocal values) reached a maximum 4 to 8 weeks post injection, at only low levels. The changes in antibody concentrations showed some relationship to the quality and dose of the immunogens. Anamnestic effects following reinjection reached maximum values within as little as 5 to 14 days. Animals not developing chorioretinitic lesions showed higher antibody titers more frequently than rabbits with clinically confirmed disease. Provided there is sufficient relevance of the humoral reactions observed, the complement-fixing antibodies we found in our experiments are more likely to have protective than pathogenic functions.

Proteins antigenically related to CoF, the anticomplementary protein of the venom of the Indian hooded cobra (Naja naja), were found in a variety of elapid and viper venoms but not in the venom of Brazilian crotalids . In keeping with this finding was the weak ability of Brazilian snake venom to convert C3 in human serum. All snake serums tested, including Brazilian crotalids , contained a beta-globulin antigenically related to CoF. This serum protein in Brazilian snake serum had a number of characteristics of mammalian C3, including conversion on storage or on incubation of this serum with endotoxin, zymosan or mammalian antigen-antibody precipitates. The serum protein did not, however, convert on incubation with hydrazine. Brazilian crotalid serum did not, as did cobra serum, have the ability to inactivate CoF's ability to activate complement in normal human serum. The crotalid serum had hemolytic activity for rabbit antibody-sensitized and unsensitized sheep red blood cells that was active in the presence of Ca++ and Mg++ or Mg++ alone but greater with Ca++ present, suggesting the presence of both classical and alternative pathways of complement activation. This activity was maximal at 37 degrees C, but was destroyed or inactive after heating at 50 degrees C for 1 hr, incubation with hydrazine or by addition of EDTA. A marked reduction of hemolytic activity of Bothrops serum occurred after removal of the CoF-like protein. These findings suggest that Brazilian snake venom has little CoF-like material, but its serum contains a CoF-like protein with many characteristics of C3.(ABSTRACT TRUNCATED AT 250 WORDS)

25 Sera from children with the clinical diagnosis: urticaria, allergic-toxic exanthema were examined for the complement components C3, C4 and C3-Activator. We applied the radial immunodiddusion. As controls served 50 helathy children and 25 children with morbili, rubeolae and scarlet fever. C3 was found to be decreased in 23 cases, C4 in 4 and C3-Activator in 19 cases of urticaria. This indicates the possiblity for the differential diagnosis, but children of the control groups did not show any consumption of one of the described complement components.

An immunoadsorbent-antibody (egg albumin-Sepharose-antibody) column was found to be suitable for the rapid separation of C1-esterase from normal human serum. About 1.54 mg of C1-esterase, with a specific activity of 447 units/mg was obtained in voer 80% yield from the 20 ml of human serum.

Complexes formed of Cobra venom factor (CVF) and activated factor B (B) by interaction of CVF and B with trypsin or factor D are capable of activating the third and fifth complement component. When incubated with sheep or guinea pig red cells and guinea pig serum in the presence of EDTA, these CVFB complexes produce "indirect lysis". Addition of a human serum factor, earlier designated as factor E (6), greatly enhances the efficiency of this lytic system. The component with this activity has been purified to homogeneity (disc and immunoelectrophoresis). In chromatographic fractionations it was inseparable from the fifth complement component, it was inactivated by and reacted with several anti-C5 antisera, and kinetics of inactivation by heat (56 degrees C) and trypsin were the same for factor E and hemolytic C5 activities. It is concluded that factor E is the fifth component of human complement. Guinea pig C5 is not capable of supporting indirect lysis in a comparable manner, for as yet unknown reasons. Some possible explanations are discussed.

A study of lymphocytes bearing C3b, C4b and C3d complement receptor (CRL) was performed on human peripheral blood from 16 healthy donors and 12 patients affected with Chronic Lymphatic Leukemia (CLL). In the latter group a clear rise of C3dCRL was demonstrated, when compared with immunoadherence receptors bearing lymphocytes. However, when compared with controls, also these latter were slightly augmented. Furthermore in CLL under treatment CRL populations showed the same profile as CLL: so it was suggested that the treatment reduced not selectively all the three types of CRL, within the population of sIg bearing lymphocytes. Here we discuss the hypothesis that, in CLL, the proliferating lymphocytes population is chiefly sIg+, C3d+/C3b-, C4b-, but also sIg+, C3d+/C3b+, C4b+.

Sixty-seven patients with systemic lupus erythematosus (SLE) were followed up for 3-19 months (mean 12) in a prospective study. The activity of SLE was estimated on clinical grounds and correlated with DNA antibody and complement levels. The disease reactivations consisted mostly of articular and cutaneous symptoms. There were 17 relapses and 22 complicating infections during the follow-up period. The levels of antibodies to native, double-stranded (ds) DNA (P less than 0.001) and antibodies to denatured, single-stranded (ss) DNA of IgG class (P less than 0.001) and C3 (P less than 0.001) correlated best with disease activity, which was estimated on the clinical symptoms and signs. These assays were not reliable, however, in predicting minor exacerbations. The levels of IgM class ss-DNA antibodies were significantly higher in SLE patients without nephritis than in SLE nephritis patients. In most cases, the combination of IgG class ss-DNA antibody and complement (C3 and CH50) determinations differentiated SLE relapse from infection.

Monoclonal antibodies obtained after immunization of mice with human C5 were screened for activity against native C5 and SC5b-9. Clone 568 reacted with the C5 beta-chain in western blotting and with purified native C5 in double antibody enzyme immunoassays. The reactivity of clone 568 against serum decreased markedly when serum was activated with zymosan. The residual activity was found to be against free C5, whereas the antibody did not react at all with the SC5b-9 complex. Thus, the antibody 568 reacts with an activation-dependent C5 epitope, which is concealed when C5 is incorporated into the terminal complement complex.

A factor chemotactic for Walker carcinosarcoma and Novikoff hepatoma tumor cells can be generated by incubation of serum with an extract from tumor cells. The chemotactic factor has no activity for neutrophilic leukocytes, and three factors chemotactic for leukocytes are not chemotactic for tumor cells. By the use of complement deficient serums as well as purified complement components, the chemotactic factor for tumor cells has been found to be a fragment of the fifth component (C5) of complement and appears to result from direct interaction of C5 with an enzyme in the extract. The chemotactic factor for tumor cells can be classified as a functionally unique fragment of C5 that may significantly influence the behavior of tumor cells in vivo.

Precipitation of the euglobulin fraction from serum separates C1q from another hydroxyproline-containing protein(s), which remains in the serum-minus-euglobulin fraction. The presence of C1q in the euglobulin fraction has been verified by radial immunodiffusion assay, by sensitivity to collagenase, by hydroxyproline content, and by electrofocusing. The presence of another hydroxyproline-containing protein(s) in the serum-minus-euglobulin fraction has been verified by the same criteria. The isoelectric pH range of C1q is 6.1-7.0; the isoelectric pH range of the other hydroxyproline-containing protein(s) is 4.2-5.5.

Isolation of the C5b-9 complex from inulin-activated whole human serum was effected by molecular sieve column chromatography employing Biogel A-15 M, preparative Pevikon block electrophoresis, and removal of low density beta-lipoproteins by flotation in CsCl. The final product was homogeneous upon cellulose acetate strip electrophoresis and analytical ultracentrifugation. Ouchterlony analyses indicated that the complex reacted with antisera to C5, C6, C7, C8, and C9 to form a continuous, circular precipitin line without spurs. The C5b-9 complex was dissociated by sodium dodecyl sulfate (SDS) in the absence of reducing agents, and analytical SDS-polyacrylamide gel electrophoresis revealed seven protein bands after straining with Coomassie Blue. Bands 1, 2, 3, and 6 were identified as C5b, C7, C6, and C9, respectively. Bands 4 and 7 were identified as two noncovalently bound subunits of C8. Molar ratios among C5b, C6, C7, C8, and C9 dissociated from the complex by SDS were estimated to be 1:1:1:1:3. Band 5 protein, which had an estimated mol wt of 88,000 and was found to occur with a molar ratio of 3, has not yet been identified. Its nature and possible biological functions are discussed.

The interaction between the complement components in human serum and the dye, Cibacron Blue F3GA, immobilized on cross-linked agarose (Affi-Gel Blue) has been studied. All nine components of the classical complement pathway bound to the dye and could be recovered using a linear salt gradient. With the exception of C5 and C8, all the components were eluted over a narrow NaCl concentration range, with the following yields: C1, 17%; C2, 69%; C3, 92%; C4, 87%; C6, 105%; C7, 109%; C9, 128%. C5 and C8 eluted throughout the NaCl gradient with yields of 103% and 14%, respectively. Since all components could be eluted without substantial contamination by albumin or IgG, this procedure may prove valuable as an initial step in the purification of complement components. In addition, the ability of immobilized Cibacron Blue F3GA to physicallly remove complement components may prove useful for both the decomplementation of serum and in elucidating the role of complement in immunological reactions.

The C5b-9 complex derived from human serum and assembled on target sheep erythrocyte membranes is a thin-walled cylinder rimmed by an annulus at one end. The total height of the cylinder is 150 A, towards which the annulus contributes 30 A. The cylinder has an apparently uniform internal diameter of 100 A. The external diameter of the annulus is 200 A. The classical complement 'rings' visualized on membranes after complement lysis represent such C5b-9 cylinders perpendicularly oriented on the membranes. The thin-walled cylinder is anchored in the membrane matrix and the annulus located in the exterior membrane glycocalyx. At the sites of attachment of the C5b-9 complexes, the continuity of the membrane bilayer is disturbed and the presence of trans-membrane pores is indicated. The data essentially support the 'doughnut' theory of complement lysis.

Affinity chromatography of serum on IgG covalently linked to Sepharose results in the retention of the proteins of the first component of complement (C1). A fraction called pool II is eluted from this column with 0.025 M EDTA and has previously been shown to contain C1s and a novel protein believed to be part of the C1 complex and called C1t. C1r has now been located in pool II and these three proteins were purified by DEAE cellulose chromatography. C1r and C1s were recovered in the proenzyme form and their identity was established by SDS polyacrylamide gel electrophoresis before and after reduction and alkylation, and on the basis of their esterolytic activities toward different substrates. The properties of C1r from pool II are contrasted with those of the protein recovered from the pool III eluate of the affinity column and previously thought to be C1r.

The fate of highly purified properdin (P) upon introduction into normal human serum or properdin-depleted serum (RP) was investigated. It was observed that, concomitant with the activation of the alternate pathway components, properdin underwent immunochemical alterations characterized by a shift in mobility from gamma2 to beta2 position and by an increase in the sedimentation rate from 5.1S to between 6.8 and 9.3S. The immunoelectrophoretic behavior of C3 was also altered with the appearance of a beta2 arc in addition to the beta1C arc. The immunochemical properties of altered P resemble those of "native" properdin in fresh serum. The principle in serum (designated factor F) mediating these changes is a euglobulin with an approximate sedimentation rate and molecular weight of 9.0S and 250,000 daltons, respectively. The alteration in the immunochemical properties of P may be due to aggregation of P molecules or a complex formation between P and a serum euglobulin (probably C3) mediated by factor F and it is associated with loss of ability of P in initiate the alternate pathway of complement activation upon interaction with serum.

A method is described for the preparation of C3d from fresh-frozen CPD plasma. A redissolved EDTA euglobulin precipitate formed from defibrinated plasma is chromatographed on DEAE cellulose. The C3-rich peak is further chromatographed by stepwise elution from hydroxy-apatite. The highly purified C3 and C3b so obtained are then treated with commercial C3b inactivator. G-200 sephadex filtration of this material produces a low molecular weight peak containing C3d greater than 95 per cent purity. The purified C3d, which contains negligible carbohydrate, has a molecular weight of 28,000 (based on SDS polyacrylamide gel electrophoresis). When added at a final concentration of less than 70 mug per ml, it completely inhibits the anti-C3d hemagglutinin reactivity of a potent anti-C3d antiglobulin serum but does not inhibit sera with reactivities against C3c, C4b, and C5b. Two of three rabbits hyperimmunized with purified C3d produced antisera exhibiting only anti-C3d precipitin activity. After absorption of heterologous antibodies, the antisera strongly agglutinated C3d-coated RBC, failed to agglutinate RBC glutinated C3d-coated RBC, failed to agglutinate RBC coated only with C4, and reacted weakly against strongly anti-Rh coated RBC. The latter reactivity was readily absorbed by whole IgG. Preliminary studies with 125I-labeled C3d indicate its potential value in assessing potency of anti-C3d reagents.

Of 33 children with membranous nephropathy screened for HBs Ag, 14 were found to be HBs Ag carriers, whereas HBs Ag was detected in 3 of 170 and 4 of 100 children with glomerular and nonglomerular kidney diseases, respectively. HBs Ag was often associated with acute hepatitis at onset (five patients) or with elevated transminases values. This high incidence and the prevalence of an unusual subtype (ayw2) suggest a relationship between HBs Ag and the glomerular lesions. Using immunofluorescence, however, HBs Ag could not be detected within the deposits, so that the nature of the relationship cannot be considered as established. The clinical outcome (50% remission), the plasma complement component disturbances, and findings by immunofluorescence did not differ from those observed in children with MGN without detectable HBs Ag.

Procedures for the isolation of the human complement proteins C6 and C7 have been described. These procedures allow isolation of the two proteins without any loss of hemolytic activity. Apparent activity gains of 160% and 140% were observed for C6 and C7, respectively, when the activity of the isolated proteins was compared with their activity in serum. The recovery of C6 was 3.5 to 11% and that of C7 was 7 to 13% of the amount present in serum. C6 has a m.w.of 128,000 and an electrophoretic mobility at pH 8.6 of -2.6 times 10(-5) cm2 s-1 v-1. C7 has a m.w. of 121,000 and an identical electrophoretic mobility. With 3 times 10(7) assay cells, 63% hemolysis was achieved with 1 ng of C6 and 3.8 ng C7. On polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate and after reduction with mercaptoethanol, C6 and C7 behaved as single polypeptide chain proteins.

The phenotypes of C3 and of Tf were determined in 818 and 576, respectively, unrelated individuals living in Liege. The gene frequencies observed are: (formula: see text) The application to disputed paternity cases and to the study of twins is discussed.

In a patient with lifelong increased susceptibility to infection and multiple abnormalities in complement-mediated functions, the infusion of normal plasma had been seen to produce a prolonged partial correction of serum abnormalities. It was subsequently shown that the patient was genetically deficient in the C3b inactivator and that immunochemical depletion of C3b inactivator from normal serum resulted in abnormalities similar to those found in the patient's serum, including alternative pathway C3 activation. Highly purified C3b inactivator was obtained from the euglobulin fraction of normal human serum, sterilized by filtration, and infused intravenously. Partial or complete correction of almost all the known serum abnormalities was obtained. C3b almost disappeared from the serum within 4-5 h, as did Factor C activity. Native C3, C5, and serum hemolytic activity rose to normal or near-normal levels over 4 days and were sustained for another week. Factor B, properdin, opsonic activity, and bactericidal activity reached a level at least two-five times that found before the infusion within 24 h and fell over the next 5 days. These observations prove the primary role of C3b inactivator deficiency in the patient's disease and demonstrate clearly the curcial role in vivo of C3b inactivator in modulating alternative pathway activity.

Localization of fluorescein-conjugated heat-aggregated IgG (FA IgG) was demonstrated by immunofluorescence in renal glomeruli of 16 of 69 patients with glomerulonephritis. FA IgG bound more frequently in kidney biopsies from patients with diffuse glomerulonephritis and depressed renal function, and localized selectively in glomeruli that contained heavy deposits of IgM, C3, and C4. The factors that caused FA IgG to bind were specifically reactive with the Fc piece of the IgG molecule and were resistant to 56 degrees C heat for 30 minutes. Localization of FA IgG in the kidney did not correlate with the presence of soluble immune complexes or detectable antiglobulin antibodies in the sera. Binding of FA IgG was also seen in glomeruli and arteries of 18 of 21 kidney allografts studied at the time of impending rejection. But the factors responsible for binding FA IgG in the allografts were heat labile and thus could have been C1q. Although the role of these "antiglobulins" in the immunobiology of glomerulonephritis remains unknown, the fact that they occurred mainly in patients with relatively severe glomerular injury suggests that they could play some part in promoting renal glomerular injury.

Most of the C1 material of euglobulin was adsorbed to heparin-Sepharose at an ionic strength of 0.265. After desorbtion at an ionic strength of 0.415 the C1 material was found to be purified six to seven-fold. Highly purified subcomponents C1q, C1r and C1s were recovered at DEAE-Sephadex chromatography from such purified C1 material after EDTA-treatment. Tests on isolated C1q, C1r and C1s disclosed in addition to the well known interaction between heparin and C1q an equally strong or even stronger interaction between heparin and C1s. Even C1r was adsorbed to heparin although by somewhat weaker ionic bonds.

The third component of complement has been purified from fresh human plasma employing an initial fractionation with poly(ethylene glycol) followed by sequential depletion of plasminogen by affinity adsorbents, chromatography on diethylaminoethylcellulose, gel filtration on agarose, and batch adsorption/desorption on hydroxylapatite. Final recoveries of C3 were 33% of the initial protein, as quantitated by radial immunodiffusion, and 31% of the initial hemolytic activity. Apparent homogeneity is indicated by immunological criteria and by polyacrylamide gel electrophoresis. A partial specific volume of 0.736 +/- 0.003 mlgm-1 was determined for C3 by the mechanical oscillator technique. "Low speed" sedimentation equilibrium yielded an apparent weight average molecular weight for the protein of 187 650 +/- 5650. Based upon this molecular weight, a molar extinction coefficient of 1.82 X 10(5) 1. mole-1 cm-1 at 280 nm was calculated from boundary-spreading experiments in the ultracentrifuge and as assumed refractive index increment. Amino acid analyses revealed no unusual or distinctive characteristics. Automated Edman degradation revealed a double N-terminal sequence, Ser-Val,Pro-Glx,Met-Lee,Tyr-Thr,Ser-Glx,Ile-Lys,Gly-Arg,Thr-Met,Pro-Asx, in agreement with the two chain structure observed on polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate, and revealing both chains are available to degradation. Serine is postulated as the initiating sequence in both chains based upon high recoveries of dinitrophenylserine upon hydrolysis of dinitrophenylated C3, and our inability to identify any other dinitrophenyl or phenylthiohydantoin derivatives in this position. Alanine is the ultimate carboxy-terminal amino acid of at least one of the chains, as indicated by the action of carboxypeptidases on C3 in the presence of sodium dodecyl sulfate.

The sera from two well-documented cases of allergic vasculitis were examined for the presence of C1q precipitins. Both sera contained material capable of precipitating C1q in agarose gel. The material from one of the sera was partially purified using ammonium sulfate precipitation. Sephadex G-200 filtration, and DEAE-cellulose chromatography. Attempts to identify the nature of C1q precipitin were unsuccessful, but it resembled the high-molecular-weight precipitins of sera from patients with systemic lupus erythematosus. In a lesion, deposition of fibrinogen, immunoglobulins, and complement were noted mainly in the vessel walls. No correlation between immunoglobulin and complement deposition in the skin and the presence of C1q precipitins in the blood could be established.

C3, GBG, and orosomucoid polymorphisms were electrophoresed in a high-voltage agarose method which permitted the typing of 15-20 samples. The increased sensitivity of the dye Coommassie blue was used to stain the protein, and yielded higher resolution than amido black. The typing of haptoglobin samples, was facilitated by devising a method which utilizes the peroxidase activity of the haptoglobin-hemoglobin complex with 90-tolidine and hydrogen peroxide as substrates and 4-chloro-1-naphthol as coupler.

Immunofluorescence (IF) studies by the direct and indirect methods demonstrate immunoglobulins and complement bound in vivo in psoriatic scales. The IF pattern is comparable to that of stratum corneum antibodies (SCAb) bound in vitro on specific substrate, as visualized by the indirect IF method. Formation of immune complexes can be responsible for the "squirting papilla" phenomenon, and conversion of the stratum corneum - which is normally an inaccessible antigen - into its reactive form seems to be brought about by proteases of polymorphonuclear leukocytes. Stimulation of protease production by polymorphonuclears appears to be an important factor in the pathogenesis of psoriasis. The stratum corneum of the epidermis is probably the target, and becomes an antigen for SCAb present in the circulation.

We have devised an enzyme-linked immunosorbent assay (ELISA) to quantitate fluid phase terminal complement pathway activation. Upon activation to form C5b-9, terminal complement components express neoantigens not present in the unassembled individual components. Expression of one of these neoantigens occurs at the step of C9 activation. C9 neoantigen is present in fluid phase SC5b-9 complexes, membrane-bound MC5b-9 complexes, and in in vitro polymerized C9. Under physiologic conditions, the presence of C9 neoantigen indicates that the terminal complement pathway is activated through the terminal component C9. In our assay for C9 neoantigen, we used rabbit antiserum to polymerized C9 rendered specific for C9 neoantigenic determinants by serial absorption with human serum, human C9, and other terminal complement components bound to Sepharose. Using the IgG from this antiserum, we devised a sandwich ELISA to bind SC5b-9 from solution onto polystyrene plates. The ELISA plates were developed with the use of goat antiserum to native C9 epitopes followed by a swine anti-goat IgG-alkaline phosphatase conjugate. Quantitation of SC5b-9 in solution was performed by comparing sample OD to a standard curve generated with human SC5b-9 that was purified from zymosan-activated serum. The assay was sensitive to as little as 100 ng of SC5b-9/ml and should be useful for screening plasma, serum, cerebrospinal fluid, or other biological fluids for the presence of terminal complement pathway activation.

This study describes the presence of a receptor for fluid phase human C3 and C3b on Raji cell membranes. The binding of C3 and C3b was demonstrated indirectly by a fluoresceinated anti-C3 serum and directly by using radioiodinated proteins. No other complement proteins or serum factors were needed to mediate binding of C3 and C3b to the receptor. The possibility of enzymatic cleavage of C3 before or after its attachment on the cell membrane was ruled out by the demonstration of antigenically intact C3 on Raji cells. Inhibition and dissociation of Raji cell-EAC1423 rosettes by C3 and C3b indicated that both of these proteins bind to the same receptor site or closely associated receptor sites on Raji cells. C3b-bearing Raji cells were immune adherence negative, indicating that C3b binding to the receptor is brought about through the immune adherence region of the molecule and not the C3d portion. The C3 receptor on Raji cell membranes is uniformly distributed and can move on the membrane plane. Approximately 4 x 10(5) molecules of C3 or C3b bind per Raji cell. The receptor had a higher affinity for C3 than C3b, as was shown by uptake experiments and inhibition of Raji cell-EAC1423 rosette formation. Apart from the described receptor for C3 and C3b another specific receptor for C3b inactivator-cleaved C3b (C3d) bound to red cells was shown to be present on Raji cells. Raji cells cultured in medium containing fresh normal human serum and cobra venom factor were lysed. Similar results were obtained when C3b-bearing Raji cells were cultured in medium with fresh normal human serum. The lytic effect could be abolished by inactivating serum C3 proactivator (C3PA) and required C6. It was concluded that C3b bound to the Raji cell membrane activates the complement system through the alternate pathway and results in membrane damage and cytolysis. It is postulated that cell destruction by this mechanism may play an important role in vivo in controlling cell growth.

Rabbit erythrocytes are lysed by cobra plasma. The membranes of the lysed rabbit erythrocytes exhibited typical ultrastructural complement lesions with an inner diameter of 72 A. Structures similar to human C5b-8 complexes were also visualized. The proteins found in membranes lysed by cobra plasma resembled closely the proteins of the human membrane attack complex in number and molecular weight. Cobra poly C9 that is resistant to reduction and boiling in SDS could also be demonstrated. These results indicate that the membrane attack complex of cobra complement is very similar to its mammalian counterpart.

Test erythrocytes (E) used to evaluate anti-complement (C') antiglobulin sera have not been adequately standardized. This report describes a previously unrecognized C4-derived antigen (temporarily called X-Ag) found on E generally believed to be coated only with the C3d subcomponent of C3, X-Ag occurred on all E coated in vitro with C' by low ionic strength-sucrose or cold agglutinin methods and on E from ten of ten patients whose cells had been C' coated in vivo. It was not removed by incubating these cells with trypsin or fresh compatible serum. This antigen was found on "C4-only-coated" red blood cells made with normal or congenitally C2-deficient serum but not on cells similarly prepared with congenitally C4-deficient serum. It was not identified on E coated with C' via the alternate pathway, normal trypsinized cells, nor cells coated only with IgG. Absorption experiments utilizing purified complement components and subcomponents and G200 Sephadex fractions of normal human serum strongly suggest that X-Ag is a subcomponent of C4(C4d). These results show that at least one C' subcomponent other than C3d occures on both in vitro and in vivo C3d-coated erythrocytes and must be taken into account when such cells are used to evaluate antiglobulin reagents.