Fit for life in the immune system? Surrogate L chain tests H chains that test L chains

Antibody molecules (Igs), built of heavy (H) and light (L) chains, are made by B-lymphocytes. L chains are composed of two μH chains of five Ig domains. The amino terminal domains of H and L chains are variable in amino acid sequence (V_H and V_L) particularly at the three hypervariable regions that form the antigen combining site. Assembled H-L chain pairs are stabilized by a disulfide bridge between the constant domain of the L chain and the first constant region domain of the H chain (C_H1). In addition, noncovalent interactions, especially between the V_H and V_L domains, are involved in H-L chain associations (1). The genes encoding Ig H and L chains are assembled by stepwise somatic rearrangements of gene segments encoding different part of the V regions (2). During B lymphocyte development in bone marrow of adult mice and humans, D_H segments first are rearranged to J_H segments, followed by V_H segments to D_HJ_H segments on the H chain loci and, finally, V_L-segment rearrangements to J_L segments on the L chain loci. Between the segments, at the joints, nontemplated N regions can be inserted, which all encode parts of the third complementary-determining regions (CDR3) of H chains, and also of L chains in humans, but not in mice. A wide variety of CDR3 amino acid sequences can be generated, as long as the nucleotide sequences remain in-frame and no stop codons were generated. It follows that these rearrangement processes also generate out-of-frame, nonfunctional H and L chain gene loci.

Different stages of pre-B cell development can be characterized by the rearrangement status of the Ig gene loci (3, 4). Even before V_H segments are rearranged to D_HJ_H segments, and well before V_L segments are rearranged to J_L segments, pre-B cells express a set of B lineage-specific genes called V_preB and λ₅. They encode two proteins that can assemble with each other to form an L-chain-like structure, the surrogate L chain (5–8). Although the three-dimensional structures of the V_preB and λ₅ proteins have not yet been determined, the nucleotide sequences of the genes encoding them have allowed predictions of their structures (Fig. 1). Thus, V_preB appears to have a V domain-like structure, but it lacks the last β-strand (β₇) of a typical V domain. Instead it has a carboxyl terminal end that shows no sequence homologies to any other proteins. On the other side, λ₅ appears to have a constant region—the domain with strong homologies to λL chains. Furthermore, it has, toward its amino terminal end, two functionally distinct regions. One, situated near Cλ₅, has strong homology to J regions of the λL chains, i.e., to the β₇-strand of Vλ domains. The other, i.e., the amino terminal region of the λ₅ protein, is unique in that it shows only marginal sequence homologies to Ig domain structures (5). From these sequence homologies, it has been proposed that the β₇ strand within the amino terminal portion of the λ₅ protein could provide the structure that is missing in V_preB to form a complete, however noncovalently assembled, Ig domain and, thus, could mediate assembly of the two proteins (Fig. 1). As reported by Minegishi, Hendershot, and Conley in this issue of the Proceedings (9), deletion of the β₇ strand in the amino terminal portion of the λ₅ protein completely abrogates the formation of the surrogate L chain. By measuring proper or improper folding of the protein components of the surrogate L chains and their mutants, Minegishi et al. conclude that complementation of the incomplete V_preB domain by the extra β₇-strand of λ₅ is necessary and sufficient for folding and assembly of the surrogate L chain (10–12).

Figure 1.

Structure of associated V_preB and λ₅ protein as proposed by Kudo et al. (6), the Ig-like portions of V_preB (β-strands 1–6 in red) and λ₅ (Cλ_5, β-strands 1–7, amino terminal pair β₇-strand in blue) are assumed to fold as Ig domains. The figure is modified from that published in ref. 6.

Minegishi et al. extend their studies to show, by deletional mutagenesis of the unique amino terminal portion of λ₅ and the carboxyl terminal portion of V_preB, that these two parts of the proteins do not mediate assembly of the surrogate L chain, as the two deleted forms of the proteins still assemble. On the other hand, the unique amino terminal region of the λ₅ protein appears to act as an intramolecular chaperone in preventing a proper folding of the Cλ₅ domain, when it is made in the absence of its partner, V_preB (Fig. 2). Observations from our own laboratory support this finding, as a whole series of μH chains do not associate with λ₅ protein expressed in the absence of V_preB protein in the same cell (T. Seidl and F.M., unpublished work). It is tempting to speculate that the putative β-strands of the unique portion of the λ₅ protein may, in fact, disturb the proper folding of the Cλ₅ domain by assembling into the Cλ₅ fold.

Figure 2.

Stepwise formation of the pre-B cell receptor. Isolated λ₅ protein has an improperly folded structure in which Cλ₅ (◊, blue) has not yet attained an Ig-domain structure (○) and is associated with the amino terminal unique portion of λ₅, which acts as an intramolecular chaperone (). The β₇ strand of the amino terminal λ₅ portion is folded in an Ig-domain- like fashion (). The V_preB protein is folded as an Ig-like domain, missing the β₇-strand with a unique, non-Ig carboxyl terminal portion (, red). The isolated λ₅-protein cannot displace BIP and associate with the C_H1-domain of the μH chain, because neither λ₅ (◊) nor C_H1 (□, green) are properly Ig-domain-folded. Association of V_preB with λ₅ induces an Ig-domain-structure in Cλ₅ and displaces the intramolecular chaperone (). V_preB interacts with V_H of the μH chain, Cλ₅ displaces BIP and induces an Ig-domain structure in C_H1 (○, green), forms a disulfide bridge and thus, the pre-B cell receptor.

Once the surrogate L chain has been assembled properly it awaits the production of a μH chain from a productively V_HD_HJ_H-rearranged H chain allele. Initially the μH chains bind the hsp70 chaperone BIP via their C_H1 domains (13, 14). This prevents the C_H1 domain from folding properly, and the μH chain is retained in the endoplasmic reticulum until BIP is replaced by a constant domain of a L chain (Hellman , unpublished work as quoted by Minegishi et al. in ref. 9). It can be expected that the properly folded Cλ₅ domain in the assembled surrogate L chain can do the same to the C_H1 domain of the μH chain, whereas free λ₅ protein, improperly folded in the absence of V_preB, cannot do so (Fig. 2).

However, not all μH chains, which initially are formed in pre-B cells can assemble with surrogate L chains (15, 16). In fact, to our surprise, nearly half of all newly generated μH chains were found not to be able to pair. The inability of a μH chain to pair may have more than one structural reason. Some CDR3 regions may either be too bulky or repulsive for proper interactions with surrogate L chains. Certain V_H segments carried in the H chain locus, such as the V_H81x and V_HQ52 segments, may have germ-line-encoded structural features within the V domains that do not allow proper pairing with surrogate L chains. At present, we have no good explanation why such nonfitting V-region segments would be inherited and used in the generation of μH chains.

Those μH chains that fit to surrogate L chains can form a pre-B cell receptor and display it on the cell surface. Once this has been achieved several events take place in that pre-B cell. The lymphocyte-specific components of rearrangement machinery, i.e., the genes encoding RAG-1, RAG-2, and TdT are turned off (17). That is part of the mechanism by which allelic exclusion is secured, ensuring that one of the two H chain alleles in a B lineage cell expresses a protein, i.e., one pre-B cell expresses only one type of pre-B cell receptor. Furthermore, the expression of the genes encoding the surrogate L chain genes is turned off, and remains turned off, at least, until a mature B cell has been made (5). That limits the number of surrogate L-chain molecules available in a pre-B cell for forming a pre-B cell receptor. If pairing or nonpairing of surrogate L chain with μH chains is not an all or nothing phenomenon, hence, if different avidities exist between individual μH chains and the surrogate L chain, then different concentrations of associated pre-B cell receptors may exist in and on individual pre-B cells.

The formation of pre-B cell receptors on the surface of the cells also induces them to enter cell cycle and divide several times. As the clones of pre-B cells expand they dilute the surrogate L chain molecules initially present in the first cell of the clone. If the number of surrogate L chain molecules has been reduced so much that a critical number of pre-B cell receptors is no longer available in and on a given pre-B cell proliferation ceases. The higher the avidity of interaction between a μH chain and the surrogate L chain, the longer pre-B cells expressing them will expand by proliferation (Fig. 3).

Figure 3.

Formation of the pre-B cell receptor and its effect on B cell development. DJ_H-rearranged pre-B I cells (5) in contact with bone marrow stromal cells undergo asymmetric divisions that yield one cell, which retains contact with stroma and, thus, pre-B I differentiation status. The other cell loses contact to stroma, is induced to differentiate to a pre-B II cell in rearranging a V_H to D_HJ_H segment on the H chain loci. Nonproductive (np) rearrangements leads to μH chain-negative cells that die (). Productive rearrangements can lead to μH chain that cannot pair with surrogate L chain (15). Cells with nonpairing μH chains cannot form a pre-B cell receptor and die. On the other hand, pairing μH chain leads to pre-B cell receptor expression on the surface, which induces pre-B II cell proliferation. The higher the avidity, the longer is the proliferative phase (high avidity  = green, low avidity = yellow). In the simplest form of this model, the surrogate L chain assumes a ligand function for the μH chain to induce proliferation, hence, an external ligand binding to V_preB and/or V_H must not necessarily be present to induce clonal expansion of pre-B cell receptor-expressing pre-B II cells.

Once the pre-B cell exits the cell cycle the rearrangement machinery is turned on again. However, now the κ and λL chain gene loci have become accessible for V_L to J_L rearrangements, while the H chain loci, which have not yet been V_HD_HJ_H-rearranged, must have previously been closed to avoid V_H to D_HJ_H rearrangements on the second allele, i.e., to maintain allelic exclusion. Once L chains are expressed from individual, productively V_LJ_L-rearranged L chain loci in clones of μH chain-expressing pre-B cells, they are tested for their capacity to pair with a given μH chain in a given cell. The properly folded C_L domain of κL or λL chains now displaces BIP on μH chain, induces proper folding of the C_H1 domain of the μH chain and finally forms a disulfide-bonded H-L chain complex, part of the Ig molecule that then is displayed as B cell receptor for antigen on the surface. In λ₅-defective mutant mice the fitness of the V_H domains of the μH chain expressed in pre-B cells cannot be screened until L chains are expressed. V_H repertoire changes in μH chains using V_H81x and V_HQ52 segments occur in λ₅-mutant mice at the transition of a precursor to an immature surface Ig-expressing B cell, as they occur in wild-type mice at the transition from pre-B I to a pre-B II cell (Fig. 3). It remains to be seen whether surrogate L chain as stereotype for an L chain can muster all V_H domains for fitness to all V_L domains.