Characterization and use of monoclonal antibodies for isolation of phosphotyrosyl proteins from retrovirus-transformed cells and growth factor-stimulated cells

Protein kinases that phosphorylate the hydroxyl group of tyrosine residues of proteins have been implicated in cell transformation by some retroviruses and in regulation of normal cell growth by some polypeptide growth factors. To facilitate the identification of tyrosine kinase substrates, we developed monoclonal antibodies to the hapten azobenzylphosphonate. One of these antibodies, MA-2G8, proved to be especially attractive in that it bound a derivative of aminophenylphosphate, a close phosphotyrosine analog, with higher affinity than it bound the corresponding derivative of aminobenzylphosphonate; however, its affinity for phosphoserine was negligible. In this paper we describe the optimal conditions for using this antibody to isolate phosphotyrosine proteins, emphasizing particularly that its interaction with phosphotyrosyl proteins is sensitive to ionic detergents and to antibody density on the immunosorbent matrix. The antibody also bound ATP citrate lyase; this enzyme lacks phosphotyrosine but contains phosphohistidine, which is similar structurally to phosphotyrosine. By attaching the antibody at high density to Sepharose beads and omitting ionic detergents from the buffers, it was possible by microbatch immunoadsorption (followed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis) to isolate the 120,000-dalton transforming protein and several other phosphotyrosyl proteins from cells transformed by Abelson murine leukemia virus. Under the same conditions, phosphotyrosyl proteins were also isolated from human epidermal carcinoma cells (A431) that had been stimulated with epidermal growth factor; most prominent among these proteins was the 170,000-dalton receptor for epidermal growth factor.

Quantitative of anti-NP (4-hydroxy-3-nitrophenyl)-acetyl idiotype expression on spleen and thymus cells

Direct binding of 125I-labeled rabbit anti-NPb idiotype antibodies (RaId) was used to quantitate the expression by immune spleen and thymus cells of NPbId, the characteristic Id of the lambda 1-containing antibodies made by C57BL/6 (B6) mice to the (4-hydroxy-3-nitrophenyl)acetyl (NP) group. Direct binding of RaId by B and T cell preparations reached a maximum of 12 ng RaId per 10(8) cells at 7 days after immunization. Spleen T cell preparations maintained similar levels of binding after positive selection for Thy-1.2+ cells and overnight culture. RaId binding was also demonstrated for immune B6 thymus cells and for spleen and thymus cells of immune SJL mice, which have the appropriate heavy chain allotype for NPbId expression but have only barely detectable serum Id. However, the NPbId of T and B cell preparations were indistinguishable by (a) the susceptibility of RaId binding by the cells to inhibition by hapten or by antibodies to the variable regions of lambda light chains (anti-V lambda) and by (b) the ability of anti-V lambda and of monoclonal antibodies to the constant region of lambda 1 chains (anti-C lambda 1) to immunoprecipitate antigen (NP10-bovine serum albumin)-binding proteins from detergent extracts of isotopically labeled cells. The results strongly imply that virtually all of the NPbId of T cell preparations is due to conventional NPbId antibody that is tightly bound to T cells. The results do not, however, exclude the possibility that the T cell preparations contain a trace amount (less than or equal to 1 ng/10(8) cells) of unusual NPbId-like molecules that lack lambda chains.

Influence of the nude and X-linked immune deficiency genes on expression of kappa and lambda light chains

The relative amounts of Ig kappa and Ig lambda1 anti-2,4-dinitrophenyl antibodies were measured at various times after immunizing mice with prototype thymus-dependent (TD), thymus-independent type 1 (TI-1) and thymus-independent type 2 (TI-2) antigens. Similar amounts of Ig lambda1 were produced after TD and TI-2 immunization and somewhat less was produced after a TI-1 stimulus. In contrast, Ig kappa levels were much greater after TD than after TI-1 or TI-2 antigen. The amount of light chain isotype produced appeared to depend on the molecular form in which the hapten was presented, although possible adjuvant effects were not ruled out. Levels of Ig kappa and Ig lambda present in nonimmune sera were measured in normal, xid and nude mice. The kappa/lambda ratio was higher in xid than in normal mice and the difference was demonstrated by F1 analysis to be due to an X-linked gene. Conversely, the kappa/lambda ratio was lower in nude than in normal mice. This was true for the CBA/Tufts (Ighj), CBA.Ighb and C57BL/10 strains. However, there were no detectable differences in the relative frequencies of surface Ig kappa- and Ig lambda-bearing B cells in adult CBA/Tufts, CBA/N and nude mice. Hence, serum ratios may reflect differences at the level of B cell triggering. Two possible explanations for these differences are discussed. Ig kappa and Ig lambda may be expressed on functionally distinct B cell subsets. (For instance Ig lambda-producing cells might be readily triggered by T1 antigens whereas Ig kappa-producing cells are more dependent on T cell signals. Such functional subsets could be determined by light chain expression). Alternatively, cells producing Ig kappa antibody are selected for because they have a higher affinity for antigen. If so, triggering of cells producing high affinity Ig kappa or their subsequent selection is T cell-dependent.

Inhibition of cytotoxic T lymphocyte activity by concanavalin A

Several lectins were tested for the ability to interfere with the recognition or lysis of target cells by cytotoxic T lymphocytes (CTL). Of those tested, concanavalin A (Con A) had the greatest effect: at 0.25 to 10 micrograms/ml, it profoundly blocked antigen-specific lysis of target cells by cloned and uncloned CTL in various allo- and syngeneic systems. Separate pretreatment of CTL and target cells showed that the effect of Con A is exercised predominantly on CTL. Various experimental results ruled out several possible explanations for the inhibitory effect of Con A, such as: CTL-CTL self-killing by nonspecific, lectin-dependent cytotoxicity, impairment of lateral mobility of CTL surface components, and cell agglutination, preventing recycling of CTL among target cells. The findings suggest that subagglutinating concentrations of Con A block CTL activity by binding to and interfering with CTL surface structures that have relatively high affinity for Con A and are critically involved in recognition or lysis of target cells.

Inhibition of cytotoxic T lymphocyte activity by concanavalin A

Several lectins were tested for the ability to interfere with the recognition or lysis of target cells by cytotoxic T lymphocytes (CTL). Of those tested, concanavalin A (Con A) had the greatest effect: at 0.25 to 10 micrograms/ml, it profoundly blocked antigen-specific lysis of target cells by cloned and uncloned CTL in various allo- and syngeneic systems. Separate pretreatment of CTL and target cells showed that the effect of Con A is exercised predominantly on CTL. Various experimental results ruled out several possible explanations for the inhibitory effect of Con A, such as: CTL-CTL self-killing by nonspecific, lectin-dependent cytotoxicity, impairment of lateral mobility of CTL surface components, and cell agglutination, preventing recycling of CTL among target cells. The findings suggest that subagglutinating concentrations of Con A block CTL activity by binding to and interfering with CTL surface structures that have relatively high affinity for Con A and are critically involved in recognition or lysis of target cells.

A small hypervariable segment in the variable domain of an immunoglobulin light chain stimulates formation of anti-idiotypic suppressor T cells

The induction in BALB/c mice of suppressor T cells that block a delayed-type hypersensitivity (DTH) response to the idiotype of M315, a myeloma protein of BALB/c origin, was examined with a variety of immunoglobulin chains and fragments whose amino acid sequences are known. Normal BALB/c mice receiving either the light chain of M315 (L315, lambda 2 isotype) or the variable (V) domain of this chain prior to sensitization with M315 showed marked suppression of DTH to the M315 idiotype. In contrast, neither the heavy chain nor the variable domain of the heavy chain of M315 affected the DTH response. Two other lambda 2 chains were tested and they also failed to suppress DTH to M315. Comparison of amino acid sequences in the three lambda 2 chains indicates that in L315 at most four V region amino acid substitutions (each resulting from a somatic mutation in the V lambda 2 germ-line gene) determine the specificity of the T-cell suppressor pathway. One of the four is in the framework and probably of negligible importance; the other three, however, are all clustered in the third hypervariable loop of the L315 V domain. The tertiary structure of L315 may also be essential, because disruption of intrachain disulfide bonds abolished the ability of the chain to induce suppression.

Synthesis of lambda 1, lambda 2, and lambda 3 light chains by mouse spleen B cells

To determine whether the infrequency of immunoglobulins with lambda 3 light chains is due to a corresponding scarcity of lambda 3 B cells, the production of the various lambda chain subtypes (lambda 1, lambda 2, and lambda 3) by normal spleen cells was compared. The results showed that lambda 1, lambda 2, and lambda 3 chains are produced in a ratio of about 1.0: 0.7 : 0.3, respectively. The argument is made that lambda 1, lambda 2, and lambda 3 B cells exist in the same ratio. Results obtained with neonatal and nude mouse spleen cells suggest that these small differences are not due to stimulatory effects of environmental antigens or regulatory T cells. The much greater disparity in the abundance of lambda subtypes in various antibody responses and serum Ig suggests that lambda 1 B cells may be more likely than lambda 2 or lambda 3 B cells to differentiate into antibody-secreting plasma cells.

Mouse myeloma proteins with lambda 2 and lambda 3 light chains

Serum and ascites fluid from mice bearing 260 different myeloma tumors were screened serologically for myeloma proteins having light chains like that of L315, the lambda 2 chain produced by myeloma MOPC-315. Four proteins, made by myelomas TEPC-952, CBPC-49, ABPC-72, and SAPC-15, were identified. Their light chains were essentially indistinguishable serologically from L315, and each of them also yielded the characteristic C-terminal amino acid (leucine) and C-terminal tryptic peptide of L315. However, amino acid sequences and other findings reported elsewhere have revealed that although the light chain of TEPC-952 is indeed a lambda 2 chain, the light chains of the others (CBPC-49, ABPC-72, and SAPC-15) represent another, slightly different type of light chain, which has been designated lambda 3.

Mouse myeloma proteins with lambda 2 and lambda 3 light chains

Serum and ascites fluid from mice bearing 260 different myeloma tumors were screened serologically for myeloma proteins having light chains like that of L315, the lambda 2 chain produced by myeloma MOPC-315. Four proteins, made by myelomas TEPC-952, CBPC-49, ABPC-72, and SAPC-15, were identified. Their light chains were essentially indistinguishable serologically from L315, and each of them also yielded the characteristic C-terminal amino acid (leucine) and C-terminal tryptic peptide of L315. However, amino acid sequences and other findings reported elsewhere have revealed that although the light chain of TEPC-952 is indeed a lambda 2 chain, the light chains of the others (CBPC-49, ABPC-72, and SAPC-15) represent another, slightly different type of light chain, which has been designated lambda 3.

Identification of a third type of lambda light chain in mouse immunoglobulins

A third type of mouse lambda chain was identified in the course of examining light chains from four myeloma proteins and two monoclonal antibodies. On the basis of their antigenic properties and a characteristic COOH-terminal tryptic peptide, the light chains from these immunoglobulins had previously been provisionally identified as lambda 2 chains. However, four of these chains, designated now as lambda 3 (or lambda III), were found to share constant region features that clearly distinguish them from the other types of mouse lambda chains (lambda 1 and lambda 2). The amino acid sequence of more than 90% of the lambda 3 constant region (from position 120 to the COOH-terminus) was determined. In this region the lambda 3 sequence differs from the lambda 2 constant region at 5 positions and from the lambda 1 constant region at about 30 positions.

Synthetic phospholipid vesicles containing a purified viral antigen and cell membrane proteins stimulate the development of cytotoxic T lymphocytes

Synthetic phospholipid vesicles (liposomes) containing the purified glycoprotein (G) of vesicular stomatitis virus (VSV) and solubilized membrane proteins from cells of the appropriate H-2 haplotype elicited H-2-restricted cytotoxic T lymphocytes (CTL) that lysed VSV-infected target cells. The CTL were elicited by intact liposomes, not by released components. Thus, when spleen cells from VSV-primed H-2d X H-2b hybrid mice were stimulated with liposomes having G protein + membrane proteins from cells with one of the parental H-2 haplotypes, the resulting CTL lysed only VSV-infected target cells with that parent's H-2 type. This result argues against the view that T cells in general recognize only processed antigenic fragments on macrophages. Moreover, liposomes were only effective when G protein and cell membrane proteins were included in the same vesicles. This result suggests that for effective interaction with CTL precursors the antigen (G protein) and products of the H-2 complex must be closer to each other than 600--1,000 angstrom, the diameter of the lipid vesicles used in this study.

Crawling movements of lymphocytes on and beneath fibroblasts in culture

Some lymphocytes become highly motile upon immunological stimulation in vivo or in vitro. When introduced into a culture of 3T3 or L cells and followed by live-cell microscopy, some of these lymphocytes were observed to crawl on top of, along the edges of, and preferentially beneath the attached fibroblasts. The crawling could be as rapid as 20 mum/min, easily detectable without a time-lapse device. The striking ability of crawling lymphocytes to penetrate beneath attached 3T3 cells provided a quantitative means to compare the crawling activity of different lymphocyte populations under various conditions. Crawling was diminished by inhibitors of energy metabolism, by agents that disrupt the cytoskeleton, and by absence of Mg(2+) and Ca(2+), but not of Ca(2+) alone. Crawling lymphocytes were virtually absent in normal thymus and spleen cells. They increased greatly in 5-day mixed lymphocyte cultures and in peritoneal exudate lymphocytes taken after mice had been immunized with allogeneic tumor cells. T cells accounted for most of the crawlers. Of two T-cell leukemias tested, R1(+) cells were crawlers whereas EL-4 cells were not. The H-2 haplotype of the 3T3 fibroblasts (i.e., whether syngeneic or allogeneic) had no apparent effect on lymphocyte crawling activity. The crawling may relate to the exploration of cell surface antigens by lymphocytes (immune surveillance), to the mode of action of cytotoxic T cells, to the migration of lymphocytes across blood vessel walls, or to the penetration of lymphocytes into "solid" masses of normal tissue or tumor cells.

Cyclical changes in susceptibility of a myeloma tumor (LPC-1) to immune destruction. I. Changes in reactivity with cytotoxic T lymphocytes and anti-H-2d sera

One of six transplantable ascites tumors of BALB/c mice was found to become periodically resistant to cytotoxic T lymphocytes (CTL). About 12 days after LPC-1, a myeloma tumor, was transplanted it became resistant to lysis by allogenic CTL (anti-H-2d) and by CTL directed to trinitrophenyl groups or minor histocompatibility antigens. Susceptibility to lysis by all of these CTL was regained within 2 to 4 days after transfer of resistant cells to a fresh BALB/c host. These changes were recurrent: in each transplantation cycle the early LPC-1 cells were susceptible and the late cells were resistant to CTL. Analyses with antisera (B10 anti-B10.D2) showed that the serologically recognized products of the H-2d haplotype were reduced about 10-fold on the LPC-1 cells that were resistant to CTL.

The natural abundance of lambda2-light chains in inbred mice

The amino acid sequence of the constant (C) domain of the light chain of the mouse myeloma protein M315 has not been identified so far in any other myeloma protein. In this study, serological analysis with antiserum to the C-domain of this light chain (L315) showed that approximately equal to 1% of Igs in normal mouse serum have L chains of the L315 type (called lambda2). Corroborative evidence was obtained by analysis of the carboxyterminal amino acid removed from normal light chains by carboxypeptidase A. A survey of 35 inbred mouse strains showed that all had lambda2; the serum level of Igs with lambda2-chains ranged from approximately equal to 140 microgram/ml in AL/N mice to approximately equal to 25 microgram/ml in SJL, BSVS, and eight other strains. In accord with the anti-Dnp activity of M315, sera from mice immunized with Dnp-KLH had three- to fivefold more lambda2 than sera from control mice immunized with KLH. It was also possible to measure serum immunoglobulin molecules bearing the lambda2 variable region of M315 (VL315). In BALB/c sera, the concentration of VL315 was about sixfold lower than that measured for lambda2. Thus, lambda2-chains are divided into at least two subsets: those whose V domain is indistinguishable from VL315 and those whose VL differs from VL315. A 10-fold increase in VL315 was obtained by immunizing BALB/c mice with Dnp-KLH. The relationship of the VL domains of normal immunoglobulin lambda2-chains to the embryonic Vlambda gene recently sequenced by Tonegawa et al., is discussed.