Assembly of the (cid:1) PreB Receptor Requires a V (cid:1) -like Protein Encoded by a Germline Transcript* □ S

By confining germline transcription as a byproduct of the mechanisms inherent to genetic rearrangements, the translation of respective mRNAs and their biological relevance might have been overlooked. Here we report the identification, cloning, and biochemical characterization of a human V (cid:1) -like protein that is encoded by a germline transcript. This surrogate protein assembles with the immunoglobulin (cid:2) heavy chain at the surface of B cell progenitors and precursors to form a (cid:1) -like antigen receptor. These findings support the notion that germline transcription is not futile and stress the flexi-bility in eukaryotic

The remarkable diversity of the eukaryotic genome stems from the distinct expression patterns of genes in different tissues. Whereas some gene products are required for basal cell functions and are constitutively expressed by most cells, a restricted transcription and translation dictate tissue-and cellspecific functions. For instance, the expression of antigen (Ag) 1 receptor genes is developmentally controlled, and it is exclusive of the immune cells. A problem with such restriction, as experienced by the adaptive immune system in mammals, is the requirement of a broad repertoire of Ag receptor genes that surpasses the encoded capacity of the genome (1). However, nature has tailored gene rearrangement mechanisms to circumvent the diversity constraints of the Ag receptor loci and to ensure a large repertoire of Ag specificities (2)(3)(4). Indeed, the paired variable (V), diversity (D), and joining (J) protein sequences that furnish Ag specificity are encoded by distinct gene segments, which in their germline configuration are distantly separated and cannot produce a contiguous transcript. Thus, during gene rearrangement, B and T lymphocyte progenitor/ precursors use one of the many V, D, and J segments that are distantly clustered, in a manner that results in a clonally unique Ag-binding sequence (5)(6)(7)(8)(9)(10). Although V[D]J rearrangement can be accomplished by both B and T lymphocytes (10), B cells rearrange the immunoglobulin (Ig) Ag receptor (BCR) genes, and T cells rearrange T cell receptor (TCR) genes.
Whereas the transcription of unrearranged/germline Ig-like genes has long emerged as a major function by which progenitor and precursor lymphoid cells can regulate V[D]J rearrangement (11,12), it was prematurely concluded that the expression of germline Ig genes was sterile and limited to transcription, with no other function than promoting chromatin changes and gene loci accessibility. However several studies in B lymphocytes have demonstrated that Ig gene rearrangement is initiated by signals that are driven through the -like PreB cell receptor, which is formed by the heavy chain (HC) in conjunction with unrearranged/germline surrogate light chains (SLC) 5 and VpreB (13)(14)(15)(16)(17).
Because single and combined inactivation of 5, VpreB 1 and VpreB 2 genes did not completely abrogate B lymphocyte maturation (18 -21), it was hypothesized that other germline PreB receptor molecules, which functioned similarly to the 5 and VpreB proteins, could take over to back-up PreB receptor function (18). In line with this prediction, a surrogate J-C (SJC) germline transcript, encoding a protein with the capacity to covalently associate with HC at the surface of Pro/PreB cells, was identified and characterized as a candidate to form the alternative -like PreB receptor (11).
As the 5 PreB receptor is formed by HC, 5, and VpreB, we postulated that by analogy, the PreB receptor should be constituted by HC, SJC, and a putative V-like molecule (11,22). In support of this hypothesis, we report here the molecular characterization of a V-like protein encoded by a germline transcript, which can associate with HC at the surface of ProB/PreB cells.
Collectively, the data presented herein and the earlier identification of the SJC molecule (11)  germline transcription is not sterile and stress the relevance of the encoded proteins in cell and developmental functions.

Cells
The PreB acute lymphocytic leukemia (ALL) cell line Blin-1 and the mature B cell line 1E8 derived from Blin-1 after in vitro rearrangement and maturation were kind gifts from Dr. Tucker LeBien. The PreB ALL cell line PreALP was previously reported for the characterization of the surrogate JC molecule (11). The mature B cell line B207 (23) was a generous gift from Dr. Max Cooper. The PreB ALL cell line Nalm 6, the Burkitt lymphoma cell lines Daudi and Namalwa, and the T cell line Jurkat were obtained from the American Type Culture Collection. Last, enrichment of normal ProB/PreB cells from human bone marrow was accomplished by negatively selecting with a custom-made antibody mixture (anti-CD2, CD3, CD13, CD14, CD16, CD36, CD56, CD66b, and glycophorin A), using magnetic-activated cell sorting (MACS) according to the manufacturer's guidelines (StemCell Technologies, Vancouver, BC).

Antibodies
Anti-IgM antibodies were obtained from BD Biosciences (San Diego, CA). Two affinity-purified anti-V-like IgGs were custom-made: Antibody p42-58, which reacts with a peptide encompassing amino acids 42-58 of the V-like protein sequence, was produced by Neosystem Laboratory (Strasbourg, France); Antibody pGL121-142 was custommade by Bethyl Laboratories (Montgomery, TX) against a germline V-specific peptide comprising the heptamer/spacer/nonamer domain between amino acids 121 and 142.

RNA Isolation
Total RNA was extracted using the TRIzol reagent (Invitrogen) and following a previously described method. Poly(A) ϩ was subsequently purified by two rounds of oligo d(T) 12-18 chromatography using the FastTract R kit (Invitrogen) according to the manufacturer's specifications.

RT-PCR
Cell Lines-A total of 0.5 g of poly(A) ϩ RNA from PreB ALL or mature B cell lines were reverse-transcribed using oligo d(T) [12][13][14][15][16][17][18] and the RT kit Superscript II TM (Invitrogen) according to the manufacturer's instructions. The use of poly(A) ϩ RNA as a template and oligo d(T) as a primer in the RT reaction was aimed at preventing PCR amplifications originating from potential genomic DNA contaminants. A total of 2 l of RT reaction product was amplified through 35 cycles (30 s at 94°C, 30 s at 60°C, and 1 min at 72°C) using a previously described method (11) and the Expand TM High Fidelity PCR system (Roche Applied Science). The 5Ј forward primers are consensus oligonucleotides covering each of the known LV gene families (24). The 3Ј reverse primer was either a germline consensus sequence at the heptamer/ spacer/nonamer segment or an oligonucleotide designed from the C region (24).
Normal Bone Marrow Pro/PreB-The RT-PCR analysis was carried out as mentioned above except that the forward (5Ј) and reverse (3Ј) primers were based on the cloned full-length V4-like cDNA.

cDNA Construction and Analysis
Generation and screening of the pre-ALP cDNA library was carried out by using the -ZAP system (Strategene, La Jolla, CA) as previously reported (11) and the RT-PCR-generated germline V cDNA probes. Sequencing was carried out by automated DNA sequencing (The University of Texas, M. D. Anderson Cancer Center DNA sequencing core facility).

Immunoprecipitation
For immunoprecipitation, 2 ϫ 10 7 cells were lysed and precipitated with anti-IgM or anti-V-like antibodies. To assess the molecular association of the HC with the V-like molecule, we performed IPs using 2 mg of protein lysates of Blin-1 or freshly isolated normal ProB/PreB cells. The immunoprecipitations were carried out using anti-IgM monoclonal antibodies coupled to agarose beads (Sigma) and anti-V-like antibodies. After overnight incubation at 4°C, the anti-V-like antibodies were pulled-down using protein A-agarose (Roche Applied Science). The IgM and V-like immunocomplexes were extensively washed with radioimmune precipitation assay buffer and analyzed by Western blots.

Western Blotting
Immunoblotting analyses were carried out following SDS-PAGE (25) using a commercial electrotransfer system (Bio-Rad) according to the supplier's protocol. Reactivity to V moieties was determined by the binding of the antipeptide sera p42-58, which reacts with the variable domain, and pGL121-142, which specifically targets the germline V molecule. Reactivity to p42-58 was directly tested by Western blotting without a precipitation step. To enhance the specificity toward Ig complexes, western binding to pGL121-142 was preceded by immunoprecipitation of the cell lysates with either protein A (known to bind Ig Fc domain) or anti-HC ϩ protein A. Western blots were revealed by autoradiography (15 min) after the chemiluminescence reaction, using the SuperSignal TM West Pico Chemiluminescent System (Pierce).

Surface Staining, Enzymatic Amplification, and Flow Cytometry
An enzymatic amplification staining (EAS) kit (FlowAmp Systems, Ltd.) was used, following the manufacturer's instructions. Briefly, 1 ϫ 10 6 Blin-1 cells were washed with staining buffer (phosphate-buffered saline (PBS), pH 7.4; 2% fetal bovine serum, and 0.1% sodium azide). Cells were then reacted with the pGL121-142 antibody (1.0 g/10 6 cells) at 4°C for 20 min in 50 l of staining buffer. The reaction was followed by two 2.0-ml washes with staining buffer, and the cell pellets were subsequently reacted by incubating them at 4°C for 20 min with a horseradish peroxidase-conjugated anti-rabbit second antibody (diluted 1:100) in 50 l of staining buffer. Cells were subsequently washed at least three times with sodium azide-free staining buffer. After a thorough cell washing to remove the sodium azide, cells were exposed to the EAS amplifier reagent (diluted 1:20) for 10 min at room temperature, followed by a 10-min incubation with the detector reagent (fluorescein isothiocyanate-conjugated streptavidin, diluted 1:100) at room temperature. Surface anti-V-like reactivity was detected by flow cytometry in an Epics Profile analyzer (Beckman/ Coulter Inc., Fullerton, CA).

Fluorescence Surface Staining
Normal ProB/PreB and Blin-1 cells were washed with PBS, followed by a 30-min incubation at 37°C with the DNA-binding fluorochrome Hoechst 33342 (Molecular Probes, Eugene, OR) to stain the nuclei. Alternatively, nuclei were stained with TO-PRO-3 marker (Molecular Probes). Typically, 1-5 ϫ 10 4 cells were cytocentrifuged onto microscope slides. Cells were fixed with 4% paraformaldehyde in PBS for 10 min at room temperature. Fixed cells were washed three times with PBS and blocked with PBS containing 5% goat sera and 2% fetal bovine serum for 1.0 h at room temperature. After the blocking of nonspecific sites, cells were washed three times with PBS and simultaneously reacted overnight at 4°C in a humidity chamber with anti-H and the pGL121-142 antibodies in PBS ϩ 2% fetal bovine serum. After the overnight incubation cells were thoroughly washed with PBS and simultaneously stained in the dark for 1.0 h at room temperature with Alexa488 (green fluorescence)-conjugated anti-mouse and Alexa594 (red fluorescence)conjugated anti-rabbit antibodies (Molecular Probes). Unbound antibody was removed by three washes with PBS, universal mount was added (Research Genetics, Huntsville, AL), and coverslips were applied. Surface anti-HC and anti-V-like reactivity was analyzed by threechannel (green, red, and blue fluorescence) confocal microscopy (Olympus 1X71; PMT 0 -900 V; Scan speed 0.005184 s/line). Alternatively, images were captured or standard fluorescence microscopy.

RESULTS
Germline V Transcription-As previously reported, (11) the PreB acute lymphocytic leukemia (PreB ALL) cell line PreALP (26) has both and gene loci in germline configuration but expresses germline Ig light chain (LC) transcripts. To determine whether PreALP does express germline V transcripts, RT-PCR analysis was designed to specifically target germline sequences. Briefly, poly(A) ϩ RNA was prepared from either PreALP or mature B cell lines by standard methods and converted into cDNA. PCR reactions were carried out using forward 5Ј degenerate V leader (LV) oligonucleotide primers in combination with consensus reverse 3Ј primers, spanning a sequence from the germline heptamer/spacer/nonamer region (Fig. 1a). The degeneracy of the designed primers allows comprehensive coverage of the known V (LV 1-6) gene families (24). As shown in Fig. 1a, a 360-bp germline V transcript was detected in the RT-PCR reactions in which degenerate primers from the LV3 and LV5 gene family pools were used. To rule out the germline V transcription as an isolated phenomenon of the cell line PreALP, RT-PCR experiments were carried out with the cell line Blin-1, also known to express germline V transcripts (27). As also shown in Fig. 1a, a comparable amplification of the 360-bp germline V transcript was obtained with Blin-1 and PreALP mRNAs, thus ruling out a PreALP cell line idiosyncrasy.
We next investigated whether the PreALP cells have spontaneously undergone LC rearrangement, by testing the expression of rearranged V-transcripts. To that end, RT-PCR experiments were carried, using the forward (5Ј) LV primer pool in combination with a reverse (3Ј) C region primer. Fig. 1b shows that whereas the mature B cells Daudi and B207 expressed the expected 450-bp rearranged V transcripts, PreALP routinely lacked conventional LC expression.
Cloning and Sequencing of Full-length Germline V cDNA-The germline V expression revealed by RT-PCR am-plification depicted transcripts that could be assigned to distinct LV gene families. Namely, one strong and predominant amplified with degenerate primer pools that related to the gene family V3 and a weaker one, whose amplifying primers related to the gene family V5 (Fig. 1a). Although the primer pools (LV 1-6 in Fig. 1a) do correspond to bona fide LV genes (24), the degeneracy of the oligonucleotides could result in cDNA amplifications with LV gene family overlap. Therefore, the identity of the transcript encoding the putative V-like molecule could only be established by the cloning and sequencing of the full-length germline V cDNA. In keeping with this rationale, a cDNA library that originated from the cell line PreALP was screened, using an equimolar ratio of the RT-PCRgenerated V3 and V5-like cDNA probes. A single cDNA of 700 bp was identified, cloned, and sequenced, revealing an open reading frame of 180 amino acids that predicted a polypeptide with a deduced molecular mass of 19 kDa (Fig. 2).
The deduced V-like protein exhibited three distinct features: (a) a variable domain from amino acids 1-120 that is identical to the V4 gene p06312 (28) and possesses a signal peptide and three complementarity determining regions (CDRs); (b) the V4 sequence similarity ends at amino acid 120 and has no identifiable J segment; and (c) it has a unique segment from amino acids 121-180 that includes the heptamer-spacer-nonamer region, which is a bona fide signature of its germline configuration. Intriguingly, the V4 identity of the V-like protein, as deduced from its full-length cDNA, appeared to be at odds with the LV3 and LV5 gene assignments of the primers that amplified the germline transcripts (Fig. 1a). However and as stressed earlier, the degeneracy of the oligonucleotide primer pools used in the amplification of V-like cDNA cannot provide unambiguous V gene identity, whereas FIG. 1. Germline V transcription by RT-PCR. a, depicts the PCR strategy and the results for the amplification of germline V transcripts, using RT-generated cDNA templates from the PreB ALL cell line PreALP (26). The 5Ј forward primers are consensus oligonucleotides covering each of the known leader V (LV 1-6) gene families (24). The 3Ј reverse primer was a germline consensus sequence, spanning a region at the heptamer/spacer/nonamer segment. The numbers 1-6 shown in each reaction lane represent the corresponding LV 1-6 (the respective primer sequences have been previously reported) (27). Lanes 7 and 8 show a comparative RT-PCR amplification of the germline V mRNA between the two PreB ALL cell lines PreALP and Blin-1, in which an equimolar 5Ј forward LV 1-6 primer mix were used in combination with the 3Ј-reverse germline primer. b, shows the RT-PCR scheme and results for the detection of rearranged LC transcripts, using mRNA purified from the PreB ALL cell line PreALP (lanes 1-7) and the mature B cell lines Daudi (lane 8), and B207 (lane 9). The 5Ј forward primers are also the consensus oligonucleotides covering each of the known leader V (LV 1-6) sequences, which were used in combination with an oligonucleotide designed from the constant (C) region. Lane 7 in panel b represents an equimolar mixture of the 5Ј LV primers used in combination with the 3Ј C primer. Controls for the compatibility of LV mix and C primers that were used for the amplification of rearranged LC transcripts are presented in lanes 8 and 9, in which mRNA from mature Daudi and B207 cell lines was efficiently amplified. The random appearance of RT-PCR fragments below the predicted 360-or 450-bp products result from secondary RNA structures that are incompletely reverse-transcribed (Frances et al. unpublished data). the full-length cDNA cloning and sequencing can. These data show that PreB cells express germline V transcripts that can potentially translate into a V-like protein and support the hypothesis of a surrogate V-like protein that could be a structural component of the PreB receptor.
Expression and Germline Identity of the V-like Protein-Although the cloned cDNA does encode a V-like protein, there is no evidence that the germline V transcript can indeed translate into the deduced protein. To assess the expression of the V-like protein and at the same time confirm its germline identity, two affinity-purified anti-V antibodies were generated. The first antibody p42-58 was raised against a peptide that spanned amino acids 42-58 from the deduced V4 sequence of the cloned V-like cDNA. In contrast, the reactivity of second antibody pGL121-142 aimed at the exclusive recognition of germline V4-like proteins by the use of a peptide within amino acids 121-142, which included the heptamer-spacernonamer region of the V4-like protein. Fig. 3a shows the results of the Western blot analysis obtained with the PreB cell lines PreALP and Blin-1, which revealed the antibody p42-58 reactivity to V-like proteins of the predicted molecular size. In contrast, the cell line 1E8, a mature B cell derivative of Blin-1 (29), expressed only trace levels of the V-like molecule. These results are consistent with the productive translation of the germline V transcripts expressed by the PreALP and Blin-1 cells and stress the relevance of their expression during B cell ontogeny. The germline nature of the V-like molecule was subsequently confirmed by the specific reactivity of the affinitypurified antibody pGL121-142, which can only detect proteins that possess the cDNA-deduced germline V4 structure (Fig.  3b). The likelihood for the existence of other V-like proteins involving distinct but closely related V genes was again prompted by the detection of additional protein bands revealed by both p42-56 and pGL121-142 antibodies. Because the anti-V p42-58 antibody may target an epitope common to other V-like proteins, its stronger reactivity could be anticipated, since the pGL121-142 antibody can only react with unique germline protein residues. Nevertheless, the anti-V pGL121-142 antibody does reveal (albeit with weaker strength) the same molecular species as p42-56 and further confirms the germline nature of the detected V-like proteins, whose different sizes could have alternatively derived from distinct posttranslational processing.
It is noteworthy that the V-like reactivity of the antibody pGL121-142 appeared indistinguishable when the Western blots were preceded by an immunoprecipitation step with either protein A alone or anti-HC ϩ protein A. These results are in line with the capacity of protein A to bind IgM in addition to most IgG subclasses (30). Because the Blin-1 cells concomitantly express HC and V-like proteins, the co-immunopre-cipitation and correlative detection of IgM and the surrogate V-like protein by Western blot was expected (Fig. 3, b and c). The specificity of the protein A-mediated IP experiments for the HC/V-like immune complex was demonstrated by the incapacity of the agarose-protein A conjugate to precipitate any detectable proteins from the T cell line lysates with either the anti-V-like pGL121-142 or the anti-IgM antibodies (lanes 3 and 2 in Fig. 3, b and c, respectively). Together these findings demonstrate that the PreB cells tested express V-like proteins that are detectable by two antibodies, which target two distinct regions of the same surrogate protein, and further support the notion that germline transcripts can be productively translated. The data additionally suggest that the V-like molecules may associate with HC at the surface of PreB cells to form the PreB receptor (11). The V-like Protein Is Expressed at the Cell Surface and Associates with HC-Previous studies have demonstrated that surrogate Ig-like proteins can covalently bind HC at the surface of PreB cells to form the PreB receptor (11). However, it remained to be determined whether the V-like molecule also assembles with HC at the surface of PreB cells. As in earlier experiments, the PreB cell line Blin-1 was chosen for subsequent studies because it represents a homogeneous cell population and expresses significant levels of the V-like protein (Fig. 3, a and b). To first determine whether the cell lines express the V-like protein at the cell membrane, Blin-1 cells were surface-stained using the germline V-like-specific antibody pGL121-142 and subsequently analyzed by flow cytom-  lanes 1 and 2) protein lysates, in which the germline (GL) identity of the V-like protein is confirmed by its detection with the GLV-specific antibody pGL121-142. c, shows that HC is associated with the V-like protein as revealed by the protein A-mediated IP and anti-IgM-specific Western blot (WB).  2. The V-like amino acid sequence. A PreALP cDNA library was constructed and screened using the PCR-generated (gel-purified) germline V probes. The cDNA sequence is identical to the one previously reported by Thompson et al. (22). Alignment of the V-like protein sequence to germline V4 gene subgroup p06312 (28) was performed by the Clustal method (DNASTAR, Inc., Madison, WI), revealing a 100% identity up to amino acid 120 (black contig). From there onward, the sequence is unique (gray contig) and depicts a germline V sequence that includes amino acids encoded by the heptamer/spacer/nonamer region (boxed). The amino acids spanning the peptides that were chosen for the generation of p42-58 and pGL121-142 anti-V-like antibodies are shown in bold letters.
etry. Because molecular hindrance can limit detection of complex receptor molecules by flow cytometry, an enzymatic amplification staining procedure (31) was used to assess the expression of the V-like molecule at the surface of Blin-1 cells. Whereas it is conceivable that the number of V-like sites at the surface of the Blin-1 cells could remain low, its expression was expected to be clonally distributed through the majority of the cells. Therefore, the surface-staining amplification procedure aimed at assessing an accurate estimate of the surface V-like ϩ cells. Fig. 4a demonstrates that in agreement with its putative role in the formation of the PreB receptor complex, the V-like protein can be expressed at the surface of PreB cells. The baseline reactivity of isotype-matched control Ig (fine histogram on the left side of Fig. 4a) stresses the specific and significant surface expression of the V-like molecule.
To investigate whether HC can form a protein complex with the V-like molecule, reciprocal IP and Western blot assays were carried out as follows: First and in order to analyze proteins complexed with HC, Blin-1 protein lysates were precipitated with a monoclonal anti-IgM antibody directly conjugated to agarose beads. The anti-IgM immune complex was subsequently analyzed by Western blotting for the presence of the V-like protein, using the anti-V-like pGL121-142 antibody. To reciprocally assess whether HC is complexed with the V-like molecule, IP was next carried out with the anti-Vlike pGL121-142 antibody and followed by anti-IgM specific Western blot. Fig. 4, b and c demonstrate that HC and the surrogate V-like protein are associated, as their reciprocal IP and Western blots revealed their concomitant presence in the respective immune complexes. The fact that the same antibody reactivities were completely negative when T cell protein lysates were used, further substantiated the specificity of HC/ V-like IP and Western blots (Fig. 4, b and c).
Focal Cell Surface Co-localization of the HC and V-like Proteins-Although the surface expression of the V-like protein and its physical association with HC could be revealed by flow cytometry and reciprocal IP/Western blots respectively, the results in Fig. 4 do not prove that the HC and V-like proteins can co-localize at the cell surface. Therefore, to determine whether the V-like protein has the capacity to associate with H at the surface of PreB cells, V-like and HC fluorescent surface staining of the cell line Blin-1 was carried out, using the monoclonal anti-IgM antibody (green fluorescence) and anti-germline V-specific antibody pGL121-142 (red fluorescence). Confocal microscopy (Fig. 5) demonstrates that indeed, HC and V-like molecules can be focally localized at the surface of PreB cells. More importantly and consistent with their reciprocal co-IP (Fig. 4, b and c), the green and red fluorescence overlap (yellow/orange image) depicted in Fig. 5c provides additional support for the molecular association of HC and V-like molecules at the surface of the PreB cell line Blin-1 (The negative controls validating the specificity of the antibody reactivities are accessible in the Supplementary Fig.  5, e-h). Collectively, the data presented in Fig. 5 support the notion of a PreB receptor, which can be formed by HC in conjunction with surrogate -like molecules.

V-like Expression by Normal Pro/PreB Cells-The preceding experiments have shown that PreB cell lines can express
the V-like protein in conjunction with HC. However, we have not proved that the surrogate -like protein can be expressed during the normal B cell ontogeny. To examine the surface expression of the V-like molecule and determine the significance of the -PreB receptor at distinct stages of the normal B cell maturation program, fresh human bone marrow aspirates from normal donors were obtained and sorted into a Vk-like enriched ProB/PreB subset (Fig. 6a). V-like ϩ cells were further phenotyped by flow cytometry without the enzymatic amplification step, since a more heterogeneous distribution of the -like PreB receptor can be anticipated. As depicted in Fig. 6b, the sorted ProB/PreB subset concomitantly express V-like and HC and are CD34 low CD19 low CD10 high , suggesting that surface V-like expression is relevant at the transition from ProB to PreB stages and that its assembly with HC could precede the conventional 5/VpreB receptor. These findings were further supported by the capacity of the ProB/PreB-sorted subset to express the V-like at the mRNA level (Fig. 7a).
V-like and HC Association and Assembly at the Surface of Normal ProB and PreB Cells-To biochemically prove that the formation of V-like/HC complex is also relevant for normal B cell ontogeny, protein lysates from the ProB/PreB subsets were IP and analyzed by Western blots. To that end, protein lysates from freshly sorted normal ProB/PreB cells were subjected to IP with the agarose-conjugated anti-IgM antibody. The IP protein complex was subsequently analyzed by Western blots to reveal the presence V-like and HC proteins by the respective reactivity of the anti-IgM and the anti V-like pGL121-142 antibodies. Fig. 7b demonstrates that the V-like and HC proteins are expressed and biochemically associated in normal ProB/PreB cells.
In support of these findings, immunofluorescence staining (Fig. 8) shows the focal expression of the V-like and HC proteins, thus confirming their capacity to assemble at the surface of normal ProB and PreB cells. It is noteworthy that among the surface V-like ϩ HC ϩ (1 in 4 per field), V-like ϩ HC Ϫ and Vlike Ϫ HC Ϫ can also be detected, indicating the presence of distinct ProB (V-like ϩ HC Ϫ ) and PreB (V-like ϩ HC ϩ ) cells. The results in Fig. 8 additionally suggest that the V-like protein is expressed at ProB cell stage prior to HC rearrangement. More-  Supplementary Fig. 5, e-h link). Images were X/Y mode captured, using an Olympus 1X71 confocal microscope with a ϫ60 objective magnification. over, the presence of V-like Ϫ HC Ϫ cells in the immunofluorescent field further emphasizes the specificity of the anti-IgM and anti V-like pGL121-142 antibodies. DISCUSSION In most eukaryotic cells the DNA content is several orders of magnitude higher than the one required for the coding of proteins. However, beyond the sequence annotations for protein coding, the human genome stands out with a complex gene expression diversity that breaches the limits of gene usage and function. Such diversity is largely contributed by genetic rearrangements, differential RNA processing, and alternative translation initiation mechanisms. In keeping with this notion, genes previously ascribed to only yield non-coding germline mRNAs, are increasingly being documented to productively translate into small and long polypeptides, which conform to the cell lineage phenotype and function (11,12,(32)(33)(34). Moreover, a fundamental property of mammalian antibody diversity is the rearrangement of their V[D]J gene segments, which requires the expression of SLC germline genes.
SLC genes 5 and VpreB were discovered as a result of their selective transcription by B cell precursors (13,14,(35)(36)(37). Structural studies have subsequently revealed that these genes have significant homology with conventional LC but differ in that they do not undergo rearrangement. The finding that SLC genes are expressed by PreB cells before LC and LC chains are rearranged is consistent with the notion of PreB Ag receptors. Whereas 5/14.1 is a homologue of the J-C genes, (14) the VpreB genes depict a significantly high sequence similarity to V (38,39). The present work reports the identification and molecular characterization of a V-like protein that, in analogy to the VpreB molecule, is the product of an unrearranged V gene. The finding that the V-like protein can also assemble with HC at the surface of PreB cells further supports the co-existence of and PreB receptors that can function in a coordinated manner to promote LC rearrangement. The present study further proposes that B cell precursors expressing either the or the PreB receptor could respond to antigen-like stimuli from the bone marrow microenvironment and selectively signal the rearrangement of LC or LC, respectively. It is noteworthy that the expression of the V-like molecule occurs in a developmentally specific manner that peaks at the transition from the ProB to PreB stages, thus suggesting that the PreB receptor plays an essential role in the mechanisms that lead to IgHC allelic exclusion (8). The recent finding that in the triple VpreB 1 /VpreB 2 /-5-deficient mice (indistinguishable from single knockouts) the IgHC locus remained allelically excluded (21), provides a strong support to this postulate.
Our working hypothesis proposes the -like PreB receptor as a back-up mechanism that operates in concert with the conventional 5 surrogate chains to ensure LC rearrangement and allelic exclusion. Whereas it is unquestionable that the number of splenic B cells is markedly reduced in the 5 T/T and the triple VpreB 1 /VpreB 2 /5-deficient mice, mature IgM ϩ , IgD ϩ expressing LC are significantly present at 2 weeks (4-fold less than the wild type) and 5-6 weeks of age (2-fold less than the wild type). Although these data could be a priori at odds with the recent observation that mutations in the human 5 gene result in B cell deficiency (40), it is conceivable that the remarkably distinct lifespan between mice and humans are responsible for the differences in the time required for the reconstitution and accumulation of mature B lymphocytes from -like compensatory mechanisms.
The in vivo biological relevance of the -like PreB receptor is supported by earlier work. For instance, gene knockin studies (41) demonstrated that while transgenic mice harboring multiple copies of germline V4 -J4 genes effectively rescued the 5 T/T -deficiency, the knockin of a rearranged VJ transgene only provided a partial reconstitution of 5 gene inactivation. Furthermore, it has been previously shown that active germline Ig transcription in the mouse is a prerequisite for LC rearrangement (42). Such findings are completely in tune with the fact that the targeted in vivo deletion of promoter elements results in reduced germline transcription and abrogation of the gene rearrangement (43). In addition, dual block of RelA and c-Rel (transactivating components of NFB) by in vivo gene targeting, leads to a potent inhibition of germline transcription and LC rearrangement (44).
The messages that emerge from the data reported in the present study are: (a) Whereas the function of germline transcription has been mainly regarded as a byproduct of the mechanisms that enhance chromatin accessibility to target gene loci, productive translation of the respective mRNAs might have been overlooked. The cloning of the V-like chain cDNA and the cellular and biochemical characterization of its deduced protein, provide strong support to the concept that germline transcripts (11) are not by definition sterile. (b) There are now substantive facts that support the co-existence of the and PreB receptors, which may work in concert to promote LC rearrangement (11,18) allelic exclusion (8) and Ag receptor editing (45,46).
In conclusion, the missing V-like structural component of PreB receptor has now been identified and its co-localization with H at the surface of Pro/PreB cells demonstrated. The data reported herein strongly support the concept of alternative and PreB receptors whose functions may represent back-up pathways to ensure the maturation of B lymphocytes.