A Unique Role for the λ5 Nonimmunoglobulin Tail in Early B Lymphocyte Development¹

Blymphocytes recognize Ags via the variable regions of their membrane-bound BCRs that consist of two covalently associated Ig H chains (HC)³ and two Ig L chains (LC). The exons encoding the variable regions of HC and LC are stepwise assembled during B lymphocyte development by DNA recombination from multiple variable (V), diversity (D), and joining (J) gene segments. Early progenitor (pro) B cells with a productive VDJ exon synthesize a μHC that can assemble with the surrogate L chain (SLC) and the signal transducers Igα-Igβ into the immature precursor (pre) BCR (pre-BCR) (1, 2, 3, 4). Pre-BCR signals enhance the proliferation of large pre-B cells and guide their differentiation through the pre-BCR checkpoint into small quiescent pre-B cells that rearrange the LC gene (5, 6).

Pre-BCR-induced proliferation signals lead to the clonal expansion of pre-B cells producing a functional Igμ HC (μHC), that is, one that can pair with SLC and with LC later in development (7, 8, 9). On the one hand, this compensates for the loss of pro-B cells with two nonproductive VDJ exons and of pre-B cells producing a dysfunctional, that is, a nonpairing μHC (10, 11, 12). On the other hand, the expansion of a single pre-B cell gives rise to a clonal progeny in which all cells produce the same μHC; yet, each of these cells will rearrange a different LC gene and thus produce a BCR with different Ag specificity. Hence, the major function of the pre-BCR is to increase the combinatorial diversity, and thus the overall size, of the primary Ab repertoire (5, 11).

The importance of pre-BCR signals for B lymphocyte development has been established in various gene-targeted and transgenic mice (13, 14, 15, 16, 17, 18, 19, 20). Surprisingly, the SLC is not absolutely required for pre-BCR function (14, 21, 22), since SLC-deficient mice generate mature B lymphocytes, albeit at lower numbers, a finding that has been attributed to the reduced proliferative expansion of large pre-B cells (7, 9). However, it is unclear how the SLC, which consists of the invariant and noncovalently associated polypeptides VpreB and λ5, controls the initiation of pre-BCR-mediated proliferation signals.

Structural motifs that distinguish the SLC from a conventional LC are the non-Ig-like (unique) tails at the C and N termini of VpreB and λ5, respectively (11, 23). As shown previously, both tails are not required for SLC assembly (24). In silico modeling and x-ray crystallography revealed that the unique tails protrude from the pre-BCR at the position where the third CDR (CDR3) would be located in the BCR (25, 26, 27) and could, therefore, be accessible for ligands. Indeed, the λ5 unique tail was required for binding of pre-BCRs to the stroma cell-associated self-Ags heparan sulfate and galectin-1, which have been proposed to induce pre-BCR signaling (28, 29, 30). Alternatively, signals could be initiated autonomously, since the pre-BCR is also reactive against itself, which leads to self-oligomerization, presumably via direct interaction of the oppositely charged unique tails of VpreB and λ5 (25, 31, 32, 33).

However, none of these in vitro studies clarified whether the unique tails of the SLC are indeed critical for pre-BCR-mediated proliferation signals in vivo. On the contrary, previously published in vivo studies rather suggest that the SLC is functionally dispensable and can be replaced by a prematurely expressed conventional LC (16, 18, 34, 35). Therefore, we established transgenic mice with mutated pre-BCRs either lacking the λ5 unique tail or bearing a modified λ5 unique tail in which all positively charged amino acids were converted to alanine. Flow cytometry analyses of B lymphoid cells in these mice revealed that the proliferative expansion of μHC-positive pre-B cells with mutated pre-BCRs was impaired and, consequently, the numbers of mature B lymphocytes were markedly reduced. Hence, the λ5 unique tail controls early B lymphocyte development and is critical to increase the primary Ab repertoire by enhancing the proliferation of pre-B cells that produce a functional μHC.

λ5 transgenic mice were generated by microinjecting linearized DNA fragments (see “Retroviral, bacterial, and transgenic expression vectors” below) into pronuclei of oocytes from FVB mice. Transgenic offspring were verified by Southern blot (see below) and genotyped by PCR (95°C, 3 min; 35 cycles (95°C, 30 s; 60°C, 30 s; 72°C, 1 min); 72°C, 5 min) with the following primers: GGACTGGATATCAGTCAGGCAGAGCTG, ACTTTGCCCCCTCCATATAACATGAA. λ5 transgenic mice, RAG2^−/− mice (36), λ5^−/− mice (14), and μHC transgenic quasimonoclonal (QM) mice (37), in which the J_H cluster was replaced with the VDJ exon from the hybridoma 17.2.25, were maintained under specific pathogen-free conditions. All animal experiments were performed in accordance with the guidelines of the Animal Care and Use Committee of the Government of Bavaria and the institutional guidelines of the University of Erlangen-Nürnberg.

BamHI-digested genomic DNA was analyzed as described (38) by Southern blotting with a 1687-bp-sized λ5 probe amplified by PCR from genomic DNA (primers: AGTCAGCTAGCAAACTCCAAATATCTGTAAAATTCCAAAGCTTCT and TGACTGCTAGCCAGCACACAAACCTATAGTGCAAACAGGTCC) and labeled with [³²P]dCTP (RadPrime kit, Invitrogen). Radioactive bands were visualized with a FLA-3000 phosphoimager (Fujifilm) equipped with BASReader software (Raytest).

The λ5 coding sequence lacking the unique tail (λ5ΔU) was constructed by inserting GCTCAGCATGGTATGTCTTTGGTGGTGGGACCCAGCTCACAATCCTAGG (Invitrogen) into BlpI and AvrII sites of the murine λ5 coding sequence. λ5Uala was constructed by site-directed mutagenesis of all basic amino acid codons (seven arginines and one lysine) in the λ5 unique tail, as follows: Briefly, the synthetic BstZ17I/NaeI-digested fragment GTATACATAAGCTACATATACGCAGAAGCAAGCGCCGCTGTGGGCCCTGGAGCTTCAGTGGGAAGCAACGCCGGC was inserted into BstZ17I/StuI-digested pCR2.1-λ5mel vector (29). Subsequently, an overlap PCR product was obtained from the modified λ5mel sequence (primers: GCCGGAGCAGGTCCCGCTTGCTCGCCCCATGCCCTTCCATCTGCACCCCAG, AGCGGGACCTGCTCCGGCTGGGATGATCTGGAACAGGAGTGCGCCGGG, TCCCGGGACCATGAAGTTC, and TCTCGAGCTAAGAACACTCAGCAGGTGAC). Finally, the mutated λ5 sequence with the murine leader was constructed by inserting into the λ5ΔU coding sequence a BlpI/AvrII-digested PCR product obtained from mutated λ5mel coding sequence (primers: CATACTTTCCCCAAGCTCAGCAGAAGCAAGCGCCGCTGTGG and CTTGGGCTGACCTAGGATTGTG).

Retroviral λ5 expression vectors were constructed by inserting λ5 coding sequences upstream of an internal ribosomal entry site-GFP cassette into the EcoRI-digested MIEV vector (39). Bacterial expression vectors encoding the GST fusion proteins were constructed by inserting the unique tail sequences of murine as well as human λ5 and VpreB into the BamHI and EcoRI restriction sites of the GST-encoding plasmid pGEX2T (Amersham Biosciences). Wild-type and mutated unique tail sequences of λ5 and VpreB were obtained by PCR with the following primers: Uλ5m forward, CGCGGATCCGAAAGGAGCAGAGCTGTGGG; Uλ5m reverse, CGGAATTCCTAAAACTGGGGCTTAGATGGAAG; UVpreBm forward, CGCGGATCCCAGGAAAAGAAGAGGATGGAG; UVpreBm reverse, CGGAATTCCTAAGATCCCAAATCTGTATACG; Uλ5h forward, CGCGGATCCTCGCAGAGCAGGGCCCTG; Uλ5h reverse, CGGAATTCCTACACTGAGTTATGCTTGGATTGA; UVpreBh forward, CGCGGATCCTCGGAGAAGGAGGAGAGGG; UVpreBh reverse, CGGAATTCTCAAGGGACACGTGTCCTGG. The mutated unique tail of murine λ5 (Uala(λ5m)) was obtained from mutated λ5mel coding sequence by PCR with the primers CGCGGATCCGAAGCAAGCGCCGCTGTGG and CGGAATTCCTAAAACTGGGGTGCAGATGGAAG. Transgenic expression vectors were constructed as follows: Mutated and wild-type murine λ5 coding sequences with a Kozak consensus motif were inserted between a splice donor (exon 2)/intron/splice acceptor (exon 3) and the polyadenylation site of rabbit β-globin gene. The cassette was inserted downstream of the murine λ5 promoter (40) and between two genomic fragments isolated from murine λ5 locus by PCR as follows: The upstream 2773-bp genomic fragment starting with TATTCTGCAGAAGTGCAGCATGCAG and ending with GGCTAGAGTTGACTTTGGACTTGAGGGTCA contains the 5′ locus control region (LCR) as well as the λ5 promoter, and the downstream 4636-bp genomic fragment starting with CTAGGGGAGACATATGCAACGTGTGCCC and ending with TGGACCTGTTTGCACTATAGGTTTGTGTGCTG contains the 3′LCR (41).

CD19⁺c-kit⁺ BM cells were isolated by magnetic sorting on autoMACS (Miltenyi Biotec) with FITC-conjugated anti-CD19 Abs (BD Biosciences) and anti-FITC Abs coupled to magnetic beads (Multisort kit, Miltenyi Biotec), followed by R-phycoerithrin (RPE)-conjugated anti-c-kit Abs (BD Biosciences) and anti-RPE Abs coupled to magnetic beads, according to the manufacturer’s instructions (Miltenyi Biotec). Retroviral supernatants were produced by transfecting the appropriate vectors into phoenix-E packaging cells maintained in DMEM (1 mM sodium pyruvate, nonessential amino acids, 50 μM 2-ME, 50 U/ml penicillin G, 50 μg/ml streptomycin) with 10% FCS. Retrovirally infected B-lymphoid cells (10⁴) were seeded in triplicate onto 10⁴ mitomycin C-treated ST-2 stroma cells and grown in a 96-well plate in complete RPMI 1640 medium (7). Cell numbers were determined with Flow Count Fluorospheres (Beckman Coulter) on FACSCalibur (BD Biosciences).

GST fusion proteins in 10 mM phosphate (pH 7), 180 mM NaCl were loaded at 1 ml/min onto a heparin column (Amersham Biosciences) and eluted with 10 mM phosphate (pH 7), 1.5 M NaCl. Fractions were analyzed by Western blotting as previously described (42) with goat anti-GST Abs (Amersham Biosciences) and HRP-conjugated rabbit anti-goat Abs (Sigma-Aldrich).

Single-cell suspensions were stained as previously described (38) with the following Abs: RPE-conjugated rat anti-CD19 (clone 1D3, BD Biosciences), rat anti-CD19 (clone 1D3F2) conjugated to Alexa Fluor 488 with a labeling kit (Molecular Probes/Invitrogen), RPE-conjugated rat anti-CD25 (clone PC61, BD Biosciences), RPE-conjugated anti-c-kit (clone ACK45, BD Biosciences), RPE-conjugated monoclonal rat anti-Igκ (clone 187.1, Southern Biotechnology Associates), RPE-conjugated goat anti-Igλ (Southern Biotechnology Associates), rat anti-λ5 (clone LM34, from F. Melchers) conjugated to Cy5 (GE Healthcare/Amersham Biosciences), FITC-conjugated anti-μHC (BD Biosciences), goat anti-μHC (Southern Biotechnology Associates) conjugated to Cy5 with a labeling kit (GE Healthcare/Amersham Biosciences), rat anti-pre-BCR (clone SL156 (43), from F. Melchers), and Cy5-conjugated goat anti-rat (The Jackson Laboratory). To detect cytoplasmic proteins, cells were fixed and permeabilized with Fix&Perm kit (An der Grub Bio Research). Cells in the lymphocyte gate were analyzed with a FACSCalibur. Binding of GST fusion proteins to ST-2 stroma cells was detected with goat anti-GST (Amersham Biosciences) and FITC-conjugated anti-goat IgG Abs (Jackson ImmunoResearch Laboratories).

The λ5 unique tail is thought to initiate signals by mediating pre-BCR self-oligomerization (33) or by engaging with stroma cell-associated self-Ags (28, 30). Regardless of the utilized mechanism, mutations of the λ5 unique tail should therefore interfere with pre-BCR signaling, and thus with the developmental transition of pre-B cells through the pre-BCR checkpoint (6, 11). To test this in vivo function of the λ5 unique tail in early B lymphocyte development, we established three different transgenic mouse strains that were backcrossed with λ5-deficient mice (λ5^−/−). The first transgene encoded a λ5 chain lacking the unique tail (λ5ΔU), the second a λ5 chain with mutated unique tail, in which all positively charged amino acids (seven arginines and one lysine) were converted to alanine (λ5Uala), and the third encoded wild-type λ5 (λ5wt) (Fig. 1,A). Genomic integration of transgenes was confirmed by Southern blotting (Fig. 1,B). Since transgenes were driven by the λ5 promoter and two λ5 locus control regions (Fig. 1,A), which can be silenced by pre-BCR signals (41, 44, 45), we expected lineage- and stage-specific transgene expression in pro-B cells and early pre-B cells, but not in later B-lymphoid stages. Indeed, flow cytometry of bone marrow cells from the mouse lines λ5ΔU^Tg1, λ5Uala^Tg1, and λ5wt^Tg13 detected transgenic λ5 in CD19⁺c-kit⁺ pro-B and early pre-B cells, but not in CD19⁺CD25⁺ late pre-B or in CD19⁺IgM⁺ B cells (Fig. 2). Additionally, the abundance of λ5 was comparable between pro-B cells from λ5 transgenic and wild-type λ5^+/− mice (Fig. 2, left), and the surface expression of pre-BCRs containing λ5ΔU or λ5Uala was only slightly higher than that of wild-type pre-BCRs (Fig. 3). Therefore, pre-BCRs with a mutated λ5 unique tail can reach the cell surface of early pre-B cells and induce signals that silence the expression of the SLC component λ5 in late pre-B and IgM⁺ B cells.

If the λ5 unique tail controls pre-BCR-induced clonal expansion signals in μHC-positive pre-B cells, we would expect that B lymphocyte development is impaired at the pre-B cell stage in mice with tail-mutated λ5. As predicted, flow cytometry revealed that frequencies and numbers of CD19⁺CD25⁺ pre-B cells in the bone marrow and of CD19⁺IgM⁺ B cells in the bone marrow as well as in the spleen were reduced in λ5ΔU and λ5Uala transgenic mice when compared with wild-type mice (λ5^+/−) (Fig. 4,A and Table I). Similar results were obtained for two other independent transgenic lines expressing mutated λ5 (data not shown). In contrast, mice with transgenic wild-type λ5 (λ5wt) showed no substantial alterations in the frequency and number of B-lineage cells when compared with wild-type mice (λ5^+/−) (Fig. 4,A and Table I). These findings exclude that random integration of λ5 transgenes adversely affected B-lineage cells and demonstrate that the λ5 unique tail is critical for developing B lymphocytes at the transition from the pro-B to the pre-B cell stage.

If pre-BCR-mediated clonal expansion signals are impaired in transgenic mice with tail-mutated λ5, we would expect a reduction in cytoplasmic μHC⁺LC⁻ pre-B cells (46). As predicted, the frequencies of these cells among all CD19⁺LC⁻ B-lineage cells (pro-B and pre-B cells) were reduced in the bone marrow of λ5ΔU and λ5Uala transgenic mice when compared with wild-type mice and mice transgenic for λ5wt (Fig. 4,B). This indicates that the λ5 unique tail is required for efficient clonal expansion and/or survival of μHC⁺ pre-B cells. Additionally, no significant differences were detected in the absolute number of CD19⁺c-kit⁺ pro-B cells in mice with mutated or wild-type λ5 (Table I), which provides further evidence that the λ5 unique tail is critical for early B lymphocyte development at the pre-B cell stage.

The reduced number of pre-B cells in transgenic mice with mutated λ5 chains could be attributed to a decrease in either proliferation or survival. To distinguish between these two possibilities, we analyzed the cell cycle by measuring the DNA content of DAPI-stained large μHC⁺LC⁻ pre-B cells. As revealed in another staining, large μHC⁺LC⁻ pre-B cells were surface pre-BCR-positive (Fig. 3), and they thus represent the earliest cells that coexpress a newly synthesized μHC and the SLC components VpreB and λ5. Compared with wild-type mice (λ5^+/−), λ5ΔU and λ5Uala transgenic mice had lower frequencies of large pre-B cells in the proliferative cell cycle phases S/G2/M, albeit the frequencies were still higher than in λ5^−/− mice (Fig. 5,A). In contrast, no significant differences were observed for the frequencies of cycling pre-B cells in λ5wt transgenic and wild-type mice (λ5^+/−). Further, the frequency of cycling pre-B cells in λ5Uala transgenic mice was slightly higher than in λ5ΔU transgenic mice, which is consistent with the finding that λ5Uala transgenic mice have more pre-B and B cells in the bone marrow when compared with λ5ΔU transgenic mice (Fig. 4 and Table I). This indicates that pre-BCRs with mutated λ5 unique tail induce some proliferative signals, which are, however, markedly reduced when compared with wild-type pre-BCRs. In contrast, apoptotic pre-B cells in the sub-G1 phase were hardly detectable in all mice (<2%; Fig. 5 A). Hence, both the deletion and mutation of the λ5 unique tail predominantly impair proliferation but do not severely affect the survival of pre-B cells.

To confirm that the λ5 unique tail controls pre-BCR-mediated cell expansion, we retrovirally transduced pre-B cells from μHC transgenic (37) RAG2/λ5-deficient mice with bicistronic vectors encoding mutated or wild-type λ5 and enhanced GFP. The cells were cultured on ST-2 stroma cells and the number of GFP-positive pre-B cells was determined over a period of 4 days. The increase in pre-B cells transduced with mutated λ5 was lower than that of pre-B cells transduced with wild-type λ5 (Fig. 5,B, top), even though pre-BCRs with mutated λ5 were properly expressed at the cell surface (data not shown). The increase in pre-B cells with mutated λ5 was, however, higher than that of pre-B cells transduced with the GFP control vector (Fig. 5,B, top), suggesting that the growth of μHC⁺ pre-B cells is enhanced by, but is not strictly dependent on, the λ5 unique tail. In contrast, only minor or no differences were observed for the growth of transduced μHC⁻ pro-B cells from RAG2/λ5-deficient mice (Fig. 5 B, bottom), which excludes nonspecific effects of mutated λ5 chains on cell growth. Additionally, propidium iodide (PI) staining revealed a higher frequency of cycling pre-B cells from μHC transgenic RAG2/λ5-deficient mice in cultures transduced with wild-type λ5 than in cultures transduced with mutated λ5 (data not shown). Therefore, the λ5 unique tail is a pre-BCR-specific motif that controls the clonal expansion of pre-B cells with a functional μHC by amplifying proliferation signals.

To explain why tail-mutated λ5 chains partially restored B lymphocyte development in λ5-deficient mice (Fig. 4), we hypothesized that pre-BCR-mediated differentiation signals, in contrast to proliferation signals, do not require the presence of an intact λ5 unique tail. For example, pre-BCRs with tail-mutated λ5 could still induce the rearrangement and expression of LCs (47). In this case, the ratio of LC⁺ B cells to small LC⁻ pre-B cells should be similar in mice with mutated and wild-type λ5 (Fig. 6,A). As expected, this ratio was comparable in λ5ΔU transgenic, λ5Uala transgenic, and wild-type mice when we analyzed cytoplasmic μHC⁺ B lineage cells for the expression of LC by flow cytometry (Fig. 6, B and C). In contrast, a much lower ratio of LC⁺ B cells to small LC⁻ pre-B cells was observed in λ5-deficient mice. These findings suggest that efficient pre-BCR-mediated differentiation signals, which induce LC expression in small pre-B cells, require the presence of the λ5 chain, but not its unique tail. In summary, we have shown that the λ5 unique tail is a critical pre-BCR-specific signaling motif that controls early B lymphocyte development by enhancing proliferation, but not differentiation, of pre-B cells.

Our finding that the clonal expansion of pre-B cells is impaired in mice with tail-mutated λ5 can be explained by the model that the positively charged amino acids in the λ5 unique tail are critical for pre-BCR self-oligomerization (33). However, our data do not exclude that the λ5 unique tail is also involved in binding of the pre-BCR to self-Ags, such as bone marrow stroma cell-associated heparan sulfate (30). Since heparan sulfate contains many negatively charged sulfate and carboxyl groups, it is a good candidate to interact with the positively charged residues in the λ5 unique tail.

To test this idea, we produced recombinant fusion proteins consisting of GST and the unique tail (U) of either murine λ5 (mλ5) or human λ5 (hλ5). Additionally, we replaced all positively charged amino acids (seven arginines and one lysine) in the λ5 unique tail with alanine, fused the mutated tail to GST (Uala(mλ5)-GST), and analyzed binding of all fusion proteins to immobilized heparin, a glycosaminoglycan closely related to heparan sulfate, by affinity chromatography. As expected, most of U(mλ5)-GST and U(hλ5)-GST were found in the elution fraction by Western blot analysis (Fig. 7,A). In contrast, Uala(mλ5)-GST was almost exclusively detected in the flow through and only a minor fraction was present in the elution fraction (Fig. 7 A). Hence, the basic amino acids in the λ5 unique tail are critical, but not absolutely required, for binding of the pre-BCR to heparan sulfate. Unconjugated GST (data not shown) and GST fused to the unique tail of murine VpreB (U(mVpreB)) or human VpreB (U(hVpreB)) were exclusively detected in the flow through, which demonstrates that only the unique tail of λ5 contains binding sites for heparan sulfate.

These findings were corroborated when we analyzed the binding of GST fusion proteins to the surface of ST-2 bone marrow stroma cells by flow cytometry. When compared with wild-type U(mλ5)-GST and U(hλ5)-GST, only weak binding of Uala(mλ5)-GST to ST-2 stroma cells could be detected with Abs against GST (Fig. 7 B). In contrast, no binding was observed for unconjugated GST, U(mVpreB)-GST, or U(hVpreB)-GST. These findings demonstrate that the positively charged amino acids in the λ5 unique tail mediate the interaction of pre-BCRs with stroma cell-associated self-Ags, such as heparan sulfate. Since that could explain the impaired clonal expansion of pre-B cells producing tail-mutated λ5, the involvement of ligands in pre-BCR signal initiation should be considered in future studies.

We determined the in vivo function of the λ5 unique tail in B lymphocyte development by establishing transgenic mice lacking this pre-BCR-specific signaling motif. This was stimulated by recent in vitro experiments demonstrating that the λ5 unique tail is critical for both autonomous pre-BCR signal initiation (25, 33) and binding of pre-BCRs to stroma cell-associated self-Ags (28, 30). Our analysis now establishes that the λ5 unique tail controls the clonal expansion of pre-B cells with functional μHCs by accelerating the cell cycle (Figs. 4,B and 5). Thus, the λ5 tail is the major structural and functional hallmark that distinguishes a pre-BCR from a mature BCR and targets proliferation signals specifically to pre-B cells, but not to mature B cells. The clonal expansion of pre-B cells leads to a progeny of cells that produce the same μHC but will rearrange different LC genes, and thus synthesize Abs with different Ag specificities. Therefore, pre-BCR-mediated proliferation signals initiated by the λ5 unique tail may be pivotal to increase the combinatorial diversity in the primary Ab repertoire.

Another important observation was that pre-BCRs lacking an intact λ5 unique tail still delivered differentiation signals that terminated stage-specific SLC expression (Fig. 2, middle and right) and induced LC production in pre-B cells (Fig. 6). Based on these findings, we propose a unified pre-BCR signal initiation model in which the C-terminal Ig-like fold domain of λ5 in concert with VpreB displace a μHC from the endoplasmic reticulum-resident chaperone BiP (48), facilitate the transport of pre-BCRs to the cell surface, but govern only basal signals that support survival and differentiation of pre-B cells (Fig. 8, left). Subsequently, crosslinking of pre-BCRs via the λ5 unique tail increases the signal amplitude, thereby reaching the threshold required to induce efficient proliferation (Fig. 8, right).

Crosslinking could be initiated, for example, by autonomous pre-BCR oligomerization via the positively and negatively charged amino acids in the unique tail of λ5 and VpreB, respectively (25, 31, 32, 33). Similarly, pre-TCR signal initiation appears to occur through self-oligomerization via charged residues in the extracellular domain of pre-Tα (49). However, as recently discussed (6, 50), pre-BCR crosslinking could also be mediated by pre-B cell-associated ligands (not depicted in Fig. 8) or by stroma cell-associated self-Ags, such as heparan sulfate and galectin-1 (28, 30), which bind to the positively charged residues in the λ5 unique tail (Fig. 7). In support of this idea, heparan sulfate and galectin-1 appear to enhance the phosphorylation of signaling molecules downstream of the pre-BCR, such as Igα and ERK (28, 51, 52). All of these models are not mutually exclusive, and ligand-mediated enhancement of autonomous pre-BCR signals may have evolved to maintain those pre-B cells that express only low amounts of the pre-BCR at their cell surface. Interestingly, heparan sulfate also interacts with the surface molecule CD19 (53), which may co-cluster pre-BCRs and CD19 to additionally enhance pre-BCR signaling (54).

In contrast, our finding that pre-BCR-mediated differentiation signals are triggered even in the absence of the λ5 unique tail (Fig. 6) indicates that some tonic signals can be initiated without pre-BCR cross-linking (55). Stroma cell-associated self-Ags are unlikely to cross-link pre-BCRs containing tail-mutated λ5, e.g., through the unique tail of VpreB, since the unique tail of VpreB did not interact with heparin or with the surface of stroma cells (Fig. 7). Tonic signals can be easily explained by our unified model for pre-BCR signal initiation, which postulates that weak signals sufficient for survival and differentiation of pre-B cells can be induced even by a pre-BCR that is unable to oligomerize (Fig. 8, left). Likewise, tonic signals in the absence of pre-BCR clustering might be triggered by truncated μHCs that are unable to associate with SLC, but nonetheless support the differentiation of pre-B cells in transgenic mice (13, 16, 18).

Compared with λ5ΔU transgenic mice, the expansion of pre-B cells was slightly more efficient in λ5Uala transgenic mice, in which all positively charged amino acids of the λ5 unique tail were replaced with alanine (Figs. 4,B and 5, Table I). Based on our finding that heparin still weakly binds to the λ5 unique tail in the absence of its positively charged amino acids (Fig. 7,A), we propose that weak engagement by stroma cell-associated heparan sulfate induces some proliferative signals through pre-BCRs carrying an uncharged λ5 unique tail. Alternatively, the alanine residues in λ5Uala might support weak pre-BCR self-oligomerization via direct hydrophobic interactions, and thus enhance basal autonomous signaling. Hence, the phenotype of λ5Uala transgenic mice can be explained by our unified model for pre-BCR signal initiation (Fig. 8).

The number of B lymphocytes in λ5-deficient mice can be restored by premature expression of a productively rearranged κLC transgene (34, 35). This suggests that a conventional LC can replace the SLC, and thus argues against a pivotal role of the λ5 unique tail in the clonal expansion of μHC-positive pre-B cells. However, the premature and constitutive expression of a transgenic κLC in all newly generated pre-B cells bypasses the developmental stage at which cells with unproductive LC genes are usually eliminated. Therefore, the increased survival of pre-B cells in κLC transgenic mice might compensate for their presumably inefficient clonal expansion. A conventional LC may thus, in analogy to the SLC lacking the λ5 unique tail, trigger only survival and differentiation, but not proliferation signals, which is in agreement with our unified model for pre-BCR signal initiation (Fig. 8, left). A critical role for the SLC in enhancing pre-BCR-mediated proliferation signals (Fig. 8, right) is also supported by recent findings that the SLC, in contrast to a conventional κLC, can efficiently induce Igα phosphorylation and Ca²⁺ signaling in μHC-positive pre-B cell lines (32, 33).

Most of the basic amino acid residues in the unique tail of λ5 are conserved in humans, mice, rats, and rabbits (33, 56). This suggests that the pro-proliferative function of the λ5 tail may have favored the asynchronous rearrangement of HC and LC genes during mammalian evolution, because random pairing of one HC with various LCs to increase Ab diversity requires that the clonal expansion of μHC-positive pre-B cells precedes LC gene rearrangements. In this view, the assembly of a μHC with the SLC replaces an older evolutionary state, when rearrangement of HC and LC genes still occurred simultaneously and before clonal expansion (57). In support of this idea, no orthologs for λ5 and VpreB1/VpreB2 genes were identified in chickens, in which HC and LC gene loci are simultaneously rearranged during early embryonic development (58, 59). Chickens may not require SLC, since Ab diversity is generated predominantly by gene conversion in the bursa of Fabricius after the IgM-positive B lymphocyte population has been expanded (60). In contrast, the evolutionary emergence of VpreB and λ5 together with the autoreactive λ5 unique tail may have provided the mammalian immune system a mechanism to build up its primary Ab repertoire by increasing the combinatorial diversity in pre-B cells. In this evolutionary perspective, the SLC may represent an ancient autoreactive LC that was hard-wired in the genome to become an invariant component of the pre-BCR (61). In accordance with this hypothesis, the expression of SLC including its autoreactive λ5 unique tail is tightly regulated and terminated by pre-BCR signals (44, 45), which resembles a mechanism of tolerance induction. Thus, the exclusive expression of SLC in early pre-B cells could ensure that transient autoreactivity can be exploited to select the primary Ab repertoire, while avoiding the generation of autoreactive B lymphocytes.

We thank Johannes Lutz and Katy Schmidt for critical reading of the manuscript, Maureen Grady for proofreading, and Dennis Castor, Rita Spannenberger, and Regina Vogelbacher for help in constructing vectors.

The authors have no financial conflicts of interest.