HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Fang, T. Articles by Roman, C. A. J. Search for Related Content PubMed PubMed Citation Articles by Fang, T. Articles by Roman, C. A. J.

Conventional and Surrogate Light Chains Differentially Regulate Ig µ and Dµ Heavy Chain Maturation and Surface Expression¹

Department of Microbiology and Immunology and Morse Institute for Molecular Genetics, State University of New York-Downstate Medical Center, Brooklyn, NY 11203

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion References

 
Positive selection of precursor (pre-) B cells by Ig membrane µ H chains (µm HC) and counterselection mediated by the truncated HC Dµ depend on the ability of each HC to form a pre-B cell receptor (pre-BCR) signaling complex with the surrogate L chain (SLC) components 5 and Vpre-B. To better understand how pre-BCR signaling output is determined by its Ig components and the SLC, we investigated the regulation of pre-BCR surface expression and HC secretory maturation in a new nonlymphoid system. We took this approach as a means to distinguish B-lineage-specific effects from pre-BCR-intrinsic properties that may influence these aspects of pre-BCR homeostasis necessary for signaling. As in pre-B cells, the SLC in nonlymphoid cells supported only a limited degree of µm HC maturation and low pre-BCR surface expression levels compared with conventional LCs, indicating that this was due to an intrinsic property of the SLC. We identified the non-Ig region of 5 as harboring the restrictive activity responsible for this phenotype. This property of 5 was also evident with Dµ, but the overall SLC- and L chain-dependent requirements for Dµ maturation and surface expression were markedly different from those for µm. Surprisingly, Dµ was modified in an unusual manner that was only dependent on Vpre-B. These results establish a novel function of 5 in limiting surface pre-BCR levels and reveal biochemical properties of Ig molecules that may underlie the diverse consequences of pre-BCR signaling in vivo by different HCs.

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

 
The expression of (Ig) H and L chains (HC³ and LC) as components of signaling complexes on the B cell surface is required to direct early B cell development through discreet stages of differentiation, serving as a quality control mechanism to monitor the success of the rearrangement process at each step (reviewed in Ref. 1). For most B cells, the µ HC is rearranged first, and the membrane isoform (µm) must associate with the invariant surrogate LC (SLC) components 5 and Vpre-B to form a complex known as the precursor (pre)-B cell receptor (pre-BCR). The pre-BCR signal directs proliferation and the reprogramming of gene expression patterns that allow progenitor-B cells to differentiate into pre-B cells. Also required in the pre-BCR are the integral membrane proteins Ig and Ig, the cytoplasmic domains of which link the Ig components to signal transduction pathways (2). Although necessary for the pre-B transition, the SLC does not guide B cell differentiation beyond the pre-B cell stage; SLC expression is down-regulated, and conventional LC rearrangement is stimulated. Only if an LC protein is produced that can associate with µm will the new signaling complex formed (the BCR) be able to direct further maturation. Failure to produce a pre-BCR due to a lack of 5, the inability to produce a µm HC, or deficiencies in Ig lead to a developmental block at the progenitor-B cell stage and thus immunodeficiency in humans and mice (3, 4, 5). In this way developmental progression in vivo is absolutely dependent upon the productive rearrangement of HC and LC genes and their association into signaling-competent complexes.

Despite its critical importance, the molecular mechanisms that underlie SLC function in pre-BCR signaling are not fully understood. Sequence homology to conventional LCs, and biochemical and genetic evidence support a minimal model for SLC function in which Vpre-B (a V_L-like protein) and 5 (a JC_L-like protein) form a LC-like complex that is required to release HCs from the endoplasmic reticulum (ER) by displacing the ER-resident folding chaperone BiP (Ig binding protein) (6, 7, 8) from C_H1 and to escort HCs to the cell surface in conjunction with Ig (9, 10), where the assembled pre-BCR can engage signal transduction pathways (11). Maturation of µm proteins is a consequence of their transport through the secretory pathway, which leads to the modification of N-linked polysaccharides on µm from high mannose ER forms to complex Golgi-dependent forms (12, 13). In support of this model, 5 and Vpre-B associate in the absence of HCs (14, 15, 16, 17), and low amounts of surface µm-SLC complexes and trans-Golgi-modified, mature HCs can be detected in normal and transformed pre-B cells (18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31). Finally, only HCs that can associate with both SLC components contribute to the adult pre-B cell pool (32, 33, 34). Based on this activity, a quantitative model for SLC function in pre-BCR-dependent positive selection has been proposed in which the limiting parameter for proliferative expansion of a pre-B cell clone is the number of pre-BCRs on its surface, proportional to the ability of the HC to associate with the SLC (35).

However, unlike LCs, 5 and Vpre-B are products of separate, nonpolymorphic, nonrearranging genes and contain non-Ig-like regions (36, 37, 38, 39) that may have a specialized function. Interestingly, part of the non-Ig region of human 5/14.1 has the unusual property of inhibiting 5 folding (40), suggesting that it may also have an important role in controlling pre-BCR homeostasis. In contrast to the BCR, surface pre-BCR levels are very low, but it has not been fully resolved whether this is a pre-BCR-intrinsic property or a B-lineage-specific effect. Given the low amounts of surface pre-BCRs, it is also possible that the SLC may not only have a quantitative role by allowing surface pre-BCR transport, but may also have a qualitative role in activating surface pre-BCR signaling (41). Although the molecular nature of the triggering mechanism is not fully understood, the developmental pre-BCR signal may be the result of an inherent constitutive activity, i.e., that it does not require cross-linking by a putative pre-BCR-ligand (2, 11, 42, 43, 44, 45). Even so, an escort activity of the SLC cannot account for all diverse consequences of HC-dependent signaling at the pre-B stage. For example, counterselection of precursors that express Dµ, a truncated HC product of DJ recombination that lacks a V_H region (46, 47), also requires 5 (48). Therefore, although Dµ can form a pre-BCR-like signaling complex (31, 49, 50), SLC association is not sufficient to promote positive selection by this HC in vivo (51). In contrast, HC allelic exclusion remains enforced in the absence of 5 and conventional LCs at the pre-B stage (33), implying that this aspect of HC-dependent signaling may not require any HC maturation and/or surface expression or may use other factors for this function. Therefore, a better understanding of the role of the SLC in the pre-BCR is necessary to elucidate the molecular basis for how different HC components within pre-BCR complexes direct diverse signaling outputs.

To address these issues, we have focused on the regulation of HC maturation and pre-BCR cell surface expression because they are fundamentally linked aspects of pre-BCR homeostasis necessary for the formation of signaling competent complexes, and surface pre-BCR density may be a primary determinant controlling pre-BCR signaling. We examined these processes in nonlymphoid cells as a way to identify pre-BCR-intrinsic vs B cell-specific properties that may influence pre-BCR homeostasis. Our results indicate that the SLC has fundamentally distinct and unique properties that distinguish it from conventional LCs in directing HC maturation and cell surface expression and identify the sequences within the SLC responsible for these differences. Furthermore, we show that µm and Dµ differ in their molecular requirements for surface expression and maturation, which may underlie the diverse biological effects of these HCs in vivo.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

 
Plasmid construction

All cDNAs described were subcloned into pEBB (52); further details of plasmid construction are available upon request. All cDNA constructs contained the natural initiating methionine. The integrity of all cDNAs created by PCR was confirmed by sequence analysis.

H chains. µm and µm1m2 (from a C57BL/6 anti-NP Ab; V_H186.2-DFL16.1-J_H2), were derived from pJ (53) (gift from Dr. S. Pillai, Massachusetts General Hospital, Boston, MA). The Dµ cDNA was obtained by RT-PCR of total RNA from the Dµ-expressing Abelson line 300-19P (46) (gift from Dr. B. Malynn, Harvard Medical School, Boston, MA). The cDNA of the human Ig HC expressed by IGSA transgenic mice (herein referred to as TGSA, 64) was amplified by RT-PCR from total RNA isolated from MOPC-21Tg⁺TGSA⁺RAG1^+/- mouse spleen (54). TGSAm1m2 is a chimeric protein in which the human extracellular domains were fused contiguously to the transmembrane domain/C-terminal tail of µm1m2 by overlap PCR via the common GFENLW peptide sequence.

L chains. 1 cDNA was obtained from pHCMV-1 (55) (gift from Dr. M. Neuberger, Medical Research Council, Cambridge, U.K.). The MOPC-21 cDNA was isolated by RT-PCR of total RNA from MOPC-21Tg⁺TGSA⁺RAG1^+/- mouse spleen. 5 cDNA was amplified from the plasmid pZ183-1A (36). JC_L proteins were created by fusing the leader sequence of 1 (aa 1–19) to the first amino acid of the respective J region via overlap PCR. The Vpre-B-1 cDNA was amplified from pZ121 (37). Vpre-BC was made by overlap PCR to generate a stop codon after amino acid 122 of Vpre-B-1.

Other expression constructs. Ig and Ig cDNAs were amplified by PCR from the plasmids pCL2-T3 (56) and pTZ18-1A94 (57), respectively, using oligonucleotide primers that restored to each their normal translation initiation sequence. pEBB-green fluorescent protein (GFP) was created by subcloning the GFP insert from MSCV-hGFP (gift from Dr. J. Jacob, Emory University, Atlanta, GA) into pEBB.

Cell culture and transient transfections

Human embryonic kidney (HEK) epithelial 293 cells were maintained in DMEM supplemented with 10% heat-inactivated FCS, 50 U/ml penicillin, and 50 µg/ml streptomycin. A total of between 5 and 15 µg of pEBB plasmid constructs were introduced into HEK 293 cell cultures (seeded at 2 x 10⁶ cells/60-mm dish) by calcium phosphate-mediated transfection as previously described (58). LC and m1m2 HC plasmids were transfected at a 1:1 ratio; LC, wild-type HC, Ig, and Ig plasmids were transfected at a 1:1:0.5:0.5 ratio; 5, Vpre-B, and HC plasmids were transfected at a 0.7:0.3:1 ratio. pEBB-GFP (0.5 µg) was included in each transfection mix.

Cell extracts and Western blots

HEK 293 cells were harvested 2 days post-transfection in ice-cold PBS supplemented with 10 mM EDTA. Approximately 2 x 10⁵ cells were removed for FACS analysis (below). Total Nonidet P-40 cell lysates were made by disruption of cells in 0.5% Nonidet P-40 buffer (14, 59) at 2–3 x 10⁶ 293 cells/ml; nuclei were pelleted by centrifugation for 3 min at 400 x g. For direct Western blot analyses, proteins in equal volumes of Nonidet P-40 lysates were resolved by SDS-PAGE (4–6 x 10⁴ cells/lane), transferred to Immobilon P membranes, and processed according to the manufacturer, using Ponceau S to confirm transfer.

The Abs used were goat anti-human µ HC, goat anti-mouse , and goat anti-mouse (Southern Biotechnology Associates, Birmingham, AL) and rabbit anti-mouse µ HC (The Jackson Laboratory, Bar Harbor, ME). Alkaline phosphatase and HRP-conjugated secondary Ab reagents were obtained from Sigma (St. Louis, MO). Protein deglycosylation with peptide:N-glycosidase F (PNGase F) and endoglycosidase H (Endo H; New England Biolabs, Beverly, MA) was performed according to that manufacturer’s protocols.

Immunoprecipitations

Equal volumes of total Nonidet P-40 cell lysates (0.5–1 x 10⁶ cells) were first precleared by a 1-h incubation with normal goat (Caltag Laboratories, South San Francisco, CA) or normal rabbit serum followed by 1 h with protein A beads at 4°C (Roche, Indianapolis, IN). Cleared supernatants were incubated with 5 µg of primary Ab overnight at 4°C, followed by protein A capture for 1 h. Pellets were washed three times in lysis buffer at 4°C for a total of 20–30 min, resuspended in 100 µl of SDS-PAGE sample buffer, and boiled; between 20 and 50 µl were analyzed. For radioactive immunoprecipitations, 2 x 10⁶ 293 cells were methionine-starved for 1 h before the addition of 0.05–0.10 mCi [³⁵S]methionine (ICN Pharmaceuticals, Irvine, CA)-supplemented medium. After 4 h, the culture medium was collected; the cells were washed in situ, then harvested and processed as described above. SDS-polyacrylamide gels were soaked in Amplify (Amersham, Arlington Heights, IL) and dried before exposure to film.

Lectin binding analyses

Lectin enrichment for glycoproteins was performed in a manner modified from Brouns et al. (30). Agglutinins from Ricin communis (RCA) and Pisum savatum (PSA) coupled to agarose beads (Sigma) were equilibrated in Nonidet P-40 lysis buffer. Equal volumes of total Nonidet P-40 cell lysates (0.5–1 x 10⁶ cells) were incubated with a 5-µl packed bead volume of RCA-agarose cells for 2 h in lysis buffer at 4°C. Supernatants were decanted to a new tube to which PSA-coupled agarose beads were added (5-µl packed bead volume) and incubated for 2 h at 4°C. RCA beads were immediately washed three times in 1 ml of ice-cold lysis buffer for between 30 and 45 min total wash time and resuspended in 40–100 µl of 1x SDS sample buffer. The PSA beads were similarly processed. One-quarter to one-half of the material recovered in sample buffer was analyzed by Western blot.

Abs and flow cytometry

Approximately 2 x 10⁵ HEK 293 cells were incubated in FACS buffer (PBS plus 1% (v/v) FCS, 1% (v/v) normal goat serum, and 0.1% (w/v) NaN₃) before the addition of primary Abs. The following Abs were used: goat anti-mouse µ-FITC (Caltag Laboratories and The Jackson Laboratory), goat anti-mouse µ-PE Fab₂ (Caltag Laboratories), and goat anti-mouse or -FITC or -PE (Southern Biotechnology Associates). Abs from different suppliers yielded equivalent results. Cells were stained in 50 µl of a 100-fold dilution of Ab in FACS buffer for 1 h on ice, followed by two washes in 100 µl of FACS buffer before analysis on a FACScan (BD Biosciences, Mountain View, CA); 10⁴ events were analyzed using CellQuest.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

 
Surrogate and conventional LCs ((S)LC) promote maturation of µm in HEK 293 cells

To establish the viability of examining pre-BCR and BCR homeostasis in a nonlymphoid system, we first determined whether LC- and Ig-dependent regulation of µm secretory maturation could be recapitulated by transient transfection of BCR components into the HEK 293 cell line (58). This epithelial cell line was chosen because cells can be transiently transfected at very high efficiency and lack components of B cell-specific Ig signaling pathways (A. Rashkit and C. Roman, unpublished observations). We used a murine µm HC containing a J558 V_H region (53). Western blot analyses of total Nonidet P-40 cell lysates from HEK 293 cells transfected with µm alone, plus Ig, or plus either 1 (55) or MOPC21 LCs (60) revealed that µm migrated primarily as a single 95-kDa band (Fig. 1A). Treatment of these extracts with Endo H, which only removes high mannose N-linked sugars, or PNGase F, which removes all N-linked moieties, indicated that the vast majority of µm proteins contained polysaccharides of the Endo H-sensitive, N-linked, high mannose ER type. Differences in the amount of total µm visualized in each lysate (5-fold; Fig. 1A) were attributable to differences in transfection efficiency (i.e., proportion of GFP⁺ cells) and protein loading (as determined by Ponceau S staining of the membrane after transfer; data not shown).

View larger version (44K): [in this window] [in a new window]   FIGURE 1. Maturation and surface expression of an Ig HC requires LCs and Ig in HEK 293 cells. A, Western blot of µm proteins (µ) in total Nonidet P-40 cell lysates from cells transfected with Ig components (indicated above each lane), prepared as described in Materials and Methods. Un, untreated extracts. Lane 1, Control extract lacking µm. EndoH, Endo H-treated extracts. Mature HCs are Endo H resistant and remain glycosylated (gly), whereas immature ER forms are Endo H sensitive and are deglycosylated (degly) by this enzyme. Lane 10, A 5-fold dilution of the untreated extract from lane 3, the glycosylated 95-kDa µm in this sample serves to highlight the faster migration of µm proteins after removal of the N-linked sugars (degly) to 75 kDa. PNGase, PNGase F-treated extracts. Lane 10, Same as from the EndoH panel. B, Western blot of lectin-bound µm proteins isolated from extracts of transfected cells (as described in Materials and Methods). RCA, µm proteins bound to RCA; mature indicates the complex, Golgi-modified µm polypeptide, and immature the high mannose, ER µm species. PSA, µm proteins bound to PSA. C, FACS analysis of transfected HEK 293 cells for surface µm expression with FITC-conjugated goat anti-mouse Ig µ polyclonal Abs (as described in Materials and Methods). Top, Dot plots comparing cells transfected with the empty expression vector pEBB (left) and with pEBB-GFP (right), mapping GFP intensity (FL1) vs forward scatter (FSC). Cells within the indicated gate (GFPlow) were analyzed for fluorescence intensity of surface µm staining (x-axis), shown in the histograms (cell number, y-axis). Bottom, Histograms comparing surface µm levels on cells transfected with the indicated Ig components. In histogram 1, the solid gray line denotes the fluorescence profile for cells transfected with µm alone (sample corresponding to lane 1, A); the filled curve represents cells transfected with pEBB-GFP. The percentages indicate the percentages of the cells that fall within the indicated marker boundaries. The number below the percentage is the mean fluorescence value for those cells. In histograms 2–8, the filled curve represents the cells transfected with µm alone, and the solid line represents the cells transfected with µm plus the cDNAs indicated above each histogram, either with (histograms 5–8) or without (histograms 1–4) Ig. A (Un) and C are representative of three separate experiments; A (EndoH and PNGase) and B are representative of two separate experiments.

The SLC, made up of 5 and Vpre-B, also promoted Ig-dependent µm maturation, although to a lesser degree than or LCs. The SLC-dependent mature µm species was revealed only after treatment of extracts with Endo H (Fig. 1A) and had the same mobility as the major LC-dependent Endo H-resistant µm species. To independently confirm its maturity, we used lectin-coupled agarose beads to enrich for glycoproteins from these extracts based on their polysaccharide composition (30). RCA preferentially binds Golgi-modified polysaccharides that contain -galactose, whereas PSA binds to -mannose residues in high mannose moieties. The slower migrating µm species was enriched in material adsorbed to RCA-coupled beads from the extract that contained µm plus Ig and 1, and between 5- and 10-fold less was recovered from the extract with µm plus Ig and the SLC, but not from other extracts (Fig. 1B). Therefore, the SLC promoted faithful HC maturation in HEK 293 cells.

The transport of µm to the cell surface also depends on LCs and Ig (61, 62, 63); therefore, we evaluated the presence of surface µm on transfected HEK 293 cells from Fig. 1A by flow cytometry (Fig. 1C). A GFP expression vector was cotransfected to normalize for protein expression levels within transfected cells, and surface expression was compared between cells with the same GFP^low intensity. Low levels of surface µm were detected when it was transfected alone, which only slightly increased with the coexpression of either LC or Ig. In contrast, 1 or LC plus Ig expression led to a dramatic increase in the amount of surface µm detected, particularly at the highest fluorescence intensity. 1 and LC surface staining also dramatically increased when all BCR components were coexpressed (data not shown). These observations were not due to differences in µm, LC, or Ig expression (Fig. 1A and data not shown). The coexpression of SLC and Ig led to a comparatively small, but reproducible, increase in the percentage of surface µm-stained cells compared with cells transfected with µm plus Ig only or µm plus the SLCs without Ig (Fig. 1C), consistent with the lower degree of HC maturation in those extracts. Qualitatively similar FACS results were obtained from cells expressing higher GFP levels (data not shown). Therefore, the degree of HC and LC cell surface expression correlated with the degree of HC maturation in the HEK 293 system.

Vpre-B and 5 are required together to promote low levels of µm maturation

To further characterize the role of 5 and Vpre-B in HC maturation, we simplified the assay system by using a µm allele that contained mutations within the transmembrane domain rendering µm surface expression Ig independent (µm1m2) (53); µm1m2 is otherwise identical with the wild-type µm used above. µm1m2 migrated predominantly at 95 kDa when expressed by itself, whereas the coexpression of 1 led to the appearance of a slower migrating species that selectively bound RCA (Fig. 2A) and was Endo H resistant and PNGase F sensitive (Fig. 2B), confirming that N-linked polysaccharide maturation of this HC allele only required LC association, as predicted. Cotransfection of µm1m2 with 5 and Vpre-B together also promoted the formation of low levels (2–5% of the 1-dependent material) of mature HCs that were readily detectable in the RCA-bound fraction (Fig. 2, A and B); maturity of the RCA-bound HC was confirmed by its resistance to Endo H digestion (Fig. 2B). Only the Endo H-sensitive ER species was detected in RCA-bound material when the HC was expressed alone or with either Vpre-B or 5 expressed individually (Fig. 2, A and B). PSA-bound HCs from all extracts were Endo H sensitive (Fig. 2B), confirming the specificity of that lectin for high mannose proteins.

View larger version (72K): [in this window] [in a new window]   FIGURE 2. Vpre-B and 5 are both required to promote low levels of HC maturation. A, Western blot of µm1m2 proteins from total Nonidet P-40 cell lysates, RCA-bound and PSA-bound fractions, and cells transfected with the indicated Ig components. Total lane 7, One-fifth the amount of extract of that sample was loaded compared with the amount loaded in lanes 2–5. RCA lane 6, A 10-fold dilution of the 1 sample from adjacent lane 7. RCA-long, A 3–4 times longer exposure of the RCA panel; m, mature µm1m2 species; i, the immature species. PSA lane 6, Undiluted material from that sample. B, Western blot of µm1m2 proteins from total cell lysates (T) and RCA-bound (R) and PSA-bound (P) fractions from A after enzymatic deglycosylation treatment with either Endo H or PNGase F. gly and degly, Glycosylated and deglycosylated µ proteins, respectively. C, FACS analysis of cells transfected with the indicated Ig components from A. Gated GFPlow cells were analyzed for surface µm1m2 expression, shown in the histograms. In histogram 1, the solid line represents cells transfected with µm1m2 only; the filled curve shows cells transfected with pEBB plus pEBB-GFP (no HC). In histograms 2–5, the filled curve indicates µm1m2 only; the solid line shows µm plus the LC component indicated above each histogram. Percentages in the upper right corner are the percentages of cells within the indicated marker boundaries; the number under it is the mean fluorescence value for those cells. D, Autoradiogram (Autorad) of immunoprecipitated (IP) µm complexes from metabolically labeled extracts from cells transfected with the Ig components indicated, using an anti-µ Ab. E and F, Western blots for TGSA (E) or TGSAm1m2 (F) protein in total extracts (Total) or RCA-bound material (RCA) from cells transfected with the indicated Ig components. i, immature, ER-modified species; m, slower mobility, Golgi-modified mature species. F, lane 6, One-fifth the amount of extract of that sample (TGSAm1m2 plus 1) was loaded onto the gel compared with the amount loaded from the other samples (lanes 2–5). A, C, and D are representative of three separate experiments; B, E, and F are representative of two separate experiments.

SLC-dependent HC maturation was much lower than that supported by the conventional LCs with both wild-type and m1m2 HCs. To understand the basis for this difference, we evaluated SLC expression and HC association by immunoprecipitation of µm-containing complexes from [³⁵S]methionine-labeled extracts of transfected HEK 293 cells (Fig. 2D). µm1m2 immunoprecipitated as a single polypeptide species when expressed alone and with SLC components. We identified the band migrating below µm1m2 in these samples as most likely BiP, based on published and our own observations (6, 7) (data not shown). When the LC was coexpressed, the slower migrating, mature µm1m2 species became apparent, and the putative BiP band diminished. Vpre-B and 5 individually or together specifically coprecipitated with µm1m2, indicating that they could associate with the HC. The BiP band was also evident in the µm1m2 plus SLC samples, consistent with the inability of µ-SLC complexes to be efficiently released from ER; however, it was not possible to distinguish whether this was due to a failure of the SLC to compete with BiP for C_H1 binding or to recruitment of BiP by the SLC to the HC. The anti-µ Abs did not precipitate Vpre-B and 5 when expressed in the absence of the HC or when mixed postlysis with extracts containing µm1m2 alone (Fig. 2D and data not shown). Differences in band intensities for the Vpre-B, 5, and the LC paralleled differences in the amount of µm1m2 precipitated in each lane. These results suggested that the low degree of SLC-dependent HC maturation was not due to an inability of the SLC components to associate with µm1m2.

The low degree of SLC-dependent HC maturation could also have been a reflection of the folding and/or associative properties of the particular V_H region within µm rather than an intrinsic property of the SLC. To address this issue, we tested the properties of another HC, TGSA, a human HC unrelated to µm, which elicits allelic exclusion and promotes B cell development in transgenic mice (54, 64). TGSA maturation required the expression of all BCR components (Fig. 2E). RCA enrichment of mature glycoproteins revealed the slower migrating mature TGSA species when it was coexpressed with Ig and the SLC, although considerably less than with Ig and 1. As was observed with the murine µm, maximal levels of TGSA surface expression (41% in this experiment; FACS analyses not shown) were only observed in cells transfected with the full complement of BCR components and to a lesser degree with all pre-BCR components (19%); in contrast, there was no effect of LC, SLC, or Ig expression by themselves on surface TGSA expression compared with TGSA alone (all at 3–6%; data not shown). Similarly, maturation of TGSAm1m2, a TGSA protein that contained the m1m2 mutations, depended only upon the coexpression of an LC (Fig. 2F). However, RCA-enrichment for mature TGSAm1m2 proteins was necessary to reveal the mature form when the HC was expressed in the presence of both Vpre-B and 5 together, but not individually. Therefore, maturation of these diverse HCs was similarly influenced by the SLC and LCs, suggesting that the low level of SLC-dependent HC maturation was due to an intrinsic SLC inefficiency and not to an idiosyncrasy of a specific V_H region.

A SLC complex lacking the 5 non-Ig region efficiently promotes µm maturation

Two models could explain the inefficient escort capability of the SLC: 1) the SLC lacks a dominant, maturation-promoting activity present in LCs; and/or 2) the SLC contains an inhibitory domain. Such activities may control the release of SLC-HC complexes from the ER. 5 and Vpre-B contain non-Ig-like sequences not present in LCs. Importantly, a portion of the non-Ig region of human 5/14.1 was shown to limit the rate of 5 folding and thus is a strong candidate for controlling HC maturation (40). Moreover, there is 40% amino acid sequence identity between the mouse and human 5 N-terminal non-Ig regions, suggesting evolutionary conservation, although to a lesser degree than the JC domains (at 65% identity). Therefore, we tested a panel of murine 5 and Vpre-B alleles that lacked the non-Ig domains (Fig. 3A). FACS analyses revealed that deletion of the C-terminal tail of Vpre-B (Vpre-BC) did not increase cell surface expression of µm1m2 above µm1m2 expressed alone or above the low degree observed with 5 and wild-type Vpre-B (Fig. 3B). However, expression of the truncated 5 protein JC5, in which the non-Ig region was replaced with the leader sequence of 1 (Fig. 3A) led to a reproducibly small increase in surface µm1m2 expression (Fig. 3B). A more substantial increase in surface µm1m2 expression was detected when JC5 was coexpressed with Vpre-B or Vpre-BC, at levels similar to the LC. JC, a truncated LC created by replacing the leader and V sequence with the leader sequence of 1 (Fig. 3A), behaved similarly to JC5 (Fig. 3B). Western blots showed that the increase in surface µ staining correlated directly with the degree of µm1m2 maturation (Fig. 3C). We also visualized similar amounts of mature µm1m2 in RCA-enriched material in extracts from cells transfected with either JC5 or both SLC components (Fig. 3D), although at levels lower than with 1 or Vpre-B plus JC5, thus consistent with the relative degree of surface µm1m2 staining. These observations suggested that 1) the interaction of a JC_L protein, either JC5 or JC, with the HC was a limiting event in HC maturation; 2) Vpre-B provided a limiting maturation activity only in conjunction with the JC_Ls; and 3) the N-terminal non-Ig region of 5 contained an inhibitory domain (thus supporting model 2).

View larger version (52K): [in this window] [in a new window]   FIGURE 3. The non-Ig region of 5 is responsible for limiting SLC-dependent HC maturation and pre-BCR surface expression. A, Schematic illustration of wild-type and truncated 5, Vpre-B, and proteins. Ig V, J, or C regions are indicated in black (1, 5, and Vpre-B alleles) or solid gray (JC). Non-Ig regions of 5 and Vpre-B are shown in hatched gray; leader sequences of 1 and Vpre-B are shown in solid or hatched white. In JC5, the non-Ig region of 5 (hatched gray) was replaced with the leader sequence from 1 (white); in Vpre-BC, the C-terminal tail of Vpre-B was replaced with a stop codon. JC is analogous to JC5. B, FACS analysis for surface µm expression in the presence or the absence of wild-type and truncated LCs and SLC components. In histogram 1 (numbered in the lower right corner), the filled curve shows the fluorescence profile of cells transfected with pEBB-GFP; the solid line shows cells transfected with µm1m2. In histograms 2–13, the filled curve shows µm1m2 only; the solid line shows µm1m2 plus the indicated LC component(s). Percentages in the upper right corner are the percentages of cells within the indicated marker boundaries; the number under it is the mean fluorescence value for those cells. C and D, Western blots of µm1m2 proteins in total Nonidet P-40 cell lysates (Total µ), with the LC components transfected indicated above each lane. m, mature µm1m2 species; i, immature species. C, Lysates analyzed are from transfected cells shown in 3B. WT, wild-type Vpre-B; C, Vpre-BC. D, Total lysates and lectin-enriched fractions analyzed were from a separate transfection from that shown in C. B and C are representative of four separate experiments, and D is representative of two; labeled as in C.

Dµ maturation is stimulated by 1, but not or the SLC, and requires only a JC_L protein

Developing B cells that express Dµ are counterselected (47). Dµ can associate with the SLC (49), and 5 is required for Dµ counterselection (48). Based on the µm paradigm, the SLC probably facilitates the formation of surface pre-BCRs that contain Dµ by providing an escort function, resulting in Dµ maturation. However, Dµ lacks a V_H region, suggesting that it might have different requirements for maturation compared with a full-length HC, which may in part underlie its deleterious effects on B cell development. In fact, it has been hypothesized that counterselection occurs because Dµ cannot associate with LCs to provide positive developmental signals (31). To directly test these hypotheses, we evaluated the regulation of Dµ maturation and cell surface display in HEK 293 cells by the panel of full-length and truncated SLC and LCs.

Flow cytometry (Fig. 4A) indicated that a low level of Dµ could be detected on the surface of HEK 293 cells when expressed alone, and coexpression of Ig, LCs, or the SLC had little effect on surface staining. Western blot of the corresponding extracts indicated that Dµ migrated primarily at 80 kDa and was Endo H and PNGase F sensitive, demonstrating that it contained primarily ER-dependent N-linked sugar modifications (Fig. 4B). RCA enrichment of Dµ from Dµ only or Dµ plus (S)LC extracts revealed the discreet Endo H-sensitive ER form, whereas low levels of Dµ species with a more diffuse migration pattern could also be detected when Dµ was coexpressed with Ig. In contrast, coexpression of 1 and Ig led to a substantial increase in surface Dµ staining (Fig. 4A). This correlated with the appearance in total untreated extracts of the diffusely migrating Dµ forms (Fig. 4B), which were Endo H resistant due to N-linked sugar modifications, and were highly enriched in RCA-bound glycoprotein fractions. A discreet faster mobility species was observed in the PSA-bound fraction (Fig. 4B), most likely due to the trimming of high mannose polysaccharides that occurs before the addition of more complex polysaccharides, as was observed with wild-type µm (Fig. 1B). Interestingly, an additional PNGase F-resistant Dµ band was observed when it was expressed with 1 and Ig (Fig. 4B); the identity of the modification responsible is not known. These data demonstrate that Dµ can biochemically mature, and its surface expression can be influenced in a 1- and Ig-dependent manner similar to but distinct from full-length HCs.

View larger version (83K): [in this window] [in a new window]   FIGURE 4. Dµ maturation and surface expression are promoted by association with 1 and JCL proteins, but not or the SLC. A, FACS analysis for surface Dµ expression, analyzed as described in Fig. 1. In histogram 1, the filled curve indicates cells transfected with pEBB plus pEBB-GFP; the solid line shows cells transfected with Dµ only. In histograms 3–8, the filled curve indicates Dµ only; the solid line shows Dµ plus indicated LC component(s). B, Western blot for Dµ from cells analyzed in A, examining untreated (Un) total Nonidet P-40 cell lysates, lysates after Endo H or PNGase F treatment, and RCA- and PSA-bound fractions. The maturity and glycosylation status of Dµ proteins are indicated as in previous figures. C, Top panels, Western blots (W) of total untreated Nonidet P-40 cell lysates for µm1m2, Dµm1m2, or LC proteins. The primary Ab for each blot is indicated left of the panel. Lower left and right panels, Western blots (W) of immunoprecipitated complexes (IP) using the indicated primary Ab reagents for each step. The arrow indicates the mobility of immature Dµ (upper right) or LC (lower right). ns, nonspecific background band from the precipitating antiserum that serves as an indicator of protein loading. D, Western blots for µm1m2, Dµm1m2, and TGSAm1m2. LC components transfected are indicated above each lane. m and i, indicate mature and immature HC species, respectively. E, FACS analyses for µm1m2 and Dµm1m2 surface display; samples are from D. In histograms 1 and 7, the filled curve indicates cells transfected with pEBB-GFP only; the solid line shows cells transfected with either µm1m2 (panel 1) or Dµm1m2 (panel 7). In histograms 2–6, the filled curves indicate µm1m2 only; histograms 8–12 show Dµm1m2 only; the solid lines indicate µm1m2 or Dµm1m2 plus the indicated LC component. A and B are representative of two separate experiments, C (upper; Western blot) is representative of four separate experiments, C (lower; IPs) is representative of two, and D is representative of three (Dµ and µm) and two (TGSA) separate experiments.

We postulated that the key difference between the MOPC21 and 1 LCs lay in their V_L domains, in that the MOPC21 V domain, unlike the V1 domain, might require the association of a V_H region to properly fold (65). In support of this idea, the LC was not secreted, in contrast to the 1 LC, which was (data not shown). Therefore, we speculated that because Dµ lacks a V_H region, it was likely that the MOPC21 LC could not properly fold. Moreover, given its lack of a V_H region, we predicted that Dµ maturation would only require the association of a JC_L domain to displace BiP from C_H1, and the deletion of the V region from the LC should reveal this activity. Indeed, both JC5 and JC were able to support Dµ maturation to the same degree as 1 (Fig. 4D), whereas 5 and the LC did not; increases in Dµm1m2 maturation correlated with increases in its surface staining (Fig. 4E). In contrast to their effect with Dµ, JC5 and JC directed much lower (µm1m2) or undetectable (TGSAm1m2) levels of maturation compared with 1 and (Fig. 4D), with corresponding differences in µm1m2 surface staining (Figs. 3, B and C, and 4E). Differences between the JC_L proteins to support Dµm1m2 vs µm1m2 maturation were not due to differences in their ability to associate with the HCs based on immunoprecipitation analyses (data not shown). From these data we conclude that the V region dominantly interfered with the escort capability of the JC domain by preventing its association with Dµ; in contrast, 5 and JC5 both associated with Dµ, but only JC5 supported Dµ maturation. In addition, these data support the model that the presence of unpaired (or unfolded) HC or LC V regions can profoundly affect the efficiency of HC maturation.

Vpre-B promotes the formation of a novel, glycosidase-resistant Dµ species

Coexpression of Vpre-B and 5 did not significantly change surface Dµ staining or maturation above what was observed with Dµ plus Ig (Fig. 4, A and B). These observations were recapitulated with the Dµm1m2 allele (Fig. 5, A and B). Surprisingly, Western blots of cell lysates revealed low levels of a discreet, slower migrating Dµm1m2 species that was dependent on Vpre-B (Fig. 5A). Yet the appearance of this species with Vpre-B or Vpre-B plus 5 did not correlate with any increase in Dµ surface staining intensity, in contrast to the coexpression of 1 (Fig. 5B). Several observations suggested that this form was fundamentally different from the mature Dµm1m2 species dependent on 1 (Fig. 5A). The appearance of the variant species was coincident with the appearance of a Dµ form both Endo H-and PNGase F-resistant (Fig. 5C), which had a faster mobility than the predominant, heterogeneous Endo H-resistant Dµ forms induced by 1. Its mobility was slower than the minor 1-dependent PNGase F-resistant species (Fig. 5C) and partitioned to both RCA- and PSA-binding fractions (Fig. 5D). Moreover, the formation of this species with wild-type Dµ was Ig-independent (data not shown), and the modification was not due to O-linked glycosylation, as it was resistant to O-linkage-specific glycosidases and chemical deglycosylation (data not shown).

View larger version (67K): [in this window] [in a new window]   FIGURE 5. Vpre-B promotes the formation of an unusual form of Dµ that is glycosidase resistant. A, Western blot of total Nonidet P-40 lysates for Dµm1m2 expression (Total Dµ) showing the mature (m) and immature (i) forms of Dµ, indicated on the left. The Ig components transfected are indicated above each lane. The novel Dµ form is indicated by the arrow with an asterisk. B, FACS analysis for surface Dµ expression from cells in A. In histogram 1, pEBB-GFP cells are represented by the filled curve, and Dµm1m2 only is shown by the solid line. In the remaining histograms, Dµm1m2 only is shown by the filled curve, and the solid line represents the cells transfected with the Ig components indicated above each. The percentages of cells within the marker boundaries are indicated. C, Total Nonidet P-40 cell lysates from D were incubated in buffer only (-) or with Endo H (E) or PNGase F (P) and analyzed by Western blot for Dµ proteins. Glycosylated and deglycosylated forms are indicated. Asterisks indicate the Endo H- and PNGase-resistant Dµm1m2 species. D, Western blot of Dµm1m2 proteins in untreated total Nonidet P-40 cell lysates (Total Dµ) and lectin-bound fractions. Total lane 7, A 5-fold dilution of the sample in lane 2. RCA lane 7, A 5-fold dilution of the RCA-bound Dµm1m2 only sample from lane 2. Lane 8, A 5-fold dilution of the Dµm1m2 plus 1 sample in lane 6. i, High mannose form of Dµ; m with bracket, the complex mature forms; *, the Vpre-B-dependent glycosidase-resistant form. PSA lane 7, A 5-fold dilution of the PSA-bound Dµm1m2 only sample from lane 2. E, Autoradiogram (Autorad) of immunoprecipitated (IP) Dµm1m2 complexes from metabolically labeled cell extracts, using an anti-murine µ Ab. The transfected Ig components for each extract are indicated above each lane. The ER form of Dµ and the Vpre-B-dependent form (Dµ*) are indicated. A, B, and C are representative of four separate experiments; D and E are representative of two.

To establish the degree to which 5 and Vpre-B could associate with Dµ in our system, Dµ complexes were immunoprecipitated from metabolically labeled transfected cell extracts (Fig. 5E). Vpre-B and 5 coprecipitated with Dµ when expressed individually; importantly, they coprecipitated with Dµm1m2 to the same degree as JC5, which, in contrast, efficiently promoted Dµ maturation. Control experiments showed that association did not occur postlysis (data not shown). Therefore, Vpre-B could associate with Dµ in the absence of 5, and the inability of the SLC to promote Dµ maturation was not due to a failure of the SLC components to associate with Dµ.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

 
The goal of the present study was to define the forces that influence pre-BCR homeostasis to better understand how signaling pathways responsive to diverse HC-(S)LC complexes may be activated and regulated during early B cell development. We investigated these processes in the non-B cell line HEK 293 as a strategy to reveal the contributions of B cell-specific effects and/or pre-BCR-intrinsic properties to pre-BCR homeostasis. In 293 cells, maximal Ig µ HC cell surface expression and its biochemical manifestation, Golgi maturation, required all pre-BCR/BCR components. However, the SLC was a poor HC escort compared with the LCs, indicating that this was due to an intrinsic property of the SLC and not to B cell-specific effects on pre-BCR homeostasis. A mutational analysis of LCs and SLC components revealed the non-Ig region of 5 as responsible for this effect. Finally, using the panel of wild-type and mutated (S)LC molecules, we showed that the molecular requirements for (S)LC-dependent µm and Dµ maturation were fundamentally different, providing important biochemical correlates with which to understand the molecular basis for the diverse biological effects of the HC molecules on B cell development.

The SLC complex directs low levels of HC maturation

Our results directly demonstrated that the SLC complex in HEK 293 cells had an active role in promoting µm maturation and pre-BCR cell surface expression, but to levels much lower than conventional LCs, consistent with observations in primary pre-B cells and cell lines (19, 20, 21, 59). Importantly, that this occurred in a non-B cell line indicated that it was a primary consequence of intrinsic properties of the SLC and was not a specific property of pre-B cells; previous studies had not definitively established the rate-limiting events responsible (27, 30, 59, 66, 67, 68). In addition, the hypothetical possibility remained that HC homeostasis in B cell lines could be influenced by the consequences of pre-BCR signaling, given that in vivo pre-BCR signaling leads to exit from the cell cycle (27, 67), countermanding a phenotype of transformation. However, our data definitively demonstrated a required, although limited, LC-like escort function for the SLC that links this activity with the positive developmental effects of pre-BCR signaling. This activity is sufficient to explain the low levels of surface pre-BCR density with positively selecting HCs.

Regulation of HC maturation by the JC_L and non-Ig regions of 5 and Vpre-B/V_H pairing

We have identified the non-Ig region of 5 as an inhibitory domain that limits the efficiency at which the SLC complex can promote µm maturation and surface expression. This observation is an important link to studies that showed that a sequence within the N-terminal region of human 5/14.1 acted as an intramolecular chaperone that controlled the folding of the C5 domain (40). In that study the interaction of human Vpre-B with 5/14.1 alleviated the restriction imposed by the 5/14.1 non-Ig region, thus allowing the formation and secretion of an SLC complex with a properly folded C5 region. Based on those observations, a model had been proposed in which the rate-limiting event for surface pre-BCR expression was autonomous SLC complex formation, followed by the efficiency of association of the folded SLC complex with the HC in the ER (10). However, it was not experimentally determined whether this represented the hierarchy of rate-limiting events for the formation of pre-BCR complexes, given that the influence of HCs on these processes was not addressed. In other systems HCs can facilitate the folding of secretion-incompetent LCs (65); thus, the possibility remained that an HC could also influence the rate of SLC-HC complex formation by acting as a cooperative scaffold for SLC folding. Here we establish that the N-terminal, non-Ig region of murine 5 limits µm maturation and surface pre-BCR expression. That the human and mouse regions share 40% amino acid identity strongly suggests that the restrictive property has been functionally conserved between mice and humans. Our data also suggest that HCs do not relieve that restriction and distinguish the properties of the 5 non-Ig region from the MOPC21 V domain.

However, in contrast to the model, the 5 non-Ig region did not inhibit the ability of the SLC or 5 by itself to associate with HCs. Incompletely folded conventional LCs can associate with HCs (69). Consistent with this observation, the non-Ig region blocked the ability of the JC5 region to promote maturation of, but not association with, Dµ. In fact, the trace amount of mature Dµ proteins detected when Dµ was expressed by itself was reduced in the presence of wild-type 5, suggesting that the 5 non-Ig region was dominant and could act intermolecularly to better sequester the associated Dµ protein in the ER (Fig. 5). Given that the wild-type and mutated SLCs associated to the same degree with both full-length and truncated HCs, our data are consistent with a model in which association of the SLC with the HC and the ability to promote HC maturation are distinct activities.

In addition to expediting the proper folding of 5, our data support a second important role for Vpre-B within the SLC complex based on the productivity of its interaction with V_H regions: to provide a structural, and thereby functional, screen (10, 34, 70). We observed a low degree of HC maturation with JC_L proteins, but none with Vpre-B, suggesting that indeed the eviction of BiP from C_H1 by the properly folded JC domain of 5 was a key rate-limiting step for HC maturation. However, and fully consistent with this model, the expression of Vpre-B with the JC_L proteins substantially augmented full-length HC maturation, thus revealing this function uncoupled from its role in facilitating 5 folding. This was not due to the poor folding/associative qualities of JC_L proteins in the absence of Vpre-B, because JC5 and JC associated with Dµ and µm at equivalent levels, and they could independently support Dµ maturation. These data suggest that within the context of a full-length HC, once BiP is evicted by JC5, the V_H domain within a full-length HC does require a folding partner in Vpre-B, and thus the productivity of their interaction becomes a rate-limiting factor for maturation. Unexpectedly, under conditions sufficient for µm, the SLC did not effectively promote Dµ maturation, suggesting that the unpaired Vpre-B protein was not an innocuous presence within the SLC-Dµ complex, but, due to the lack of a V_H region, it was limiting maturation, given that in the context of Dµ there was no V_H domain with which to pair. These observations underscore the importance of cooperativity and compatibility between V_H and V_L regions in HC maturation and extend this function to Vpre-B. Moreover, these data suggest that Vpre-B controls HC maturation and pre-BCR surface transport not only by facilitating V_H folding, but also by requiring V_H pairing for its own folding.

Roles of LCs and the SLC proteins in Dµ maturation

It has been proposed that Dµ-expressing cells are counterselected due to an inability to form a productive BCR complex, based on the observation that Dµ could not associate with a LC (31). Although we have extended this observation using a different LC, we have also demonstrated that Dµ can mature and form a BCR-like complex with 1. The key difference between 1 and the MOPC21 LC lay in their V_L regions: the JC protein could efficiently promote Dµ maturation, suggesting that the particular MOPC21 V domain dominantly interfered with LC folding and association with an HC, as has been demonstrated with some other LCs (65, 71). This also correlated with LC secretion; we observed that 1 was efficiently secreted, but the MOPC21 LC was not (data not shown), implying that the particular V1 did not require a V_H region for folding. These results suggest that Dµ maturation and the formation of signaling-competent BCR-like complexes with Dµ is possible, but requires an LC that contains a V_L region that can autonomously fold.

However, it is not clear whether putative Dµ-1 or Dµ-JC_L complexes could promote positive selection of mature B cells during Ag-independent development. A Dµ transgene suppressed normal B cell development in wild-type mice despite the potential availability of a pool of LCs that could autonomously fold (51), suggesting that Dµ-LC complexes cannot support mature B cell development. In contrast, signaling by truncated HC proteins that lack the C_H1 domain no longer are LC dependent and can direct the progenitor-B to pre-B transition in the absence of 5 (42, 43). Therefore, the Dµ complexes we describe here are likely to be signaling competent, at least at the pre-B cell stage. Interestingly, germline JC proteins have been described in human pre-B cells (72). The ability of the JC_L proteins described in this study to promote Dµ maturation, a degree of HC maturation independently, and enhanced levels of µm maturation cooperatively with Vpre-B, lend support to the idea that such alternate HC-SLC complexes may have functional roles in vivo and may account for aspects of HC-dependent signaling that remain in the absence of 5 and LCs.

Although 5 is important for Dµ counterselection in vivo, the SLC did not promote Dµ maturation or surface expression in a manner comparable to the full-length HCs in our system. We offer the following explanations to reconcile these observations. 1) The basal (LC-independent) surface Dµ levels we observed are sufficient for signaling, but the SLC may still be required to trigger a signal. 2) There may be a small increase in surface Dµ expression above basal levels mediated by the SLC, but it is below our detection levels. 3) Other factors or different conditions may be important for allowing higher levels of SLC-dependent surface Dµ if higher levels are necessary for signaling. Indeed, Dµ was detected on the surface of an SLC-expressing pre-B cell line, but not in mature B cell lines (31), suggesting that the SLC is required for surface expression. However, Dµ was not detected on the surface of every pre-B cell line in which it was expressed (49), suggesting that other factors besides the SLC, differentially expressed in individual B cell lines, may be important. A candidate proposed by Tsubata et al. (49) was a germline V_H protein, the function of which would be to provide a folding V_H partner for Vpre-B. The possibility that discontiguous V and C regions cooperatively and productively interact is evident in the maturation-promoting activity of Vpre-B and 5/JC5 with µm. The resolution of these observations is the goal of future studies.

An unexpected finding was the appearance of a glycosidase-resistant, Vpre-B-dependent form of Dµ. However, the functional significance of this form and the means by which the two proteins associate given that Dµ lacks a V_H region are not known. However, there exist provocative similarities between Dµ and V_H11, a HC with a V_H region associated with the fetal B cell repertoire that does not assemble efficiently with SLC components and is under-represented in adult conventional B cells (73). Interestingly, V_H11, expressed as a transgene, disrupted B cell development, and the V_H11 HC protein, isolated from transgenic B cells, carried unusual post-translational modifications that were resistant to enzymatic and chemical deglycosylation (73). These observations are reminiscent of the Vpre-B-dependent forms of Dµ shown here, and as such suggest that both may represent negative signaling forms of HCs. Furthermore, the enforcement of allelic exclusion in 5-deficient mice has lead to a model in which HC signaling leading to counterselection can occur in the absence of 5 via Vpre-B-HC complexes (1, 33). Based on our results, putative Vpre-B-HC complexes could, at most, mature at only basal levels, and this would imply that all HC-dependent signals, such as negative regulation during early development, would not require normal secretory maturation.

The quantitative and qualitative differences we observed in the abilities of LCs and the SLC to promote Dµ vs µm maturation may be functionally significant. Our data are consistent with a simple quantitative model to account for differences in the outcomes of Dµ vs µm signaling: very low levels of mature surface Dµ with the SLC in pre-BCR complexes is sufficient to signal allelic exclusion, but higher pre-BCR levels achieved by normal HCs can signal allelic exclusion plus differentiation and proliferation. This model predicts that thresholds for activating these signaling pathways are different. Yet even if higher levels of surface Dµ complexes were reached, there is no evidence that Dµ signaling could support B cell differentiation beyond the pre-B cell stage in vivo, where at some frequency LCs that fold should be available. In this context, perhaps the novel Dµ forms we observed are indicative of qualities that preclude this activity. Finally, the evolutionary purpose for maintaining the non-Ig region of 5 and its function in restricting the ability of the SLC to promote HC maturation remain unknown, given the apparent contradiction that promoting surface expression of pre-BCR complexes is necessary for signaling. Our system will provide a new tool to understand these issues, thus leading to a better understanding of how receptor signaling is regulated and B cell homeostasis is achieved.

Note. While this manuscript was under consideration, it was reported that pre-T is important for, but limits, cell surface expression of the pre-TCR, a function parallel to that we propose for 5 in the pre-BCR (74).

	Acknowledgments

 
We thank members of the Roman laboratory for advice and support; Drs. Camilo Parada, Sara Cherry, Konstantina Alexandropoulous, and Ilana Stancovski for critically reading the manuscript and for many helpful discussions; Dr. Joshy Jacob for advice and reagents; and Drs. Michael Neuberger, Shiv Pillai, and Zheng-Sheng Yi for reagents. With much gratitude we thank Dr. David Baltimore, in whose laboratory some aspects of this work were initiated. This paper is dedicated to the memory of Dr. Eugenia Spanopoulou.

	Footnotes

 
¹ This work was supported in part by grants from the American Cancer Society and the New York City Council Speaker’s Fund (to C.A.J.R.) and the New York Academy of Medicine (to B.P.S.).

² Address correspondence and reprint requests to Dr. Christopher A. J. Roman, Department of Microbiology and Immunology and Morse Institute for Molecular Genetics, 450 Clarkson Avenue, Box 44, State University of New York-Downstate Medical Center, Brooklyn, NY 11203. E-mail address: croman{at}netmail.hscbklyn.edu

³ Abbreviations used in this paper: HC, H chain; BCR, B cell receptor; Endo H, endoglycosidase H; ER, endoplasmic reticulum; GFP, green fluorescent protein; HEK, human embryonic kidney; LC, L chain; µm, membrane µ; pre-, precursor; PNGase F, peptide:N-glycosidase F; PSA, Pisum savatum agglutinin; RCA, Ricin communis agglutinin; SLC, surrogate LC; (S)LC, surrogate and conventional LC.

Received for publication March 19, 2001. Accepted for publication July 30, 2001.

	References