The requirement of light chain for the surface deposition of the heavy chain of immunoglobulin M

The dispensability of VH-VL pairing and the indispensability of VL domain integrity in the IgG1 secretion process

Introduction: The CH1 domain of IgG antibodies controls assembly and secretion, mediated by the molecular chaperone BiP via the endoplasmic reticulum protein quality control (ERQC) mechanism. However, it is not clear whether the variable domains are necessary for this process. Methods: Here, we generated IgG1 antibodies in which the V domain (VH and/or VL) was either removed or replaced, and then assessed expression, assembly, and secretion in HEK293 cells. Results: All Ig variants formed a covalent linkage between the Cγ1 and Cκ, were successfully secreted in an assembled form. Replacement of the cognate Vκ with a non-secretory pseudo Vκ (ψVκ) hindered secretion of individual or assembled secretion of neither heavy chains (HCs) nor light chains (LCs). The ψLC (ψVκ-Cκ) exhibited a less folded structure compared to the wild type (wt) LC, as evidenced by enhanced stable binding to the molecular chaperone BiP and susceptibility to proteolytic degradation. Molecular dynamics simulation demonstrated dramatic alterations in overall structure of ψFab (Fd-ψLC) from wt Fab. Discussion: These findings suggest that V domains do not initiate HC:LC assembly and secretion; instead, the critical factor governing IgG assembly and secretion is the CH-CL pairing. Additionally, the structural integrity of the VL domain is crucial for IgG secretion. These data offer valuable insight into the design of bioactive molecules based on an IgG backbone.

Specific protein interactions between rice members of the GT43 and GT47 families form various central cores of putative xylan synthase complexes

Members of the glycosyltransferase (GT)43 and GT47 families have been associated with heteroxylan synthesis in both dicots and monocots and are thought to assemble into central cores of putative xylan synthase complexes (XSCs). Currently, it is unknown whether protein-protein interactions within these central cores are specific, how many such complexes exist, and whether these complexes are functionally redundant. Here, we used gene association network and co-expression approaches in rice to identify four OsGT43s and four OsGT47s that assemble into different GT43/GT47 complexes. Using two independent methods, we showed that (i) these GTs assemble into at least six unique complexes through specific protein-protein interactions and (ii) the proteins interact directly in vitro. Confocal microscopy showed that, when alone, all OsGT43s were retained in the endoplasmic reticulum (ER), while all OsGT47s were localized in the Golgi. co-expression of OsGT43s and OsGT47s displayed complexes that form in the ER but accumulate in Golgi. ER-to-Golgi trafficking appears to require interactions between OsGT43s and OsGT47s. Comparison of the central cores of the three putative rice OsXSCs to wheat, asparagus, and Arabidopsis XSCs, showed great variation in GT43/GT47 combinations, which makes the identification of orthologous central cores between grasses and dicots challenging. However, the emerging picture is that all central cores from these species seem to have at least one member of the IRX10/IRX10-L clade in the GT47 family in common, suggesting greater functional importance for this family in xylan synthesis. Our findings provide a new framework for future investigation of heteroxylan biosynthesis and function in monocots.

Augmenting recombinant antibody production in HEK293E cells: Optimising transfection and culture parameters

Background

Optimising recombinant antibody production is important for cost-effective therapeutics and diagnostics. With impact on commercialisation, higher productivity beyond laboratory scales is highly sought, where efficient production can also accelerate antibody characterisations and investigations.

Methods

Investigating HEK293E cells for mammalian antibody production, various transfection and culture parameters were systematically analysed for antibody light chain production before evaluating them for whole antibody production. Transfection parameters investigated include seeding cell density, the concentration of the transfection reagent and DNA, complexation time, temperature, and volume, as well as culture parameters such as medium replacement, serum deprivation, use of cell maintenance antibiotic, incubation temperature, medium volume, post-transfection harvest day and common nutrient supplements.

Results

Using 2 mL adherent HEK293E cell culture transfections with 25 kDa linear Polyethylenimine in the most optimised parameters, we demonstrated a ~ 2-fold production increase for light chain alone and for whole antibody production reaching 536 and 49 μg respectively in a cost-effective manner. With the addition of peptone, κ light chain increased by ~ 4-fold to 1032 μg while whole antibody increased to a lesser extent by ~ 2.5-fold to 51 μg, with benefits potentially for antibodies limited by their light chains in production.

Conclusions

Our optimised findings show promise for a more efficient and convenient antibody production method through transfection and culture optimisations that can be incorporated to scale up processes and with potential transferability to other mammalian-based recombinant protein production using HEK293E cells.

Statement of Significance

Recombinant antibody production is crucial for antibody research and development. Systematically investigating transfection and culture parameters such as PEI/DNA concentrations, complexation time, volume, and temperature, supplements, etc., we demonstrated a ~ 4-fold light chain alone production increase to 1032 μg and a 2.5-fold whole antibody production increase to 51 μg from 2 mL transfections.

High-Throughput Generation of Bipod (Fab × scFv) Bispecific Antibodies Exploits Differential Chain Expression and Affinity Capture

Generation of bispecific antibodies (BsAbs) having two unique Fab domains requires heterodimerization of the two heavy chains and pairing of each heavy chain with its cognate light chain. An alternative bispecific scaffold (Bipod) comprising an scFv and a Fab on a heterodimeric Fc eliminates the possibility of light chain mispairing. However, unpredictable levels of chain expression and scFv-induced aggregation can complicate purification and reduce the yield of desired Bipod. Here, we describe a high-throughput method for generation of Bipods based on protein A and CH1 domain affinity capture. This method exploits over-expression of the scFv chain to maximize heterodimer yield. Bipods purified by this method have purity suitable for cell-based functional assays and in vivo studies.

Production of Autoreactive Heavy Chain-Only Antibodies in Systemic Lupus Erythematosus

Systemic lupus erythematosus (SLE) is characterized by the overproduction of high-affinity autoreactive antibodies. Here, we show that more than 65.8% of 222 recombinant antibodies derived from 8 SLE patients can be secreted as heavy chain-only antibodies (HCAbs) when expressed in HEK-293T cells. The secretion of HCAbs follows the conventional endoplasmic reticulum-Golgi apparatus pathway, despite triggering a weaker unfolded protein response (UPR). Many of the purified SLE HCAbs remain autoreactive and have an even higher affinity for dsDNA, Sm, nucleosome, and cardiolipin than HCAbs from healthy individuals. Extended analyses of the CDR3 region and the heavy chain variable (VH) region of HCAb F3 show that the VH region is responsible for IgH secretion, while the CDR3 region determines its reactivity. Such a high frequency of HCAb secretion cannot fully concur with our current understanding of antibody assembly and secretion. The presence of a large proportion of autoreactive HCAbs in SLE reveals a novel mechanism for the generation of autoreactive antibodies in lupus.

IgG light chain-independent secretion of heavy chain dimers: consequence for therapeutic antibody production and design

Rodent monoclonal antibodies with specificity towards important biological targets are developed for therapeutic use by a process of humanisation. This process involves the creation of molecules, which retain the specificity of the rodent antibody but contain predominantly human coding sequence. Here, we show that some humanised heavy chains (HCs) can fold, form dimers and be secreted even in the absence of a light chain (LC). Quality control of recombinant antibody assembly in vivo is thought to rely upon folding of the HC C_H1 domain. This domain acts as a switch for secretion, only folding upon interaction with the LC C_L domain. We show that the secreted heavy-chain dimers contain folded C_H1 domains and contribute to the heterogeneity of antibody species secreted during the expression of therapeutic antibodies. This subversion of the normal quality control process is dependent on the HC variable domain, is prevalent with engineered antibodies and can occur when only the Fab fragments are expressed. This discovery will have an impact on the efficient production of both humanised antibodies and the design of novel antibody formats.

Composition, Assembly, and Trafficking of a Wheat Xylan Synthase Complex

Xylans play an important role in plant cell wall integrity and have many industrial applications. Characterization of xylan synthase (XS) complexes responsible for the synthesis of these polymers is currently lacking. We recently purified XS activity from etiolated wheat (Triticum aestivum) seedlings. To further characterize this purified activity, we analyzed its protein composition and assembly. Proteomic analysis identified six main proteins: two glycosyltransferases (GTs) TaGT43-4 and TaGT47-13; two putative mutases (TaGT75-3 and TaGT75-4) and two non-GTs; a germin-like protein (TaGLP); and a vernalization related protein (TaVER2). Coexpression of TaGT43-4, TaGT47-13, TaGT75-3, and TaGT75-4 in Pichia pastoris confirmed that these proteins form a complex. Confocal microscopy showed that all these proteins interact in the endoplasmic reticulum (ER) but the complexes accumulate in Golgi, and TaGT43-4 acts as a scaffold protein that holds the other proteins. Furthermore, ER export of the complexes is dependent of the interaction between TaGT43-4 and TaGT47-13. Immunogold electron microscopy data support the conclusion that complex assembly occurs at specific areas of the ER before export to the Golgi. A di-Arg motif and a long sequence motif within the transmembrane domains were found conserved at the NH2-terminal ends of TaGT43-4 and homologous proteins from diverse taxa. These conserved motifs may control the forward trafficking of the complexes and their accumulation in the Golgi. Our findings indicate that xylan synthesis in grasses may involve a new regulatory mechanism linking complex assembly with forward trafficking and provide new insights that advance our understanding of xylan biosynthesis and regulation in plants.

How antibodies fold

B cells use unconventional strategies for the production of a seemingly unlimited number of antibodies from a very limited amount of DNA. These methods dramatically increase the likelihood of producing proteins that cannot fold or assemble appropriately. B cells are therefore particularly dependent on 'quality control' mechanisms to oversee antibody production. Recent in vitro experiments demonstrate that Ig domains have evolved diverse folding strategies ranging from robust spontaneous folding to intrinsically disordered domains that require assembly with their partner domains to fold; in vivo experiments reveal that these different folding characteristics form the basis for cellular checkpoints in Ig transport. Taken together, these reports provide a detailed understanding of how B cells monitor and ensure the functional fidelity of Ig proteins.

Novel mechanism of G-CSF refractoriness in patients with severe congenital neutropenia

Severe congenital neutropenia (SCN) is a rare disease diagnosed at or soon after birth, characterized by a myeloid maturation arrest in the bone marrow, ineffective neutrophil production, and recurrent infections. Most patients respond to treatment with granulocyte colony-stimulating factor (G-CSF), and the majority harbor mutations in the neutrophil elastase gene. In the subset of patients with SCN transforming to acute myeloid leukemia (AML), mutations that truncate the cytoplasmic tail of the G-CSF receptor (G-CSFR) have been detected. Here, we report a novel mutation in the extracellular portion of the G-CSFR within the WSXWS motif in a patient with SCN without AML who was refractory to G-CSF treatment. The mutation affected a single allele and introduced a premature stop codon that deletes the distal extracellular region and the entire transmembrane and cytoplasmic portions of the G-CSFR. Expression of the mutant receptor in either myeloid or lymphoid cells was shown to alter subcellular trafficking of the wild-type (WT) G-CSFR by constitutively heterodimerizing with it. WT/mutant G-CSFR heterodimers appeared to be retained in the endoplasmic reticulum and/or Golgi and accumulate intracellularly. These findings together with 2 previous case reports of extracellular mutations in the G-CSFR in patients with SCN unresponsive to G-CSF suggest a common mechanism underlying G-CSF refractoriness.

Novel NEMO/IkappaB kinase and NF-kappa B target genes at the pre-B to immature B cell transition

The IkappaB kinase (IKK) signaling complex is responsible for activating NF-kappaB-dependent gene expression programs. Even though NF-kappaB-responsive genes are known to orchestrate stress-like responses, critical gaps in our knowledge remain about the global effects of NF-kappaB activation on cellular physiology. DNA microarrays were used to compare gene expression programs in a model system of 70Z/3 murine pre-B cells versus their IKK signaling-defective 1.3E2 variant with lipopolysaccharide (LPS), interleukin-1 (IL-1), or a combination of LPS + phorbol 12-myristate 13-acetate under brief (2 h) or long term (12 h) stimulation. 70Z/3-1.3E2 cells lack expression of NEMO/IKKgamma/IKKAP-1/FIP-3, an essential positive effector of the IKK complex. Some stimulated hits were known NF-kappaB target genes, but remarkably, the vast majority of the up-modulated genes and an unexpected class of repressed genes were all novel targets of this signaling pathway, encoding transcription factors, receptors, extracellular ligands, and intracellular signaling factors. Thirteen stimulated (B-ATF, Pim-2, MyD118, Pea-15/MAT1, CD82, CD40L, Wnt10a, Notch 1, R-ras, Rgs-16, PAC-1, ISG15, and CD36) and five repressed (CCR2, VpreB, lambda5, SLPI, and CMAP/Cystatin7) genes, respectively, were bona fide NF-kappaB targets by virtue of their response to a transdominant IkappaBalphaSR (super repressor). MyD118 and ISG15, although directly induced by LPS stimulation, were unaffected by IL-1, revealing the existence of direct NF-kappaB target genes, which are not co-induced by the LPS and IL-1 Toll-like receptors.

Fate of the mutated IgG2 heavy chain: lack of expression of mutated membrane-bound IgG2 on the B cell surface in selective IgG2 deficiency

IgG2 deficiency is clinically characterized by sinopulmonary infections caused by pneumococcus and Hemophilus. We reported homozygous one-base insertion (1793insG) in the C(gamma)2 gene in two Japanese siblings in whom serum IgG2 levels were under detection limits. The 1793insG was present in exon 4, just upstream from the alternative splice site for M exons; the result being a complete amino acid change in transmembrane and cytosolic parts of membrane-bound gamma2 heavy chain (m gamma 2HC). To determine why this mutation caused selective and complete IgG2 deficiency, we constructed expression vectors of normal and mutant membrane-bound chimeric IgG heavy chain cDNAs. Stable transformants, Ag8N-L and Ag8M-L, expressing either normal and mutant chimeric IgG heavy chain with light chain respectively were obtained using P3X63Ag8653 as recipient cells. Of the Ag8N-L, 22.1% were surface IgG+; however, none of the Ag8M-L were surface IgG+. Addition of an anti-human IgG antibody induced cell death of Ag8N-L and we considered that the expressed chimeric IgG protein on Ag8N-L might function as the Ig receptor for signal transduction. However, Ag8M-L did not express mutant IgG on its surface nor did it secrete this mutant into culture medium. The mutant chimeric IgG protein was rapidly degraded within Ag8M-L. Thus, the mutated IgG2 heavy chain in our patient could not be expressed on the cell surface because of loss of the transmembrane domain and the evolutionally conserved cytoplasmic domain. In humans, B cells expressing surface IgG are indispensable for secretion of IgG.

Possible stage-specific function of NF-kappaB during pre-B cell differentiation

Lipopolysaccharide (LPS)-induced differentiation of the murine pre-B cell line 70Z/3 is a model for pre-B to B cell differentiation and has been used to show that the transcription factor NF-kappaB is essential to induce the expression of the Ig kappa gene. We have investigated the mechanism involved in late stages of the process when all cells have reached a more mature B phenotype, i.e. beyond 48 up to 96 h of LPS treatment. NF-kappaB binding activity was induced at early times by LPS treatment, but its DNA binding activity disappeared after 84 h of LPS treatment. Accumulation of IkappaB alpha protein in the nucleus correlated with the disappearance of NF-kappaB activity at 72, 84 and 96 h, and treatment of nuclear extracts of 72-96 h LPS-treated cells with Na-deoxycholate restored NF-kappaB binding activity. The data indicate that NF-kappaB, while important to initiate the process of Ig kappa gene transcription in 70Z/3 pre-B cells, is no longer required for its maintenance in differentiated 70Z/3 cells.

Heavy chain dimers as well as complete antibodies are efficiently formed and secreted from Drosophila via a BiP-mediated pathway

We have constructed a stable Drosophila cell line co-expressing heavy chain (HC) and light chain (LC) immunoglobulins of a humanized monoclonal antibody (mAb) that recognizes the F antigen of respiratory syncytial virus (Tempest, P. R., Bremmer, P., Lambert, M., Taylor, G., Furze, J. M., Carr, F. J., and Harris, W. J. (1991) Bio/Technology 9, 266-271. These cells efficiently secrete antibody with substrate binding activity indistinguishable from that produced from vertebrate cell lines. Significantly, the Drosophila homologue of the immunoglobulin binding chaperone protein (BiP), hsc72, was found to interact specifically with the immunoglobulin HC in an ATP-dependent fashion, similar to the BiP-HC interaction known to occur in vertebrate cells. This is, in fact, the first substrate ever shown to interact specifically with Drosophila hsc72. Most surprisingly, expression of heavy chains in the absence of LC led to the efficient secretion of heavy chain dimers. Moreover, this secretion occurred in association with hsc72. This dramatically contrasts with what is seen in vertebrate cells where in the absence of LC, HC remains sequestered inside the cell in stable association with BiP. Our results clearly suggest that Drosophila BiP can substitute for its mammalian counterpart and chaperone the secretion of active IgG. However, the finding that Drosophila BiP can also uniquely chaperone heavy chain dimers indicates mechanistic differences that may relate to the evolved need for retaining immature IgGs in vertebrates.

Weak bases increase surface IgM expression in 70Z/3 B lymphoid cell line without increasing kappa gene expression

A murine B lymphoid cell line, 70Z/3, has proven to be a useful model to study the developmental transition from membrane immunoglobulin (mIgM)-negative to mIgM-positive B cells. 70Z/3 cells have normal intracellular levels of mu heavy chain but kappa mRNA is transcribed at a very low rate. Because both kappa and mu chains are required for IgM to localize on the external membrane the cells remain essentially mIgM negative. A number of biologic agents can stimulate kappa mRNA transcription leading to increased surface IgM expression. Pharmacologic agents which cause cytoplasmic alkalinization have also been shown to increase the level of surface IgM expression when assessed by FITC-labeled anti-IgM antibodies. This increase has been used to argue that cytoplasmic alkalinization is sufficient to stimulate 70Z/3 differentiation but previous studies did not measure kappa mRNA levels or rate of synthesis. In this study we set out to determine whether the increased FITC staining that follows cytoplasmic alkalinization results from increased surface IgM expression and whether that increase results from activation of kappa mRNA transcription. When cells were treated with methylamine, NH4Cl, or monensin there was transient cytoplasmic alkalinization. The levels of surface IgM expression were measured by flow cytometry of cells stained with FITC-labeled anti-mouse kappa and mu. Using this criterion, a 24-hr treatment with either weak bases or monensin increased the level of surface IgM. The increase in fluorescence was not the result of nonspecific binding or uptake by the cell because there was no increase in fluorescence when methylamine-treated cells were stained with FITC-labeled antibody directed against an antigen not found on these cells. Iodination of surface proteins confirmed that the increase in fluorescence was the result of increased levels of IgM protein on the cell surface. However, exposure of the cells to weak bases or monensin caused no increase in either the steady-state level of kappa light chain mRNA, or in the level of kappa protein. We conclude that the transient alkalinization is not sufficient to induce differentiation of the 70Z/3 cells.

Slow response variant of the B lymphoma 70Z/3 defective in LPS activation of NF-kappa B

The mouse B cell lymphoma 70Z/3 is membrane immunoglobulin M (mIgM) negative, but treatment of the cells with bacterial lipopolysaccharide (LPS) induces the expression of kappa (kappa) light chain synthesis, and the cells become mIgM+. In wild type cells, this reaction is maximal after 24 h; we have isolated a variant, 1B8, which becomes mIgM+ only after a more prolonged incubation with LPS. This delayed response results from a reduced rate of accumulation of (kappa) mRNA and protein. The transcription factor, NF-kappa B is present in the cytoplasm of both the wild type and the variant cells in its inactive form. The delay in kappa expression is correlated with the failure of NF-kappa B to be activated and translocated to the nucleus. Although NF-kappa B cannot be activated by LPS, it can be activated by treatment with phorbol ester (PMA). In contrast to the clear defect in NF-kappa B, LPS treatment of 1B8 cells causes the octamer-binding factor OTF-2 to increase normally. We conclude that the defect in 1B8 cells is in an early part of the LPS activation pathway, prior to the activation of NF-kappa B, but after the signal for OTF-2 induction. The phenotype of 1B8 demonstrates that an increase in OTF-2 alone is sufficient to cause a large increase in kappa transcription in 70Z/3 cells, but that without NF-kappa B, the response is slow to develop. In this view, NF-kappa B functions to facilitate kappa transcription and to speed its rate of increase, but is not required for the long-term response of 70Z/3 cells to LPS.

A modified storage protein is synthesized, processed, and degraded in the seeds of transgenic plants

In vitro mutagenesis was used to supplement the sulfur amino acid codon content of a gene encoding β-phaseolin, a Phaseolus vulgaris storage protein. The number of methionine codons in the phaseolin gene was increased from three to nine by insertion of a 45 base pair (bp) synthetic duplex. Either modified or normal phaseolin genes were integrated into the genome of tobacco plants through Agrobacterium tumefaciens-mediated transformation. Although similar levels of phaseolin RNA are detected in seeds of plants transformed with either the normal or modified (himet) gene, the quantity of himet protein is consistently much lower than normal β-phaseolin. Himet phaseolin is expressed in a temporal- and organ-specific fashion, and is N-glycosylated and assembled into trimers in the manner of normal phaseolin. After germination, both types of phaseolin are hydrolyzed, but the himet protein is more quickly degraded. Electron microscopic immunocytochemical observations of developing seeds indicate that the himet protein is primarily localized in the endoplasmic reticulum (ER) and in Golgi apparatus secretion vesicles. Himet phaseolin is absent from protein storage vacuoles, termed protein bodies, where normal phaseolin is deposited in transgenic tobacco. We interpret the immunocytochemical data to indicate that himet phasolin is transported through the ER and Golgi apparatus and is then degraded in Golgi secretion vesicles or the protein bodies.

Molecular analysis of immunoglobulin expression in variants of murine B lymphoma, 70Z/3

We have used a genetic approach to study the differentiation of B lymphocytes. Our model system is the induction of membrane immunoglobulin M-positive (mIgM+) cells in the murine B cell tumor, 70Z/3 by three extracellular mediators: lipopolysaccharide (LPS), supernatants from concanavalin A-stimulated rat spleen cells (CAS) and gamma interferon (IFN). The wild-type 70Z/3 cells synthesize constitutively the mu immunoglobulin heavy chain, but the kappa (kappa) light chain is expressed at extremely low levels. Treatment with these three inducers markedly increases kappa synthesis and allows the expression of IgM on the cell surface. We have selected variants which respond aberrantly to LPS and have analyzed their responses to the other inducers. We have analyzed mIgM expression, mu and kappa mRNA and protein levels. Our results show that the level of kappa mRNA is the most sensitive indicator of cellular response to an inducer. The independence of the variant phenotypes demonstrates that the pathways are not identical.

Developmentally regulated rearrangement and expression of genes encoding the T cell receptor-T3 complex

Human leukemic cells corresponding to the earliest identifiable stages of intrathymic T cell differentiation lack cell surface expression of the T cell receptor(TCR alpha/beta)-T3 complex but transcribe TCR beta mRNA from either germ-line configuration (1/13) or partially (DJ) or fully (VDJ) rearranged (12/13) genes. These cells do not produce TCR alpha mRNA, but do contain T3 delta and T3 epsilon mRNA and accumulate T3 polypeptides, primarily in the perinuclear envelope. Equivalent normal T cells isolated from thymus have a predominantly germ-line configuration of TCR beta but contain intracellular T3 proteins. T3 gene expression is therefore a very early event in T cell differentiation. TCR alpha chain production appears to be the limiting maturation-linked event in the transport, assembly, and cell surface membrane insertion of the TCR alpha/beta-T3 complex.

Cytokine effects on a B-cell line

The murine pre-B-like cell line, 70Z /3, has been shown to undergo differentiation in response to supernatants derived from concanavalin-A (Con A) stimulated rat spleen cells or ultraviolet light treated P388D1 macrophage cell line. After culture for 24 h with supernatant factors, normally membrane IgM negative 70Z /3 cells are induced to synthesize light chain and become high level membrane IgM expressors . The cytokine (s) responsible for inducing 70Z /3 cells to differentiate is different from interleukin-2 (IL-2), B-cell growth factor (BCGF), colony-stimulating factor (CSF) I and II, but one of the cytokines is either very similar or identical to interleukin-1 (IL-1). These results demonstrate that IL-1 or a molecule with very similar physical properties can act directly on a B-cell line and thus probably also on normal B cells to influence differentiation.

LPS-nonresponsive variants of mouse B cell lymphoma, 70Z/3: isolation and characterization

We have used a genetic approach to study the differentiation of B lymphocytes. The cultured murine cell line 70Z/3 resembles pre-B cells in containing the heavy chain of the immunoglobulin IgM, mu, as an internal protein in the absence of light chain, L. However, overnight incubation with the B cell mitogen lipopolysaccharide (LPS) induces the cells to mature to a B lymphocyte-like state by the induction of L chain synthesis and the appearance of IgM on the cell surface. We have used immunoselection against surface-bound IgM to isolate LPS uninducible variants of 70Z/3. These fall into two complementation groups, LPS A and LPS B. LPS A variants predominated and were found at a frequency of 1/1200. These cells were completely unresponsive to LPS. LPS B was represented by a single variant in which a subset of cells was induced to display wild-type levels of membrane-bound IgM, and the proportion of induced cells increased with prolonged incubation with LPS. We detected no structural defects in either variant group, but LPS B may represent a defect in the decision to differentiate in response to LPS.