Molecular Biology of the Cell - Volume 3 - Alberts B., Bray D., Lewis J., Raff M., Roberts K., Watson J. 1994

From Cells to Multicellular Organisms
The Immune System
How Antibody Diversity Is Generated

According to available estimates, even in the absence of antigenic stimulation, a mouse can produce many millions of different antibody molecules—this constitutes its preimmune repertoire. Because the antigen-binding sites of many Antibodies can cross-react with a multitude of similar yet distinct antigenic determinants, the preimmune repertoire is presumably large enough to provide a suitable antigen-binding site for almost every potential antigenic determinant.

Antibodies are Proteins, and proteins are encoded by genes. Consequently, antibody diversity poses a challenging genetic problem: how can the number of types of antibodies produced in an Organism exceed the number of genes in its genome? (It is estimated, for example, that The Human Genome contains fewer than 10⁵ genes.) However, the problem is not as daunting as it might seem at first glance. Because both L and H chains contribute to the antigen-binding site, an animal with 1000 genes encoding L chains and 1000 genes encoding H chains could produce 1000 × 1000 different combinations from the products of these genes, i.e., 10⁶ different antigen-binding sites (assuming that any L chain can pair with any H chain). Nevertheless, The Immune System has evolved unique genetic mechanisms that enable it to generate a nearly unlimited number of diverse L and H chains. These mechanisms can vary depending on the animal species; for instance, they differ considerably between chickens and mammals. Below we will focus on the mechanisms operating in mammals.

Class="center">

Fig. 18-30. An experiment that directly demonstrated that DNA undergoes rearrangement during B-Cell development. DNA was extracted from a mouse plasma cell tumor (myeloma) synthesizing a specific Ig light chain, and from a 13-day mouse embryo, which does not yet produce antibodies. Both DNA samples were digested with a restriction endonuclease, and the resulting fragments were separated by agarose gel Electrophoresis. Fragments containing sequences that encode the C region of the L chain and fragments encoding a specific V region of the L chain were then detected (using Southern blotting with two radioactive DNA probes: one complementary to the coding sequence for the V region and the other to the C region of the specific myeloma L-chain mRNA) (Section 18.3). In the myeloma cell DNA, the sequences encoding the C and V regions were found within the same DNA fragments, whereas in the embryonic DNA they were located on separate fragments (as was also true for DNA from another myeloma tumor producing a different L chain; not shown in the diagram).

18.4.1. During B-cell development, antibody genes are assembled from individual Gene segments [20]

Studies of the Amino acid sequences of myeloma proteins led to the hypothesis that the variable (V) and constant (C) regions of each Ig chain might be encoded by two separate gene segments that are somehow joined together in the DNA prior to their expression. The first direct evidence for DNA rearrangement during B-cell development was obtained in 1976 by comparing DNA from early mouse embryos, which are incapable of producing antibodies, with DNA from a mouse myeloma cell line producing antibodies. As the experiments showed, the specific sequences encoding the V and C regions used by the myeloma Cells resided on the same restriction fragment in these cells, but on two different restriction fragments in the embryos. Consequently, at some stage of B-Cell Differentiation, the DNA sequences encoding antibody molecules undergo rearrangement (Fig. 18-30).

It is now known that for each type of Ig chain—κ light chains, λ light chains, and heavy chains—There is a separate "pool" (an extensive array) of gene segments that can be utilized in the synthesis of an individual polypeptide chain. These pools are located on different Chromosomes, and each typically contains A large number of gene segments encoding the V region of the Ig chain and a somewhat smaller number of segments encoding the C region. During B-cell development, the complete coding sequence for each of the two Ig chains to be synthesized is assembled via Site-Specific Recombination (Section 5.4.7), joining the sequences encoding the V and C regions together. In addition to uniting individual gene segments, these rearrangements also activate METABOLISM/31.html">Transcription from the gene promoter—likely by altering the relative positions of enhancers and silencers that act on the promoter (Section 10.2.7). Therefore, an Ig chain can be synthesized only after DNA rearrangement has taken place. The process of gene segment joining increases The Diversity of antigen-binding sites, driven by several factors that we examine below.

18.4.2. Each V region is encoded by more than one gene segment [21]

When the DNA sequences encoding the V and C regions were examined, it turned out that the C region of a given Ig chain is encoded by only a single gene segment (the C segment), whereas each V region is encoded by more than one segment. The V region of each L chain is encoded by a DNA sequence assembled from two gene segments—a long V segment and a short joining J segment (not to be confused with the protein J chain, which is encoded elsewhere in The Genome; see Section 18.2.5). Figure 18-31 illustrates the genetic mechanisms involved in L-chain formation.

Fig. 18-31. The V-J joining process associated with The formation of the κ light chain in the mouse. In germline DNA (where immunoglobulin genes are not expressed and therefore not rearranged), a cluster of four J gene segments is separated from the C gene segment by a short intron, and from approximately 300 V gene segments by thousands of nucleotide pairs. The J segment encodes about 15 C-terminal Amino Acids of the V region; the site of joining between the V and J segments coincides with the third hypervariable region. The V segments on the chromosome are often organized into clusters of homologous families, which appear to span a region of more than 50,000 nucleotide pairs. During B-cell development, a specific V segment (in the case shown here, V₃) is transposed precisely adjacent to one of the J segments (here, J₃). The "extra" J genes (J₄) and the intron are transcribed (along with the joined V₃, J₃, and C segments) and subsequently excised via RNA splicing. This yields mRNA molecules in which the V₃, J₃, and C sequences lie adjacent to one another. These mRNAs then

The V region of Each H chain is encoded by a DNA sequence assembled from three gene segments: a V segment, a J segment, and a D segment (diversity gene segment). The Organization OF THE gene segments involved in H-chain synthesis is shown in Fig. 18-32.

The large number of inherited V, J, and D segments encoding Ig chains contributes significantly to antibody diversity in its own right. However, this contribution is further amplified by the combinatorial association of different segments (so-called combinatorial diversity).

For example, in the mouse, each of the roughly 300 V segments in the κ light-chain pool can join with any of the four J segments (see Fig. 18-31), meaning this pool can encode at least 1,200 (300 × 4) different κ V regions. (The mouse λ light-chain pool contains only two V segments; in humans, it contains considerably more.) Similarly, any of the approximately 1,000 V segments in the mouse H-chain pool can combine with any of the four J segments and at least 12 D segments, yielding a minimum of 48,000 (1000 × 4 × 12) different heavy-chain V regions. These estimates are very approximate because the exact number of V gene segments in these pools remains unknown.

The most important mechanism generating antibody antigen-binding site diversity is combinatorial diversity, which arises from the assembly of inherited V, J, and D gene segments in various combinations. As we have seen, in the mouse this mechanism alone makes it possible to produce at least 1,000 different V_L regions and roughly 50,000 different V_H regions. These two sets can subsequently generate 5 × 10⁷ combinations, i.e., distinct antigen-binding sites. In addition to this, the joining mechanism itself, as we shall see, vastly multiplies the number of combinations—probably by a factor of more than 1,000. As a result, this total number far exceeds the total number of B lymphocytes in a mouse (~10⁸).

Fig. 18-32. In the mouse, the gene segment pool for the H chain appears to contain about 1,000 V segments, at least 12 D segments, 4 J segments, and an ordered cluster of C segments, each encoding an H chain of a specific class. The D segment encodes from 1 to 15 amino acids within the third hypervariable region of the V region. The figure is not drawn to a uniform scale, and many details are omitted. For example:

1) there are four C_γ segments (C_γ1, C_γ2a, C_γ2b, and C_γ3);

2) each C segment is composed of multiple exons (see Fig. 18-27);

3) the J_H1 and C_μ segments are separated by approximately 200,000 nucleotide pairs; much like the V_κ segments, the V_H segments are organized on the chromosome into clusters of homologous families. H-chain formation involves the same genetic mechanisms as L-chain formation (as shown in Fig. 18-31), except that two DNA rearrangements are required instead of one: first, a D segment joins to a J segment, and then a V segment joins to this complex.

18.4.3. Inprecise joining of gene segments increases V-region diversity [21]

Gene segments that can be separated by hundreds of thousands of nucleotide pairs join to form functional sequences encoding V_L or V_H regions. The details of this mechanism are not fully understood. Flanking each gene segment are conserved DNA sequences that likely serve as recognition sites for site-specific recombination Enzymes (Section 5.4.7); this ensures that only appropriate segments recombine with one another. For example, a V segment always joins exclusively to a J or D segment, never to another V segment (Fig. 18-33).

In most instances of site-specific recombination, DNA joining is precise, but this is not the case for antibody gene segments. A variable number of NUCLEOTIDES are frequently lost from the ends of the recombining segments; in the case of H chains, one or more randomly chosen nucleotides may also be inserted. Such random deletions and insertions of nucleotides at the junctions between segments immensely amplify the diversity of V-region-coding sequences generated by recombination, particularly within the third hypervariable region. In this context, increasing diversity comes at a high price, because it frequently causes a reading frameshift and, consequently, the creation of a nonfunctional gene. During B-cell development, such "nonproductive" joining events occur very frequently.

The Fate of the intermediate DNA segment upon joining two Ig gene segments depends on their relative orientation. If the transcriptional orientation of the two segments is the same, the intermediate DNA is excised as a circular molecule and discarded (deletion joining), as shown in the figure; if the segments are oriented in opposite directions, the intervening DNA is retained in an inverted orientation (inversion joining). Biochemically, the processes are identical in both cases.

Fig. 18-33. Two Types of recognition signals used in joining V-region gene segments in the L- and H-chain gene segment pools. A. Localization of the signals. For simplicity, only two gene segments of each type are shown. Signals of both types consist of conserved sequences (shown at the bottom) 7 and 9 Base Pairs (bp) long, separated by a non-conserved spacer; in one case the spacer length is 21-23 bp (approximately two turns of the DNA double helix), and in the other

11-12 bp (approximately one turn). Only compatible gene segments can join: this occurs exclusively when a "one-turn" spacer combines with a "two-turn" spacer (the so-called 12/23 rule). B. The recognition signals are thought to be recognized, in turn, by the site-specific recombination enzyme (recombinase) that catalyzes the joining.

18.4.4. Antigen-driven somatic hypermutation fine-tunes antibody production [22]

Following immunization, there is typically a gradual increase in the affinity of antibodies for the immunizing antigen over time. This phenomenon is known as affinity maturation, and it results from the accumulation of somatic Mutations in the V-region-coding sequences following the antigenic stimulation of B lymphocytes. This was most conclusively demonstrated by studying The production of oligoclonal antibodies (Section 18.1.6) in inbred mice, in which the V regions of L or H chains are encoded predominantly by a single combination of gene segments. Because the initial sequences encoding the V region are identical, mutations occurring within them over time can be readily detected. In such studies, activated B cells harvested from individual mice at various times after immunization are hybridized with non-secreting myeloma cells, yielding hybridoma cells, each of which produces a single specific type of antibody (Section 4.5.4). These immortal cells serve as an unlimited source of RNA and DNA encoding antibody V regions. The Nucleic Acids can then be sequenced to identify Changes in the original V-region-coding sequences. It has been shown that point mutations rapidly accumulate in such sequences over time following repeated immunization. The frequency of somatic mutations in these sequences is estimated to be 10^-3 per base pair per cell generation, which is about a million times greater than the spontaneous mutation frequency in other genes. Consequently, this process was named somatic hypermutation. It is believed to occur during the antigen-driven activation of B cells into memory cells rather than during activation into antibody-secreting cells. Memory B cells are formed primarily in the lymphoid follicles of Secondary Lymphoid Organs (see Fig. 18-8).

Because B-cell proliferation is stimulated by antigen binding, any mutation occurring during the Immune Response that increases the affinity of the membrane-bound antibody molecule will promote the preferential proliferation of the B cell synthesizing that antibody, especially as the antigen concentration decreases over time following immunization. Thus, affinity maturation is the result of repeated cycles of somatic hypermutation followed by antigen-driven Selection during the humoral response.

18.4.5. The joining of antibody gene segments is regulated to ensure B-cell monospecificity [23]

Numerous experimental findings have demonstrated that, as predicted by the clonal selection theory, each individual B cell produces antibodies with only one type of antigen-binding site; that is, B cells are monospecific.

B-cell monospecificity can be important for at least two reasons. First, if each cell could produce more than one type of antigen-binding site, some of them might generate both useful antibodies and antibodies directed against self Antigens (autoantibodies); this would hinder the selection for reactivity against foreign antigens while maintaining tolerance to self antigens. Second, monospecificity ensures that all antibodies produced by a single cell are composed of two identical halves and therefore contain two identical antigen-binding sites; this enables secreted antibodies to form extensive cross-linked antigen networks, facilitating antigen elimination (see Fig. 18-13).

The requirement for monospecificity implies that a mechanism must exist whereby, upon activation of Ig genes during development, each B cell generates only one type of V_L region and one type of V_H region. Because B cells (like all somatic cells) are diploid, each possesses six gene segment pools encoding antibodies: two H-chain pools, one from each parent, and four

L-chain pools—one ϰ and one λ from each parent. Thus, DNA rearrangements could in principle occur independently in each H- and L-chain pool; by assembling different V-region-coding sequences in different pools, a single cell could potentially produce up to eight Different types of antibodies, each with its own antigen-binding site. In reality, however, each B cell utilizes only two of the six pools: one of the four L-chain gene pools and one of the two H-chain pools (Fig. 18-34). Thus, each B cell must choose

not only between the ϰ and λ L-chain pools, but also between the maternal and paternal gene pools. The expression of only the maternal or only the paternal allele of a given gene in a cell is called allelic exclusion; this phenomenon appears to be unique to genes encoding antibodies (as well as genes encoding the closely related T-cell receptor proteins, see Section 18.6.1). Maternal and paternal autosomal genes for other proteins are expressed at approximately equal levels in a cell.

The mechanisms of allelic exclusion and the selection of ϰ or λ L chains during B-cell development remain unclear. Available evidence suggests that this occurs via feedback regulation of DNA rearrangement: apparently, a functional rearrangement in one gene segment pool suppresses rearrangements in the remaining pools encoding a polypeptide chain of the same type. Some of this evidence comes from experiments in which previously assembled cloned genes were injected into The Nucleus of a fertilized mouse egg. This yields Transgenic Mice (Section 5.6.10) whose lymphocytes all contain the rearranged gene. For instance, in B-cell clones isolated from transgenic mice carrying a rearranged L-chain gene, the rearrangement of endogenous L-chain genes is typically suppressed. Similarly, the Introduction of a rearranged μ gene generally suppresses the rearrangement of endogenous H-chain genes.

Fig. 18-34. Sequential selection of activated Ig genes that must occur in developing B cells to ensure they produce antibodies with only one type of antigen-binding site.

The Cell must select one of the four L-chain gene segment pools and one of the two H-chain gene segment pools. During development, a single H-chain gene pool is first activated in the progenitor cell, which becomes a pre-B cell producing only free heavy chains of the μ class. Following a period of intense proliferation, one light chain pool (ϰ or λ) is activated in the pre-B cell, and it matures into a B cell producing a unique IgM molecule.

The assembly of V-region-coding sequences in a developing B cell appears to occur in a strict order, beginning with the H-chain pool and proceeding segment by segment. First, on both parental chromosomes, D segments join with J_H segments. Next, V_H joins to DJ_H on one of these chromosomes. If this rearrangement yields a functional gene, the subsequent synthesis of a complete μ chain (which is always produced first among H chains) halts further rearrangements of V_H-region gene segments and triggers the initiation of V_L rearrangements. First, V_L joins to J_L within the ϰ gene segment pool. If this is unsuccessful, the other ϰ pool rearranges. If this rearrangement also fails, joining occurs in one λ pool and subsequently in the other. Finally, if a successful joining event occurs at any stage and yields L chains, they assemble with the previously formed μ chains of IgM antibody molecules, which turns off any further assembly of V_L-region sequences.

To shut off further assembly of V_H sequences, the formation of an intact μ chain apparently suffices. However, halting the assembly of V_L sequences likely requires the synthesis of a complete antibody molecule. If a developing B cell fails to assemble functional sequences encoding both V_H and V_L regions, it is unable to produce antibody molecules and presumably dies.

Although no biological differences between ϰ and λ light chains have been discovered, the presence of two separate L-chain gene segment pools clearly serves a purpose: it increases the probability that a pre-B cell that has successfully assembled a V_H-coding sequence will also successfully assemble a V_L sequence and mature into a B cell.

18.4.6. Switching between membrane-bound and secreted forms of the same antibody occurs via alterations in H-chain RNA transcripts [24]

Having examined the genetic mechanisms governing The Structure of the antigen-binding site, we now turn to the mechanisms determining the biological Properties of the antibody, namely the class of the heavy-chain constant region. While the initial choice of specific gene segments to encode the antigen-binding site is permanent for both the B cell and its progeny, the class of the synthesized C_H region changes during B-cell development. These changes are of two types: the transition from the membrane-bound form to the secreted form of the same C_H region, and A change in the antibody class (isotype switching).

Antibodies of all classes can be synthesized in both membrane-bound and soluble, secreted forms. The membrane-bound form serves as an antigen receptor On the surface of B cells. The soluble form is produced only after the cell is stimulated by antigen and needs to secrete antibodies. The two forms of antibodies differ only in the C-terminal region of the H chain: for example, in membrane-bound IgM molecules, the H chain terminates in a hydrophobic region that anchors it in The Lipid Bilayer of the B-cell Plasma Membrane, whereas the H chains of secreted IgM molecules instead possess a hydrophilic C-terminus that allows the molecules to leave the cell. The ability of a B cell to produce μ chains with two different types of constant regions initially seemed paradoxical, since a B cell contains only a single copy of the C_μ gene segment per haploid genome and utilizes only one of its two available H-chain gene pools during antibody production. The paradox was resolved when it was discovered that antigen activation of B cells leads to an alteration in the RNA Processing pathway for μ-chain transcripts in the nucleus, as shown in Fig. 18-35 (see also Section 10.4.5). A similar mechanism is involved in switching from the membrane-bound to the secreted form of antibodies of other classes.

Fig. 18-35. Upon antigen activation, a B cell switches from synthesizing plasma membrane-bound antibodies to producing the secreted form of the same antibodies by altering the H-chain mRNA transcripts generated within the cell. This is thought to result from a change in the Cleavage and 3'-polyadenylation patterns of primary H-chain RNA transcripts. The two forms of the H chain differ only in their C-terminal region: the membrane-bound form has a hydrophobic "tail" that anchors it in the membrane, whereas the secreted form has a hydrophilic tail that allows it to exit the cell. The long polyadenylated RNA transcript specifying the membrane-bound H-chain form contains splice donor and acceptor sites, making it possible to remove the RNA sequence encoding the hydrophilic tail of the secreted form. In contrast, the short polyadenylated RNA transcript specifying the secreted form possesses only a splice donor site; consequently, the RNA sequence generated by cross-cleavage and re-ligation cannot be removed.

18.4.7. B cells can switch from producing one class of antibodies to producing another [25]

During their development, many B cells switch from producing antibodies of one class to producing antibodies of other classes, a process known as class switching. All B cells begin their antibody-synthesizing journey by producing IgM molecules, which become embedded in The Plasma Membrane and serve as antigen receptors. Then, even before encountering an antigen, the majority of B cells shift to simultaneously synthesizing IgM and IgD molecules, which are used as membrane-bound antigen receptors. Upon antigen stimulation, some of these cells become activated and begin secreting IgM antibodies, which predominate in the primary humoral response (Section 18.2.5). Other antigen-stimulated cells switch to producing IgG, IgE, or IgA antibodies; memory cells carry these antibodies on their surface (often alongside IgM), while active B cells secrete them. IgG, IgE, and IgA molecules are collectively referred to as secondary-class antibodies because they apparently form only after antigenic stimulation and predominate in secondary humoral responses.

The class of an antibody is determined by the constant region of its H chain (Section 18.3.3). Therefore, the ability of B cells to alter the class of antibodies they produce without changing the antigen-binding site means that the same assembled sequence encoding the V_H region can associate with different C_H gene segments at different times. This capability of B cells has profound functional significance: it allows a specific antigen-binding site, selected by environmental antigens, to be distributed among all classes of IMMUNOGLOBULINS within an individual, thereby acquiring all the biological properties characteristic of each class.

Fig. 18-36. In B cells producing both membrane-bound IgM and IgD antibodies with identical antigen-binding sites, long RNA transcripts are synthesized containing both C_μ and C_δ sequences. These transcripts undergo splicing in two different ways, yielding mRNA molecules in which the same V_H sequence is joined to either the C_μ or the C_δ sequence. It is possible that the RNA transcripts synthesized in such cells are even longer and contain all the different C_H sequences.

Class switching occurs via two distinct molecular mechanisms. When a naive B cell transitions from producing solely membrane-bound IgM to simultaneously synthesizing membrane-bound IgM and IgD, the switch likely occurs through altered RNA processing. The cells produce long primary RNA transcripts containing both C_μ and C_δ sequences alongside the assembled V_H region sequence. Subsequent Alternative Splicing of these transcripts generates IgM and IgD molecules (Fig. 18-36). Apparently, the same mechanism operates during the switch to other classes of membrane-bound Ig when naive B cells are stimulated by antigen and mature into memory cells bearing surface IgG, IgE, or IgA as antigen receptors.

Terminal maturation of an active B cell secreting antibodies of a secondary class proceeds differently: it is accompanied by an irreversible change at the DNA level—a process termed switch recombination. This involves the deletion of all C_H segments located 'upstream' (i.e., on the 5' side along the coding strand) from the specific segment designated for expression in the cell (Fig. 18-37). Evidence that this stage of class switching involves DNA deletion was obtained from experiments with myeloma cells, which showed that cells secreting IgG lack the DNA encoding the C_μ and C_δ regions, whereas cells secreting IgA lack the DNA for the H-chain C regions of all other classes.

Fig. 18-37. An example of DNA rearrangement occurring during C_H region switching. When a B cell producing IgM antibodies via an assembled VDJ sequence is stimulated by antigen and matures into an IgA-secreting cell, the DNA segment between the VDJ sequence and the C_α gene segment is excised. Specific DNA regions (switch sequences, shown as colored circles) located upstream of each C_H segment (except C_δ) recombine with one another, resulting in the deletion of the intervening DNA segment.

18.4.8. Idiotopes of antibody molecules form The basis of the immunological network [26]

Antibodies not only protect the organism against infections but also play a crucial role in regulating immune responses themselves. The termination of a humoral response to an antigen is partly driven by secreted antibodies binding to the antigen, thereby preventing it from attaching to B cell receptors and halting further B cell stimulation. Alongside this simple feedback inhibition, antibodies can participate in a more subtle mechanism of immune regulation by acting as part of a complex immunological network.

Immunoglobulins can themselves act as antigens, and antibodies can be generated that recognize antigenic determinants in both the constant and variable regions of Ig chains. The antigenic determinants (epitopes) of the L- and H-chain variable regions located within the antibody's antigen-binding site are called idiotopes (Fig. 18-38). Each specific antigen-binding site possesses its own characteristic set of idiotopes; thus, an animal possessing millions of different antigen-binding sites will also have millions of different idiotopes. Because each individual idiotope is present in very low amounts within the body, the animal is not tolerant to its own idiotopes, and if properly immunized with one of its own antibodies, the animal's organism will mount both T- and B-cell immune responses.

One would expect an animal immunized with antigen A to first produce large amounts of antibodies against A, followed by antibodies against the idiotopes of those anti-A antibodies, and subsequently antibodies against the anti-idiotypic antibodies, and so on. Indeed, this type of reaction, yielding an antibody 'network,' has been demonstrated in certain immune responses, though its exact role in immunoregulation remains unclear.

As we will see later, T-cell receptors share much in common with membrane-bound antibodies: they exist in millions of forms, each featuring its own unique antigen-binding site and set of idiotopes. Apparently, an animal's body contains both antibodies and T-cell receptors that recognize a large portion of the organism's own idiotopes. On average, an antigen-binding site likely recognizes at least one idiotope of the organism's own immune system. Thus, the antigen-binding sites of the immune system are potentially linked in a complex network of anti-idiotypic interactions (Fig. 18-39). Much like Neurons in The Nervous system, many lymphocytes may interact more intensely with one another than with the outside world, meaning the immune response could be viewed not merely as a reaction of independent antigen-responding lymphocytes, but as a reverberating perturbation of the immunological network.

Fig. 18-38. An idiotope is an antigenic determinant of an antibody molecule located within or near the antigen-binding site. It may be formed by the V_H region, the V_L region, or both. Each distinct antigen-binding site has its own unique set of idiotopes, which constitutes its idiotype. Idiotopes are also associated with the antigen-binding sites of T-cell receptors.

Summary

Antibodies are synthesized through the participation of three pools of gene segments encoding the κ-, λ-, and H-chains, respectively. Within each pool, individual gene segments encoding different PARTS OF THE L- and H-chain variable regions are joined via site-specific recombination during B-cell differentiation. The L-chain pools contain one or more constant (C) gene segments and sets of variable (V) and joining (J) gene segments. The H-chain pool contains a set of C segments and sets of V, D, and J segments. For an antibody molecule to be synthesized, a V_L segment must join to a J_L segment to form a DNA sequence encoding the light-chain V region, while a V_H segment must join with D and J_H segments to form a DNA segment encoding the heavy-chain V region. Each assembled gene segment is then cotranscribed along with the corresponding C-region sequence, yielding an mRNA molecule that encodes the entire polypeptide chain. By combining inherited gene segments encoding the V_L and V_H regions in various ways, vertebrates can produce thousands of different L chains and thousands of different H chains, which can combine to form millions of diverse antigen-binding sites. This number is further expanded by nucleotide deletions and insertions during gene segment joining, and by somatic mutations occurring at a very high frequency within these segments following antigenic stimulation.

All B cells initially produce IgM antibodies. Some subsequently generate antibodies of other classes that nonetheless share the same antigen-binding site as the original IgM antibodies. This class switching allows identical antigen-binding sites to be distributed among antibodies with diverse biological properties.