Isolation of membrane associated immunoglobulins from T lymphocytes by non-ionic detergents

An unusual form of lipid linkage to the CD45 peptide

Some protein kinases and phosphatases are myristoylated on their amino terminus, which perhaps contributes to subcellular localization or regulation. Glycoprotein CD45, a hematopoietic tyrosine phosphatase, was examined for fatty acid content. The CD45 protein incorporated [3H]myristate, but little [3H]palmitate. The label was not metabolized and reincorporated into amino acids or saccharides, as revealed by peptide maps of CD45 labeled with [3H]myristate, 14C-labeled amino acids, [35S]methionine, or 125I, and glycosidase treatments, respectively. The myristate label was resistant to mild alkaline methanolysis and was found in fatty acid and sphingosine, indicating an unusual form of lipid attachment to CD45.

Effects of colchicine and cytochalasin B on distribution of concanavalin A receptors in isolated and cultured guinea pig epidermal cells

Regulation of the distribution of concanavalin A (Con A)/receptor complexes by the cytoskeletal contracture system was studied in guinea pig epidermal cells in suspension and culture using the fluorescence double staining method. After treatment with 100 micrograms/ml of Con A at 37 degrees C for 30 min lectin/receptor complexes were endocytosed by the less-differentiated cells in suspension and by the adherent cells in 1- and 3-day cultures that represent a growing cell fraction. The same treatment resulted in diffuse surface distribution of the complexes in the well-differentiated cells in suspension. Colchicine (10(-5) and 10(-6) M) inhibited internalization of the complexes with resultant diffuse distribution in 60% of the adherent cells in culture. Cytochalasin B (5 and 10 micrograms/ml) not only inhibited endocytosis but promoted formation of surface patchy clumps of the complexes in suspended, less-differentiated cells and cultured adherent cells. The distribution profile was not influenced by these drug treatments in the well-differentiated cells. SDS polyacrylamide gel electrophoresis and autoradiography of 125I-labelled epidermal membranes revealed several Con A-reactive polypeptides common to the cells at various differentiation steps. The progressive decrease in endocytosis and mobility of Con A/receptor complexes was suggested to occur with differentiation. In the germinative cells the distribution of lectin/receptor complexes seemed to be regulated by microfilaments and microtubules.

Sialylation patterns of lymphocyte function-associated antigen 1 (LFA-1) differ between T and B lymphocytes

Lymphocyte function-associated antigen 1 (LFA-1) was immunoprecipitated from various types of surface-radioiodinated murine lymphocytes, and analyzed by two-dimensional polyacrylamide gel electrophoresis. LFA-1 alpha and beta chains from splenic B lymphocytes had the same apparent molecular weights as, but distinct isoelectrofocusing patterns from, their counterparts from thymocytes or splenic T lymphocytes. The splenic B lymphocytes lacked a basically charged population of alpha chain, while the thymocytes and the splenic T lymphocytes showed both the acidic and the basic portions. Furthermore, the beta chain of the former migrated more towards the acidic end than that of the latter. No difference was found between LFA-1 molecules of the same lineage of cells from several strains of mice whose H-2 haplotypes were different from one another. When murine lymphocyte lines were examined, LFA-1 with various isoelectrofocusing patterns were recognized. The charge difference again reflected the difference in lymphocyte lineage, but in a more exaggerated manner than that seen with cells from mice. The average acidity of both chains of LFA-1 decreased in the order of B cell lines, pre-B cell lines and T cell lines. The lineage-dependent charge difference of either chain disappeared after neuraminidase treatment of LFA-1, indicating that lymphocyte differentiation was accompanied by changes in LFA-1 sialylation.

Direct purification of polyadenylated RNAs from isolated polysome fractions

We report a simplified and improved method for obtaining polyadenylated RNAs (poly(A) RNAs) from polysome fractions. Isolated polysomes were subjected directly to poly(U)-Sephadex column chromatography without conventional purification of polysomal RNAs by phenol extraction followed by ethanol precipitation. The yield of poly(A) RNAs by this direct purification method was about twice that obtained by the conventional method. When the two poly(A) RNA preparations were used in two-dimensional polyacrylamide gel electrophoretic analysis of cell-free translation products and cDNA synthesis, biological activity of the directly purified poly(A) RNA was equal to or even better than that of conventionally purified poly(A) RNA.

Two-dimensional peptide mapping by polyacrylamide-gel electrophoresis with limited proteolysis in SDS

The peptide mapping method described by Cleveland, et al. (1) was improved to a two-dimensional analysis applicable to minute amounts (less than 0.5 microgram) of proteins. Radioiodinated proteins for analysis were purified by electrophoretic elution of the proteins from polyacrylamide gels into buffer containing 0.1% sodium dodecyl sulfate. The proteins were digested enzymatically in the presence of 0.1% sodium dodecyl sulfate and an excess of nonlabeled bovine serum albumin (0.2 mg/ml) relative to labeled substrate in order to attain reproducibility by maintaining a consistent substrate concentration among different samples. The peptides of these limited proteolytic products were resolved by two-dimensional polyacrylamide gel electrophoresis (isoelectric focusing followed by SDS-gels). The resulting 2D-peptide maps of murine and bovine albumin and a murine lymphocyte membrane protein, Tp100, showed excellent resolution and reproducibility.

Regulatory idiotopes. Induction of idiotype-recognizing helper T cells by free light and heavy chains

Previously, we have demonstrated the induction of T helper cells that recognize idiotype by antigen (19), idiotype (20), and antiidiotype (12). The T cell population has been characterized and found to recognize both the T15 and M167 myeloma proteins, which share PC binding specificity but differ in idiotypic specificities. In the present work, we used isolated heavy and light chains of T15 and M167 to generate T helper cells, and examined the response to trinitrophenyl (TNP)-T15 and TNP-M167. We found that the heavy chains induced a dose-dependent response to TNP-T15 and TNP-M167, while the light chain priming was ineffective. When isolated chains of a monoclonal anti-T15 antibody (F6-3) were used to induce idiotype-recognizing T cells, only the F6-3 light chains generated T cell help for TNP-T15 and TNP-M167. Evidently, the idiotypic determinant that is recognized by the T cells is not dependent upon the conformation of combined heavy and light chains. These data show that the Th2 helper cells for the T15/M167 idiotopes are induced by free heavy chains of T15 and M167; the Th1 T cells that interact with the Th2 population, of T15 and M167; the Th1 T cells that interact with the Th2 population, however, can be triggered by free light chains of an antiidiotypic hybridoma antibody. These provocative findings suggest a new model for the T helper cell network.

Analysis of normal and neoplastic lymphocyte surface-labeled proteins by two-dimensional polyacrylamide gel electrophoresis

Normal and neoplastic murine and human lymphocytes were surface-labeled by lactoperoxidase-catalyzed radioiodination, and the cell lysates were subjected to two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) analyses, combining isoelectric focusing in the first dimension and sodium dodecyl sulfate-PAGE (SDS-PAGE) in the second dimension. 2D-PAGE autoradiogram patterns were reproducible and reflected differences in cell types. A string of spots with a Mr of 100K was tentatively identified as a new T-cell marker (Tp100) which was present in all murine and human T cells examined including human T lymphomas. Murine and human B cells displayed markers characteristic to B cells of each species with some similarities between them. Human lymphomas and murine cell lines showed markers which were absent or only weakly visible in normal cells. Thus, 2D-PAGE analysis of lymphocyte surface proteins proved to be a method useful for searching for various markers.

T cells produce an antigen-binding factor with in vivo activity analogous to IgE antibody

T cell-dependent activation of resident tissue mast cells is required for the elicitation of delayed-type hypersensitivity skin reactions in mice. A T cell-derived antigen-binding factor that transfers the ability to elicit an immediate hypersensitivity-like skin reaction is described and compared with a hybridoma IgE antibody. Both the T cell factor and IgE mediate reactions with increased vascular permeability and both are mast cell dependent, as they are inactive in two different types of mast cell deficient mice (W/Wv and Sl/Sld). The T cell factor was distinguished from IgE by affinity chromatography using specific anti-IgE and anti-factor antibodies and by a shorter duration of passive sensitization. The T cell factor is a suitable candidate for participation in the mechanism by which T cells activate mast cells in delayed-type hypersensitivity.

Properties of reticulum cell sarcomas in SJL/J mice. VIII. Prominent role of RCS cell I-A antigens in the stimulation of syngeneic T cells

While T cells from SJL and from F1 hybrids of SJL that do not express I-E antigens give strong proliferative responses to RCS, T cells from F1 hybrids expressing surface I-E do not. The nature of the stimulating antigen on the RCS cell surface was examined using monoclonal antibodies. Complete inhibition of the T-cell proliferative response was obtained with antibodies to I-A antigens, whereas antibodies to I-E antigens did not inhibit at all. This inhibition was mediated via an effect of the antibodies on the stimulating cells. Biochemical characterization of immunoprecipitated 125I-and 35S-labeled RCS antigens was performed using two-dimensional gel electrophoresis. Using this technique, I-A antigens were readily detected. However, neither Ia.7-specific antibodies nor antibodies specific for E alpha: E beta complexes precipitated any E alpha or E beta chains. Comparison of I-A antigens from RCS and normal SJL spleen cells revealed minor mobility differences in the gels, possibly due to differences in glycosylation, the significance of which needs to be further evaluated. Examination of RNA extracted from RCS, using E alpha and A alpha cDNA probes showed that RCS cells do not transcribe the E alpha gene as has been shown previously for normal H-2s cells. Furthermore, DNA from RCS cells showed a defect in the E alpha gene similar to that known to exist in normal H-2s cells. Our findings exclude the presence of E alpha on RCS cells and suggest a major role for I-A, either alone or in conjunction with another as yet unidentified cell surface antigen, in the stimulation of T cells.

The isolation of an antigen binding factor from cultured murine T-lymphocytes

BALB/c mice were immunized with the random sequence polypeptide [Glu60 Ala40] (GA). 125I-lymphocyte suspensions were prepared from lymph nodes, radiolabelled with 125I and cultured for 18 hours. Supernatant fluids were collected and passed over an immunoabsorbant composed of [D-Glu60 D-Ala40] (D-GA) bound to Sepharose. The non-adherent fraction was then applied to GA Sepharose and the adherent material eluted with 2 M NaSCN. The eluate was relabelled with 125I, reabsorbed to GA Sepharose and eluted. This fraction bound well to GA-Sepharose (50-80%) but not to the enantiomorphic D-GA Sepharose (5-10%). This material did not react with goat antisera directed against murine IgG, IgM or IgA. It was capable of binding to rabbit antiidiotypic sera raised against BALB/c anti-GA, but not to those antibodies raised against BALB/c anti-GAT. It also showed some reaction with rabbit antiserum to a rat T cell factor (TCF). Polyacrylamide gel electrophoresis revealed a major component with an approximate molecular weight of 63,000 which was not affected by reduction and alkylation.

The still elusive T cell receptor: on the possibility of a common V-gene pool for B- and T-cell-antigen receptor molecules

The contention that VH constitutes a part of T-cell receptors for antigens was probed by purifying rabbit T cells and analysing these cells for non-immunoglobulin VH, i.e. VH not associated with L chain. A number of anti-VH antisera were employed for this purpose, the most important being goat antiserum, reacting with common a1 allotype determinants (allotype determinants expressed on free VH and H chain as well as on intact immunoglobulins), rat antibody against common non-allotype VH determinants (VH framework determinants expressed on VH and H chain as well as on intact immunoglobulins) and chicken antibody against unmasked non-allotype determinants (VH determinants accessible only in the absence of L chain). VH and L chain was quantified by radioimmunoassays on extracts and supernatants from unstimulated T cells as well as from T cells stimulated by concanavalin A and by allogeneic cells. Absolute depletion of Ig-containing and -producing cells was not achieved but in no case was an excess of VH over L chain observed. This indicated that all detected VH originated from cells of the B lineage. The cells were also cultured in the presence of labelled amino acids followed by analysis of detergent extracts and supernatants by immunoadsorption and sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) fluorography. Again, no evidence for T-cell VH could be found. Affinity purified anit-VH antibody was used to label viable rabbit T cells through the use of secondary fluorescence-labelled anti-immunoglobulin antibody. No VH-specific labelling of T cells could be observed. Mixed lymphocyte cultures were carried out in the presence of affinity-purified anti-VH antibodies. No inhibition of the reaction could be discerned. The failure to detect T-cell VH is in agreement with the recent finding that the VH-genome in T cells is not rearranged in a functional manner similar to that in B cells.

Transfer of memory cells into antigen-pretreated hosts. II. Influence of localized antigen on the migration of specific memory B cells

Following i.v. injection, 2,4,6-trinitrophenol (TNP)-primed memory cells localized in recipient lymph nodes draining a footpad injection of TNP-hemocyanin (TNP-KLH) in greater numbers than in contralateral nodes draining a p-azobenzenearsonate-coupled KLH injection. Such hapten-specific, unilateral memory B cell localization was still observed in immunosuppressed mice when antigen injections were given as long as 4 days prior to the memory cell transfer. The memory cells could be challenged to form plaque-forming cells by footpad injections of TNP-labeled Brucella abortus at 5 days, but not one day, after cell transfer. The present studies further clarify some parameters of this adoptive memory, as a model for the study of persistent local memory. Measures that promoted the unilateral lymph node retention of 125I-labeled antigen also facilitated unilateral accumulation of TNP-specific memory cells. Such measures included pretreatment of the recipients with cyclophosphamide, rather than gamma irradiation, injection of anti-carrier antibody the day before antigen, or use of small doses of preformed immune complexes instead of antigen alone. In general, a high ratio of lymph node-to-spleen and lymph node-to-blood concentration of antigen in recipients appeared crucial for unilateral localization of memory B cells. Splenectomy of recipients prior to cell transfer enhanced the difference in plaque-forming cell responses between draining and contralateral nodes, but decreased their difference when performed 1 day after cell transfer, suggesting that the spleen may have served as a trap for memory cells. I.v. injection of antigen at the time of B cell transfer also interfered with unilateral localization. The results demonstrate that in the presence of persisting depots of antigen within lymph nodes (and absence of significant amounts of antigen elsewhere), memory B cells can be retained locally without activation into antibody-secreting cells. This mechanism may, therefore, by responsible for the phenomenon of local, humoral, immunological memory.

Characterization of T-cell surface proteins bound by heterologous antisera to antigen-specific T-cell products

Heterologous antisera specific for murine T-cell antigen-recognition molecules were prepared by immunization of rabbits with dinitrophenyl-specific murine T-cell suppressor factors that had been purified by hapten-affinity chromatography. The antisera (i) bind to antigen-specific T-cell products that differ in their antigen-recognizing specificity; (ii) absorb the specific suppressor activity in preparations containing suppressor factors; (iii) stain all Lyt2+ T cells brightly in indirect immunofluorescence examination, stain some Lyt1+ cells (with low intensity), and do not stain B cells; (iv) precipitate cell membrane proteins from T cells that bear striking structural resemblance to the antigen-specific molecules used for immunization. These results suggest that, like B cells, there is a commonality between antigen-specific effector molecules released by T cells and their membrane-associated receptors.

Isolation and partial characterization of an antigen-specific T-cell factor associated with the suppression of delayed type hypersensitivity

Antigen-specific factors associated with immunosuppressive activity, released by cultured T cells from mice tolerant to the haptens trinitrophenyl, dinitrophenyl and oxazolone, were purified by hapten affinity chromatography. Their binding specificity for antigens paralleled their immunoregulatory activity. Like some immunoglobulin molecules, these factors had blocked NH2 termini and could be bound to Fc-like receptors on macrophages. However, neither immunoglobulin constant region determinants (isotypes) nor antigens encoded by the major histocompatibility complex were detected on the suppressive factors. The purified factors occurred as 68,000-dalton proteins and non-covalently linked dimers. No associated immunoglobulin light chain molecules were detected. The factors showed a marked propensity toward degradation with major breakdown products of 45,000-50,000 and 25,000-30,000 daltons. These results suggest that these molecules are the T-cell products analogous to B-cell immunoglobulin (equivalent to heavy chains) and that they may be the antigen-specific components which act in conjunction with major histocompatibility-controlled gene products to perform antigen-specific suppression.

Immunoglobulin mu- and gamma-ribonucleic acid sequences in thymocytes and splenocytes from normal and hyperimmune mice

The immunoglobulin heavy-chain ribonucleic acid (RNA) repertoire of mouse thymocytes was examined. Previously, this laboratory reported immunoglobulin alpha-chain RNA sequences in mouse thymocytes [Near, R. I., & Storb, U. (1979 Biochemistry, 18, 964]. We have extended these studies to encompass mu, gamma 2b, and gamma 1 heavy-chain RNA sequences, mu-, gamma 2b-, and gamma 1-messenger RNAs (mRNAs) were purified from myelomas to 45, 22, and 54% purity, respectively. Each of these mRNAs faithfully translated into the appropriate immunoprecipitable protein in a reticulocyte lysate translation system. The gamma 1-mRNA translated into two major immunoprecipitable products of about 52 500 and 51 000 daltons while mu- and gamma 2b-mRNAs yielded only a single major protein. Complementary deoxyribonucleic acids (cDNAs) prepared from the mRNAs were used as hybridization probes and revealed the presence of about 70 mu-RNA sequences per average thymocyte as determined by hybridization kinetics, while gamma 1 and gamma 2b sequences were at the limits of detection. The mu-RNA sequences are present in the cytoplasm and are greater than 50% polyadenylated. Upon hyperimmunization of mice with sheep red blood cells, gamma 1-RNA in splenocytes increased by about 100-fold while only slightly increasing in thymocytes. mu and gamma 2b increased 2-3-fold in splenocytes and only slightly in thymocytes. The results argue against RNA sequences appearing in thymocytes due to contamination with peripheral confirmed with cloned cDNA probes. Thymocyte RNA analyzed by Northern blots displayed bands of the same size as those in splenocyte RNA or in purified mRNA when hybridized to mu, gamma 2b and alpha cloned probes. Also, K light-chain RNAs of the same size were found in spleen and thymus by using a cloned K-DNA probe. The results are consistent with the thymus containing mu-, alpha-, and K- and small amounts of gamma 1- or gamma 2b-RNAs coding for heavy- and light-chain-like proteins which may play a role in T-cell function.

Selective turnover and shedding of H-2K and H-2D antigens is controlled by the major histocompatibility complex. Implications for H-2-restricted recognition

Lactoperoxidase-catalyzed cell surface radioiodination and incorporation of [3H]-leucine were employed to radiolabel H-2K and H-2D antigens of murine spleen cells. The fate of H-2 antigens was monitored by in vitro culture of labeled cells and isolation of labeled antigens from detergent lysates of the cells and culture supernates obtained at different times during culture. H-2Kk antigens were found to be rapidly turned over and shed by CBA/J cells, whereas the turnover of H-2Dk antigens was extremely slow. Analysis of the membrane residence times of surface-labeled H-2K and H-2D antigens on spleen cells from various H-2-congenic and -recombinant strains demonstrated variations in the shedding rates of H-2K and H-2D antigens, which were controlled by genes mapping in the major histocompatibility complex. These variations show a striking correlation with published, genetically controlled quantitative variations in the cytotoxic response of T lymphocytes to chemically modified or virus-infected syngeneic cells.

Surface component of primate thymus-derived lymphocytes related to a heavy chain variable region

In a study designed to determine whether T cells of man and higher primates express a surface component related to the variable region of immunoglobulin heavy chain (VH), chickens were immunized with the purified VH fragment of a monoclonal Waldenström macroglobulin. The antibody preparation reacted with a mu chain determinant contained in the Fd fragment and with individual determinants characteristic of the orginal Waldenström protein. As estimated by immunofluorescence analysis, a subpopulation of normal human peripheral T cells (approximately 30%) bound the anti-VH antibody. B-Cell lymphoma lines grown in vitro, as well as some T-cell leukemia lines of the cotton-topped marmoset (Sagiunus oedipus), also bound the anti-VH antibody. The VH-bearing component of the T-cell line 70-N-2 was labeled biosynthetically by incorporation of [3H]leucine and was precipitated specifically by anti-VH antibody. This component was characterized by an apparent mass of 70,000 daltons as assessed by polyacrylamide gel electrophoresis under reducing conditions in buffers containing sodium dodecyl sulfate. These data provide direct support for the hypothesis that some T cells express and synthesize a component related to immunoglobulin heavy chains.

Solubilization of complexes between tumor-associated and H-2 antigens from a tumor cell plasma membrane

The tumor-associated cell surface antigens (TAA) gp 71 and AgI cocap H-2, but are not associated with H-2 following solubilization of membrane with detergent. We have attempted to solubilize intact TAA/H-2 complexes from surface-labeled L cells using decreasing concentrations of the nonionic detergent Nonidet-P 40 (NP 40). Associations between gp 71 and H-2 and between AgI and H-2 in solution were detected using sequential immunoprecipitation. gp 71/H-2 and AgI/H-2 complexes were solubilized at the lowest detergent concentration used, 0.01% NP 40. No coprecipitation of these antigen pairs was seen at any higher NP 40 concentration tested. Sodium dodecyl sulfate polyacrylamide gel electrophoresis revealed lthat precipitation of antigens solulbilized in 0.01% NP 40 leads to coprecipitation of a small number of other protein peaks and does not appear to involve wholesale inclusion of labeled membrane proteins. These results, together with previous observations of cocapping between TAA and H-2 antigens, suggest that the molecules are associated with one another in the membrane. Detergent solubilization of membranes with all but the most gentle conditions leads to disruption of these complexes.

Differential radiolabelling of lymphocyte membrane alloantigens and immunoglobulins: variation of H2O2 concentration during lactoperoxidase catalyzed cell surface radio-iodination

The effect of H2O2 concentration on lactoperoxidase catalyzed cell surface radio-iodination and subsequent isolation of Murine splenic lymphocyte Ia, H-2K and Lyb-3 surface antigens and membrane immunoglobulins was studied. For most membrane polypeptides analyzed 0.3 mM H2O2 proved to be optimal for the recovery of radiolabelled antigens from detergent lysates of labelled cells by immunoprecipitation. Marked variations among surface antigens and membrane immunoglobulin polypeptide chains were observed for the iodination and recovery of these proteins above and below the optimal peroxide concentration. The results suggest that cell surface radio-iodination conditions should be standardized to the requirements of the particular membrane protein being studied. The differential iodination and recovery of discrete membrane components above and below optimal conditions may prove useful in the analysis of surface membrane protein structure and membrane association.

Molecular properties of T-lymphoma immunoglobulin. II. Peptide composition of the heavy chain

In order to obtain structural information and to facilitate studies of the covalent structure of T-cell immunoglobulin, we have performed investigations of the peptide fragments of the heavy chain of this molecule, on a comparative basis, with heavy chains of serum immunoglobulins. T-cell immunoglobulin was isolated from 125I-labelled culture medium of monoclonal continuously cultured T-lymphoma cells by means of immunoadsorbents. Peptides were prepared from the purified 125I-labelled heavy chain by means of cleavage with cyanogen bromide or digestion with trypsin. We then resolved these peptides and compared them with those peptides derived from 131I-labelled murine mu, gamma and alpha chains separately and in mixed label experiments by means of polyacrylamide gel electrophoresis in sodium dodecyl sulfate containing buffers, 2 dimensional peptide mapping and ion exchange chromatography. The profiles of the radiolabelled peptides obtained from the T-lymphoma heavy chains were quite distinct from those of murine gamma chains but indicated a structural similarity to both mu and alpha chains, which share some common peptides. These results are consistent with the antigenic and physicochemical data available and suggest that T-cell immunoglobulin is a new isotype that shows similarities to IgA and IgM, but of which the precise nature of the constant region has yet to be delineated.

Molecular properties of T-lymphoma immunoglobulin. III. Peptide composition of the light chain

Structural studies of T-cell immunoglobulin light chains have been carried out in order to ascertain whether they possess a unique structure or if they resemble standard kappa or lambda isotypes. T-cell immunoglobulin was isolated from 125I-labelled culture medium of monoclonal, continuously cultured T-lymphoma cells, and the purified 125I-labelled light chains were subjected to either cleavage by cyanogen bromide or digestion with trypsin. These peptides were then resolved and compared with those derived from 125I- and 131I-labelled murine kappa and lambda chains in mixed label experiments, by means of polyacrylamide gel electrophoresis in sodium dodecyl sulfate-containing buffers and ion exchange chromatography. The profiles obtained suggested that T-lymphoma immunoglobulin light chains are immunoglobulin polypeptides and most closely resemble kappa chains. These results support available antigenic and mRNA hybridization data, which also suggest that T cells bear kappa-like light chains.

Molecular properties of T lymphoma immunoglobulin. I. Serological and general physicochemical properties

Immunoglobulin (Ig) released into the medium by monoclonal continuously cultured murine T lymphoma cells of the lines WEHI-22 and WEHI-7 was isolated by serological precipitation or solid-phase immunoadsorption techniques. The intact immunoglobulin had an electrophoretic mobility on sodium dodecyl-sulphate (SDS) containing polyacrylamide gels comparable to that of IgG (mass 150,000). This mobility was significantly faster than that of '7S' IgM of murine B lymphocyte surfaces. The T lymphoma immunoglobulin consisted of a pair of heavy chains linked by disulphide bonds and light chains non-covalently bound to the heavy chains. The isolated heavy chains migrated slightly faster than the mu chains of MOPC 104E IgM. Some, but not all, antisera directed against mu chains of normal mouse serum IgM bound T lymphoma immunoglobulin apparently via a cross-reaction localized to the Fd fragment. These data indicate that immunoglobulin of T lymphoma cells and, presumably normal T lymphocytes, represents an immunoglobulin isotype which is distinct from those immunoglobulins found on the B cell surface.