Peptides of human thyroglobulin reactive with sera of patients with autoimmune thyroid disease.

Autoantibodies to thyroglobulin (Tg) are a prominent feature of the two autoimmune thyroid diseases, chronic lymphocytic (Hashimoto's) thyroiditis and Graves' disease. Similar autoantibodies are found in the serum of many normal individuals without evidence of thyroid disease. Previous studies have indicated that patients with autoimmune thyroid disease recognize epitopes of Tg which are not usually recognized by normal individuals. The goal of this investigation was to identify peptide fragments of Tg bearing these disease-associated epitopes. For this purpose, we utilized a panel of mAbs that bind to different epitopes of the Tg molecule. One of these mAbs (137C1) reacted with an epitope that was also recognized by the sera of patients with autoimmune thyroiditis. In the present study, we show that two peptides (15 and 23 kDa) that reacted with mAb 137C1 are located in different parts of the Tg molecule. Each peptide inhibited the binding of mAb 137C1 to the other peptide and to the intact Tg, indicating that the same epitope was represented on the two peptides. Loops and helices of the secondary structure of the two peptides might be involved in the conformational epitope recognized by mAb 137C1. A striking finding of this study is that two apparently unrelated fragments of the Tg molecule bind to the same mAb. These findings may have important ramifications with regard to epitope spread and the progression of the autoimmune response to disease.

Linking iodine with autoimmune thyroiditis.

A great deal of circumstantial evidence has linked iodine with the rising incidence of autoimmune thyroiditis in the United States. In our investigations, we have shown directly that T cells from humans with chronic lymphocytic thyroiditis proliferate in the presence of iodinated but not in the presence of noniodinated human thyroglobulin. Moreover, the proliferative response is restored when the thyroglobulin is iodinated artificially in vitro. Using a panel of monoclonal antibodies, we found evidence that the presence of iodine induces a number of stereochemical changes in the conformation of the molecule, resulting in the loss of some antigenic determinants and the appearance of others. One prominent determinant was associated with the iodine-containing amino acid thyroxine. Both the number and position of the iodine substituents determine the precise specificity of this epitope. A new model for the study of the role of iodine in inducing thyroid autoimmunity has become available in the form of the nonobese diabetic (NOD)-H2(h4) mouse. This animal develops autoimmune thyroiditis spontaneously but in relatively low prevalence. However, if iodine is added to the drinking water, the prevalence and severity of the thyroid lesions increase markedly. The immune response is specific for thyroglobulin, both in terms of the antibody response and T-cell proliferation. In fact, the appearance of lesions can be predicted by the presence of thyroglobulin-specific IgG2b antibody. The disease, moreover, can be transferred adoptively, using spleen cells from iodine-fed donors treated in vitro with iodinated thyroglobulin. The effects of iodine feeding are greater in conventional animals compared with those maintained under specific pathogen-free conditions. Based on T-cell proliferation, it appears that the NOD-H2(h4) strain of mice has innately a greater response to murine thyroglobulin than do other mouse strains and that the proliferation is increased even more by feeding iodine. We suggest, therefore, that the presence of iodine increases the autoantigenic potency of thyroglobulin, a major pathogenic antigen in the induction of autoimmune thyroiditis. This animal model provides a unique opportunity for investigating in detail the mechanisms by which an environmental agent can trigger a pathogenic autoimmune response in a susceptible host.

Iodination of human thyroglobulin (Tg) alters its immunoreactivity. I. Iodination alters multiple epitopes of human Tg.

Human Tg, the site of synthesis of thyroid hormones, thyroxine (T4) and triiodothyronine (T3), is one of the major autoantigens in autoimmune thyroiditis. The degree of iodination of Tg may have a major impact on its immunological properties by changing its antigenicity with respect to antibody binding. We have previously prepared a panel of MoAbs that bind to different epitopes of the Tg molecule. In the present study, we show that iodination alters the conformation of Tg molecule in such a way that it is recognized differently by different MoAbs. Monoclonal antibody 137C1 recognizes Tg regardless of its iodine content. Monoclonal antibody 42C3 recognizes Tg only if the Tg is iodinated either in vitro or in vivo. Monoclonal antibody 133B1 recognizes both in vivo iodinated Tg and non-iodinated Tg, but this MoAb did not recognize Tg following in vitro iodination. Monoclonal antibody 41A5 recognizes intact Tg and tryptic peptides of normal (in vivo) iodinated and non-iodinated Tg, but did not bind the tryptic peptides of artificially (in vitro) iodinated Tg. From the results of these experiments, we conclude that iodination of Tg by either in vivo or in vitro methods changes its conformation in such a way that some natural epitopes are 'lost' and some 'new' epitopes are generated. The generation of new epitopes may be important in the generation of autoimmune responses leading to autoimmune disease.

Iodination of human thyroglobulin (Tg) alters its immunoreactivity. II. Fine specificity of a monoclonal antibody that recognizes iodinated Tg.

In a previous investigation, we found that murine MoAb 42C3, raised against human Tg, recognized Tg differently depending upon its level of iodination of Tg. A possible explanation for this finding is that iodine is directly involved with the specific epitope recognized by MoAb 42C3. In the present study, we report that the binding of MoAb 42C3 to iodinated Tg is inhibited by T4, T3, reverse T3 (rT3), triiodothyroacetic acid (triac), diiodothyronine (T2), diiodotyrosine (DIT), but not by thyronine (TO) or tyrosine. The order of inhibition of these iodinated compounds is T4 > T3 > rT3 > triac > T2 > DIT. The MoAb 42C3 does not have the same specificity as the T3, T4-receptor since the order of binding of these iodinated compounds on the receptor differed from the order of their inhibition of this MoAb. Monoclonal antibody 42C3 also recognized non-iodinated Tg that was subsequently iodinated in vitro. It failed to recognize another protein, bovine serum albumin, that was iodinated in vitro by the same method. These results suggest that iodinated tyrosines and thyronines determine the binding specificity of MoAb 42C3. The inhibitory effects of these compounds on MoAb 42C3 depend on their iodine content as well as location of iodine in the aromatic ring.

Iodine is essential for human T cell recognition of human thyroglobulin.

Here we describe for the first time that recognition by human T cells of human thyroglobulin depends upon its iodine content. We have examined the proliferation of lymphocytes from blood of autoimmune thyroiditis patients and normal individuals to thyroglobulin preparations containing different amounts of iodine. A minimal degree of iodination was required to elicit the proliferative response of both patients and normal individuals since thyroglobulin preparations containing no detectable iodine did not induce proliferation. A non-iodinated thyroglobulin preparation that was iodinated in vitro produced significant proliferation of both patient and normal lymphocytes. Addition of IL-2 to the culture medium enhanced proliferation but did not change the pattern of response.

The role of iodine in autoimmune thyroiditis.

Like most cancers, autoimmune diseases generally are due to the interaction of a number of genetic traits with an environmental trigger. Autoimmune thyroiditis, a model of organ-specific autoimmune disease, is associated with iodine as a precipitating environmental factor. T cells from patients with chronic thyroiditis proliferate in response to normal human thyroglobulin, but fail to react with non-iodinated thyroglobulin. Using a selected monoclonal antibody, we were able to identify a binding site on thyroglobulin containing iodinated thyronine. The greatest affinity was for tetraiodothyronine and binding depended upon the number as well as the positions of iodines. We have also studied an inbred strain of mice, NOD-H2h4, that developed thyroiditis spontaneously. The onset of disease was hastened in a dose-dependent manner by adding iodine to the drinking water. The occurrence of disease was greater in conventional than in specific pathogen-free mice and correlated with T-cell proliferation and IgG2b antibody to thyroglobulin.

Cyclosporine therapy suppresses ocular and lacrimal gland disease in MRL/Mp-lpr/lpr mice.

PURPOSE: MRL/Mp-lpr/lpr (MRL/lpr) mice spontaneously develop an autoimmune disease characterized by lymphoproliferation, vasculitis, glomerulonephritis, autoantibody production, and ocular and lacrimal gland inflammation. Lacrimal gland lesions in MRL/lpr mice are a model for the human disorder Sjögren's syndrome. The target organ lesions in MRL/lpr mice, including those in the eye and lacrimal gland, are composed largely of CD4+ T cells, with lesser numbers of CD8+ T cells and B cells. Cyclosporine therapy was evaluated for its effect on the autoimmune disease, particularly in the eye and lacrimal gland. METHODS: MRL/lpr mice were administered cyclosporine intraperitoneally at a dosage of 2 mg daily from age 1 to 5 months. Animals were killed at 5 months and evaluated for the presence of autoimmune disease. Control groups consisted of animals given daily injections with either saline or the cyclosporine diluent. RESULTS: Cyclosporine therapy was effective in reducing the ocular and lacrimal gland disease. Intraocular inflammation was present in 73% of control animals but in only 15% of cyclosporine-treated animals (P < 0.003). Multifocal lacrimal gland inflammatory infiltrates were present in 100% of controls but in only 23% of cyclosporine-treated animals (P < 0.0001). Mean percent area involved by lacrimal gland inflammation was reduced from 19.7% to 4.7% by cyclosporine therapy (P = 0.0003). Systemic autoimmune disease manifestations, including lymphoproliferation, vasculitis, glomerulonephritis, and serologic abnormalities, also were improved. CONCLUSIONS: Chronic cyclosporine therapy, started at an early age, is effective in controlling the autoimmune disease in MRL/lpr mice, including the ocular and lacrimal gland lesions.

Amino acid sequence of a tryptic peptide of human thyroglobulin reactive with sera of patients with thyroid diseases.

Autoantibodies to human thyroglobulin (hTg) are found in the sera of many patients with thyroid diseases. To localize epitopes recognized by these autoantibodies, hTg was incubated with tryspin for 4 hours at 37 degrees C under non-reducing conditions. Releasing peptides from hTg in their natural conformation. These peptides were then analyzed by western immunoblot using either autoantibodies from patients with autoimmune thyroiditis or murine monoclonal antibodies (mAb) produced against hTg. The autoantibodies reacted primarily with two low molecular weight peptides with apparent molecular weights (MWap) of 15 and 20 kDa. The pattern of tryptic peptides recognized by these autoantibodies resembled that of one of the mAbs (137C1), as shown by immunoblots in either one or two dimensional SDS-PAGE. To characterize these peptides further, they were separated by a high performance liquid chromatography (HPLC). The column separated the 4-hour tryptic digest of hTg into multiple peptide peaks. Further analysis by SDS-PAGE showed that one of these peaks contained the 15 kDa peptide. The 15 amino acid sequence at the amino-terminus of this peptide was determined. This amino acid sequence (KVPTFATPWPDFVPR) corresponds to a unique sequence near the carboxyl-terminal end of hTg, starting with amino acid 2657. The size of the peptide indicates that it extends to the carboxyl-terminal end of hTg. This fragment contains one of the antigenic sites of hTg that binds autoantibodies from patients with autoimmune thyroid disease.

Tryptic peptides of human thyroglobulin: II. Immunoreactivity with sera from patients with thyroid diseases.

Tryptic peptides of human thyroglobulin (Tg) were analysed by Western immunoblot for their reactivity to circulating autoantibodies from patients with Hashimoto's thyroiditis (HT), Graves' disease (GD) and thyroid carcinoma, and from normal human controls. Low molecular weight peptides were released after 4 h incubation of Tg with trypsin. The sera of thyroid disease patients reacted with several peptides, but predominantly bound three peptides with apparent molecular weights (MWap) of 25 kD, 20 kD, and 15 kD; the sera of normal individuals did not bind these fragments of Tg. The pattern of tryptic peptides recognized by the majority of sera from GD patients differed from that recognized by sera from most patients with HT. Autoantibodies from both groups of patients recognized a 15-kD peptide with a high frequency, but the sera from 26/43 (60%) GD patients also recognized a peptide with MWap of 25 kD, whereas the sera from 22/35 (63%) of HT patients recognized a 20-kD peptide. A few sera from patients with thyroid carcinoma reacted with peptides with MWap of 15 and 20-kD, and none bound the 25-kD peptide. The immunoreactivity of autoantibodies in HT sera to the 20-kD peptide paralleled the competitive inhibition of the MoAb 137C1 by these sera. In addition, MoAb 137C1 and Hashimoto's sera showed the same Western immunoblot-binding pattern to Tg tryptic peptides, suggesting that a Hashimoto-associated epitope and the 137C1-binding site are found on the same peptide. These findings suggest that distinct peptides are recognized by Tg autoantibodies from patients with different thyroid diseases.

Tryptic peptides of human thyroglobulin: I. Immunoreactivity with murine monoclonal antibodies.

Human thyroglobulin (Tg) was treated with trypsin at different concentrations of trypsin/Tg for various incubation times at 37 degrees C using non-reducing conditions. A ratio of trypsin to Tg of 1:100 (w/w) was optimal to release small peptides that were reactive to murine MoAbs to human Tg. Most peptides were released after only 1 h incubation with trypsin, but these peptides were further degraded at longer incubation times. However, a few small peptides, the largest of which with an apparent molecular weight (MWap) of 40 kD, resisted tryptic digestion up to at least 12 h of incubation. These resistant peptides were further degraded by trypsin at 18-24 h of incubation. Tryptic peptides of Tg, released at 1 h and 4 h of incubation, were analysed for their immunoreactivity to 16 well characterized anti-Tg MoAbs by Western immunoblot. Patterns of peptide recognition of these MoAbs were generally unique. Eight MoAbs reacted with peptides of MWap of 10-25 kD and above. Four other MoAbs reacted with peptides of MWap of 25-43 kD and above, and the remaining four reacted with peptides of MWap > 43 kD. Nine of these MoAbs failed to recognize peptides after reduction, suggesting that the MoAbs bind conformation-dependent epitopes. The above information will promote the development of models relating the structure of Tg to the autoimmune process, and may provide an understanding of those regions of Tg responsible for the induction of autoimmune thyroiditis.

Immunoreactivity of multiple molecular forms of human thyroglobulin.

Human thyroglobulin (Tg) was purified from thyroids of normal individuals and of patients with Graves' disease using gel filtration (Sephacryl S-400) and ion-exchange (DEAE) column chromatography. We isolated five protein peaks of Tg from the DEAE column, using a step gradient, characterized them for protein and iodine content, and assessed their immunological properties by reactivity to polyclonal and monoclonal antibodies (mAbs). Normal and Graves' Tgs differed in the relative protein content of these five protein peaks from the DEAE column. In the case of normal Tg, the majority of Tg was eluted in peak 2 but in the Graves' Tg, most of the protein was eluted in peak 1 of this column. The immunoreactivity of these five protein peaks of Tg was studied using 11 mouse mAbs prepared against human Tg, sera from patients with autoimmune thyroiditis and polyclonal antibody from a rabbit immunized with human Tg. All of five protein peaks of Tg reacted equally with rabbit antibody. The sera of five thyroiditis patients showed greater binding to peak 1 of Graves' Tg than peak 1 of normal Tg. Similarly, most mAbs showed greater binding to peak 1 of Graves' Tg than the peak 1 of normal Tg. Of particular interest was one mAb (42C3) which reacted only with Tgs containing iodine. The immunoreactivity of this mAb paralleled the iodine content of Tg. This mAb might be useful for evaluating the role of iodine in the antigenicity of human Tg.

Effects of early and late treatment with anti-CD4 monoclonal antibody on autoimmune disease in MRL/MP-lpr/lpr mice.

MRL/Mp-lpr/lpr (MRL/lpr) mice spontaneously develop a systemic autoimmune disease, characterized by vasculitis, lymphadenopathy, glomerulonephritis, and autoantibody formation, with target organ inflammatory lesions composed largely of CD4+ (helper) T cells. Previous reports have demonstrated that anti-CD4 monoclonal antibody (mAb) treatment of MRL/lpr mice from 1 to 5 months of age resulted in a dramatic reduction in both the frequency and the severity of autoimmune disease. In order to investigate the effects of early, short-course and late, short-course anti-CD4 mAb therapy on the autoimmune disease in MRL/lpr mice, groups of 12 to 15 animals were treated with weekly intraperitoneal injections according to one of four regimens: (i) anti-CD4 mAb from age 1 to 5 months (continuous treatment); (ii) anti-CD4 mAb from age 1 to 3 months (early treatment); (iii) anti-CD4 mAb from age 3 to 5 months (late treatment); and (iv) either normal saline or rat immunoglobulin (control treatment). Continuous treatment resulted in a dramatic reduction of both frequency and severity of the autoimmune disease, as demonstrated histologically and serologically. Early treatment also resulted in a significant reduction in autoimmune disease, while late treatment had little effect. Glomerulonephritis was detected in none of the animals in the continuously treated group (P < 0.05), 38% of those in the early-treated group (P = < 0.05), 92% of the late-treated group, and 100% of controls. The titer of antinuclear antibodies, of anti-dsDNA antibodies, and total immunoglobulin levels were all significantly reduced in the continuous-treatment and early-treatment groups, but not in the late-treatment group. Murine antibodies to rat anti-CD4 mAb were present in the late-treatment group. These results indicate that early short-course anti-CD4 mAb treatment of MRL/lpr mice is effective in ameliorating the autoimmune disease in this model, while late-treatment is ineffective, probably due to the induction of antibody directed against anti-CD4 mAb itself.

Structural requirements for the inhibition of human monocyte carboxylesterase by organophosphorus compounds.

Human blood monocyte carboxylesterase (CBE) is inhibited by a variety of organophosphorus compounds including arylphosphates and arylphosphites and some alkylphosphites. Triphenyl phosphate and triphenyl phosphite with Ki values of 8 x 10(-9) M and 4.8 x 10(-8) M, respectively, are the most potent inhibitors of this enzyme evaluated by this study. The arylphosphates vary in their capacity to inhibit carboxylesterase activity. Diphenyl phosphate with its strong negative charge is not a potent inhibitor (Ki = 1 x 10(-4) M), whereas if its negative charge is neutralized, as in diphenyl methyl phosphate, its capacity to inhibit carboxylesterase is significantly increased. Compounds with increased bulk, such as trinaphthyl phosphate, only inhibit the enzyme at concentrations of 10(-5) M or greater. Arylphosphites have inhibitory capacities similar to the arylphosphates. Alkylphosphites (tributyl phosphite/triethyl phosphite) inhibit carboxylesterase activity, whereas alkylphosphates (tributyl phosphate/triethyl phosphate) have no inhibitory effect. Arylphosphines and arylphosphine oxides do not inhibit carboxylesterase activity. This study demonstrates that organophosphates and organophosphites are relatively effective inhibitors of human monocyte CBE activity with the exception of the alkylphosphates which have no inhibitory activity. We conclude that molecular bulk and charge have a significant role in determining the potency of organophosphorus inhibitors of monocyte CBE. The observed variations in the degree of esterase inhibition by organophosphorus compounds as well as the differences in the pathological expression of neuropathic disorders associated with such chemicals suggest that different esterase enzymes derived from the family of esterase genes may mediate the different neuropathies observed with organophosphorus exposures. Such data also provide the rationale for the kinetic analyses of esterases and the design of non-toxic organophosphorus compounds with low or no monocyte CBE inhibitory capacity to reduce the potential of these commonly used chemicals for human toxicity.

Human monocyte carboxylesterase. Purification and kinetics.

Human peripheral blood monocytes were isolated by density gradient centrifugation and purified by counterflow centrifugation elutriation. Membrane-localized carboxylesterase (CBE) was extracted with nonionic detergent (Triton X-100) and purified by ion exchange (DEAE-cellulose), gel filtration (Sephacryl S-300), hydroxylapatite column, and high performance liquid chromatography. The purified enzyme migrated on 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis as a single protein band with a molecular weight of 60,000. Under nondenaturing conditions, monocyte CBE formed a trimer and eluted from a gel filtration column as a protein with an approximate molecular weight of 200,000. Electrophoretic patterns of the enzyme on polyacrylamide gels run a neutral pH did not vary during enzyme purification. At least four major isoenzymes of human monocyte CBE were observed with isoelectric points between 7.5 and 7.8. Pure human monocyte CBE hydrolyzed short chain alpha-naphthyl, o-nitrophenyl, and p-nitrophenyl esters. Amide esters and thioesters were not hydrolyzed by the enzyme. Short chain alcohols activated the enzyme and organophosphorus compounds, diphenyl carbonate, sodium fluoride, and phenylmethylsulfonyl fluoride inhibited the enzyme. EDTA and sulfhydryl reagents had no effect on enzyme activity. The amino acid content of the enzyme was consistent with other CBEs. Inhibitors reacted either with the active or effector site of the enzyme. Purified enzyme now permits the characterization of CBE structure and regulation.

Isolation of proteins related to the Rh polypeptides from nonhuman erythrocytes.

It is thought that the Rh antigens may be important in maintaining normal erythrocyte membrane integrity. Despite their name, Rh antigens are serologically present only on human erythrocytes. Rh structural polymorphisms are known to reside within a family of nonglycosylated Mr 32,000 integral membrane proteins that can be purified by hydroxylapatite chromatography. Mr 32,000 integral membrane proteins were purified similarly from erythrocyte membrane vesicles prepared from rhesus monkeys, cows, cats, and rats, but could not be purified from human Rhmod erythrocytes, a rare syndrome lacking Rh antigens. The purified Mr 32,000 polypeptides were labeled with 125I, digested with chymotrypsin, and found to be 30-60% identical to human Rh polypeptides when compared by two-dimensional iodopeptide mapping. The physiologic function of the Rh polypeptides remains to be identified; however, the existence of related proteins in nonhuman erythrocytes supports the concept that the Rh polypeptides are erythrocyte membrane components of fundamental significance.

Two-dimensional iodopeptide mapping demonstrates that erythrocyte Rh D, c, and E polypeptides are structurally homologous but nonidentical.

The 32,000 molecular weight (mol wt) erythrocyte Rh D, c, and E polypeptides were separately purified from cDE/cDE erythrocytes by monoclonal immunoprecipitations and compared by two-dimensional iodopeptide mapping. Digestions of the isolated Rh polypeptides with alpha-chymotrypsin revealed a high degree of structural homology between c and E (13/14 iodopeptides were identical) and less striking homology between D and c or E (8/19 identical). The iodopeptide maps of Rh proteins purified by a nonimmunologic protocol from cDE/cDE erythrocytes were virtually identical to the composite pattern (D + c + E) deduced from the individual maps of Rh D, c, and E iodopeptides. Digestions of the isolated Rh polypeptides with trypsin revealed an overall homology of approximately 50% between iodopeptides derived from D, c, and E. These data indicate that the erythrocyte Rh D, c, and E antigens are carried by homologous but distinct molecular species; c and E appear more closely related to each other than to D.

Polymorphism in the Mr 32,000 Rh protein purified from Rh(D)-positive and -negative erythrocytes.

A Mr 32,000 integral membrane protein has previously been identified on erythrocytes bearing the Rh(D) antigen and is thought to contain the antigenic variations responsible for the different Rh phenotypes. To study it on a biochemical level, a simple large-scale method was developed to purify the Mr 32,000 Rh protein from multiple units of Rh(D)-positive and -negative blood. Erythrocyte membrane vesicles were solubilized in NaDodSO4, and a tracer of immunoprecipitated 125I surface-labeled Rh protein was added. The Rh protein was purified to homogeneity by hydroxylapatite chromatography followed by preparative NaDodSO4/PAGE. Approximately 25 nmol of pure Rh protein was recovered from each unit of Rh(D)-positive and -negative blood. Rh protein purified from both Rh phenotypes appeared similar by one-dimensional NaDodSO4/PAGE, and the N-terminal amino acid sequences for the first 20 residues were identical. Rh proteins purified from Rh(D)-positive and -negative blood were compared by two-dimensional iodopeptide mapping after 125I-labeling and alpha-chymotrypsin digestion. The peptide maps were very similar; however, at least two additional iodopeptides were consistently noted in the Rh proteins purified from Rh(D)-positive erythrocytes. These data indicate that a similar core Rh protein (or group of related proteins) exists in both Rh(D)-positive and -negative erythrocytes, and the Rh proteins from erythrocytes with different Rh phenotypes contain distinct structural polymorphisms.

Purification and partial characterization of the Mr 30,000 integral membrane protein associated with the erythrocyte Rh(D) antigen.

Erythrocytes bearing the Rh(D) antigen have an Mr 30,000 integral membrane protein which can be surface-labeled with 125I and can be quantitatively immunoprecipitated from Triton X-100-solubilized spectrin-depleted membrane vesicles. The 125I-labeled Rh(D)-associated protein was purified to radiochemical homogeneity from membrane skeletons solubilized in sodium dodecyl sulfate and urea by hydroxylapatite chromatography, gel filtration, and preparative polyacrylamide gel electrophoresis. The Rh(D)-associated protein was purified nearly 200-fold from 2 units of erythrocytes from DD individuals by employing similar methods on a large scale using the purified 125I-labeled Rh(D)-associated protein as a tracer. The product appeared to be greater than 95% pure and migrated as a diffuse band of Mr approximately 30,000-32,000 on silver-stained sodium dodecyl sulfate electrophoresis gels poured from 12% acrylamide. It is estimated that the Rh(D)-associated protein makes up approximately 0.5% of the original membrane protein. When concentrated, partially purified Rh(D)-associated protein forms dimers and larger oligomers which are stable in sodium dodecyl sulfate and urea. The Rh(D)-associated protein was protected from degradation when intact erythrocytes or inside out membrane vesicles were enzymatically digested. These studies indicate that the Mr 30,000 protein associated with the Rh(D) antigen is linked to the membrane skeleton, resides within the lipid bilayer with minimal extra- or intracellular protrusions, exists normally as an oligomer, and can be purified in denatured form.