Proc. Natl Acad. Sci. USA
`Vol. 79, p. 6976-6978, November 1982 »
`Immuno ogy
`
`A theory -of allelic. and isotypic exclusion for immunoglobulin genes
`(heavy chain binding protein/heavy chain toxicity/‘DNA~rearrangement)
`
`MATTHIAS WABL* AND CHARLES STEINBERCI
`
`*Friedrich-Miescher-Laboratorium der Max-Planck-Gesellschatt, Spemannstrasse. 37-39, D-7400 Tiibingen, Federal Republic of Cennany; and
`*Basel Institute for Immunology, Postfach CH-4005, Basel, Switzerland
`
`Communicated by N. K. Jeme, july 29, 1982
`
`Heavy - (H) chain binding protein (BiP), which
`ABSTRACT
`binds to free immunoglobulin H chain of the p. and 7 classes, can
`be demonstratedin pre-B-cells. It is proposed that the displace-
`ment of BiP from H chain by light (L) chain terminates the activity
`of the enzyme system, L-generase, which catalyzes DNA‘rear- -
`rangement at the L chain loci, generatingthe complete gene which
`may or may not -be functional. This ensures allelic and isotypic
`exclusion for the L chain loci.’ It is further proposedthat those cells
`that productively rearrange both alleles atthe H chain locus are
`eliminated by the “H chain toxicity” effect.
`
`At each of the loci encoding Igs, only one (at most) of the two
`alleles is functional in any one lymphocyte (1, 2); this is called
`allelic exclusion, and it ensures that all.of the antibody mole-
`cules produced by a cell have the same specificity. Further-
`more, in a given lymphocyte, either K or A light (L) chain, but
`not both, can combine with heavy (H) chain to form a complete
`Ig molecule; this is called L chain isotypic exclusion. A simple
`way to achieve allelic exclusion would be to inactivate one of
`the homologous chromosomes, as is done with the X chromo-
`some in female mammals (3). This is thought not to be the case;
`in any event, it is clear that both homologs of a chromosome
`cannot be completely inactivated, as would be necessary to ex-
`plain L chain isotypic exclusion. For the H chain locus, it has
`been shown that somatic segregation (4) is not the mechanism
`for allelic exclusion (5). A model explaining allelic exclusion as
`a stochastic (chance) process. of DNA rearrangement to create
`functional genes for Ig chains has been proposed (6), but this
`assumption does not, by itself, explain how the cell would stop
`rearrangement when it is able to synthesize a complete lg
`molecule.
`
`A note on nomenclature
`
`We find it difficult to discuss the present problem without ex-
`tending, and slightly modifying, current terminology. The rec-
`ognition that Ig variable (V) and constant (C) regions must be
`encoded by different genetic units led to the slogan, "two genes,
`one polypeptide chain” (7). The folly of considering the units
`encoding V and C regions to be genes became fully. apparent
`only much laterwhen the information for a H chain was found
`to be encoded by four such units: V, D, j, and C. Are we to call
`them all genes? No, the only reasonable course seems to be to
`retreat back to “one gene, one polypeptide chain."* The H
`chain gene is then created by joining V, D, ], and C gene seg-
`ments (or DNA segments). This is consistent with traditional
`genetical usage in which genes are defined by functional tests
`for alle1ism—e. g. , cis-trans tests. We propose that this process
`of gene creation be called “geniture. ” Geniture involves DNA
`rearrangement, just as chromosomal inversions, deletions, and
`
`The publication costs ofthis article were defrayed in part by page charge
`payment. This article must therefore be hereby marked “advertise-
`ment" in accordance with 18 U. S. C. §1734 solely to indicate this fact.
`
`6976
`
`translocafions are termed chromosomal rearrangements. It is,
`of course, possible that DNA rearrangements arise by recom-
`bination between sister chromatids. Geniture must be me-
`diated by enzymes. We propose that these enzymes be called
`“generases. " L-generase and H-generase mediate the geniture
`of L and H chain genes.
`In addition to thesemolecular genetical neologisms, we find
`the need to increase the vocabulary of cellular immunology by
`one. In the course of B-lymphocyte differentiation,.a B cell, by
`definition, makes complete Ig molecules. The B cellis derived
`from apre-B-cell, which, by definition, makes cytoplasmic H
`chain only. We assume that. there exists a type of cell which is-
`already committed to the B lineage but which makes neither
`H nor L chain, and we propose that such a cell be called an
`"ur-B-cell. "
`
`Experimental basis of the theory
`
`Unrearranged L Chain Loci. Although both alleles at the H
`chain locus are, as a rule, rearranged in a B lymphocyte, this
`is not so for the L chain loci (reviewed in ref. 8). In K chain-
`producing myelomas, for instance, 50% have an unrearranged
`K chain allele, and in most of these, both sets of A chain alleles
`are in the germ-line configuration. (6).
`H Chain Binding Protein (BiP). Recently there has been
`found, in some cells, a protein called BiP that binds to free H
`chain of the p. and 7 classes. BiP can be demonstrated in a
`murine pre-B-cell line, in some myelomas, and in hybridomas
`derived from them (unpublished data). It also can be seen in.
`the polyacrylamide gel patterns of proteins from some mouse
`myeloma variants (9).
`H Chain Toxicity. Although it is easy to isolate variants of
`plasma cell hybridomas that have lost expression of H chain,
`variants that have lost L chain expression are rare (10-12). This
`observation suggested that free H chain is toxic to the cell (11,
`12). However, there are important exceptions to this rule.
`Plasma cells of people with H chain disease synthesize H chain
`only, as do murine pre-B-cells and hybridomas derived from
`them (13). In the mouse, some myeloma variants induced by
`chemical mutagens also produce H chain but no L chain (9). In
`hybridomas producing two L and. two H chains, a L chain is
`rarely lost before a H chain, but in hybridomas producing three
`L and three H chains,.chain loss is random (12).
`
`The basic theory
`
`L Chain. It is clear that L-generase activity, which promotes
`geniture at the L chain loci, ceases after L chain synthesis be-
`
`Abbreviations: H, heavy; BiP, H chain binding protein; L, light; V,
`variable; C, constant; Ill-H, pseudo-H; H- or L-generase, enzyme Apro-
`moting gene rearrangement at the H or L chain locus, respectively.
`* Ifwe consider what takes place at the RNA level, we might even have
`to advance in reverse—i.e., one gene, many polypeptide chains.
`
`Mylan v. Genentech
`Mylan V. Genentech
`IPR2016-00710
`IPR2016-00710
`Genentech Exhibit 2073
`
`Genentech Exhibit 2073
`
`

`
`Immunology: Wabl and Steinberg
`
`Proc. Natl. Acad. Sci. USA 79 (1982)
`
`6977
`
`gins. How does a cell know that a complete Ig molecule is being
`produced? This leads us to the first postulate of our theory.
`Postulate 1. Displacement of the BiPfrom the H chain by the
`L chain terminates L-generase activity.
`There are several possibilities for the precise relationship
`between BiP binding and L-generase activity. Perhaps free BiP
`inhibits L-generase. A more interesting possibility is that the
`BiP—H chain complex is the L-generase. In any event, geniture
`at the L chain loci ceases when there is made a L chain that can
`combine with the preexisting H chain. In our theory, regulation
`of allelic exclusion for the L chain loci takes place at the genetic
`level rather than at the cellular level.
`
`H Chain. Having found a putative “signal” to terminate L-
`generase activity, the obvious course would be to look for an-
`other signal to terminate H-generase activity. But the experi-
`mental facts discussed above seem to say (to us, at least) that
`there is an asymmetry or nonequivalence between H and L
`chains. Indeed, because unrearranged H chain alleles are rare
`or nonexistent in B cells, there is no reason to postulate H-gen-
`erase inactivation after geniture on one homolog is achieved.
`The extreme form of the stochastic theory regards allelic exclu-
`sion at the H chain locus as the statistical consequence of a high
`error rate. Our theory accounts for allelic exclusion at this locus
`with the second postulate.
`Postulate II. Those cells that productively rearrange both
`alleles at the H chain locus are eliminated by "H chain toxicity.”
`Iffree H chain is toxic to the cell, it seems odd for H chain
`synthesis to begin first (13). Perhaps the concentration of H
`chain in pre-B-cells is simply too low to be toxic. We know,
`however, that at least some of the H chain in pre-B-cells is not
`free—it is bound to BiP. An interesting possibility is that BiP
`neutralizes the toxic effect of H chain. If this is so, then the
`maximum amount of BiP that a cell can make must be sufficient
`
`to neutralize the amount of H chain that can be produced from
`one active allele but not enough to neutralize two alleles’ worth
`of H chain. Thus, when H chain from only one homolog is being
`produced, BiP would prevent it from being toxic. But if a cell
`produced H chain from both homologs, not enough BiP would
`be available to prevent the toxic effect of the free H chain, and,
`as a consequence, the cell would be eliminated and not be seen
`among the population of B cells.
`There are some experimental findings that are consistent
`with the notion that BiP is involved in neutralizing H chain tox-
`icity. Pre-B-cell hybridomas produce large amounts of H chain;
`nevertheless, they grow normally and show no signs of H chain
`toxicity. These lines also produce large amounts of BiP, which
`can be precipitated along with H chain (unpublished data). BiP
`also was found in myeloma variants that produce H but not L
`chain (9). Thus, it is reasonable to assume that in all of these
`cells, H chain is not, in fact, free but rather is bound to BiP.
`In any event, according to our theory, the regulation of allelic
`exclusion for the H chain locus takes place at the cellular level
`rather than at the genetic level. That is, those cells that do not
`achieve allelic exclusion at this locus are eliminated from the
`population of B lymphocytes.
`
`A detailed model
`
`The two postulates discussed above constitute the bare bones
`of our theory. We will now flesh them out in order to present
`a concrete model for the physiology of Ig gene rearrangement
`(Fig. 1). We imagine that an ur-B-cell makes mRNA for BiP,
`but that nascent BiP remains attached to the ribosomal complex
`until liberated by attachment to a H chain. We further imagine
`that the ur-B-cell is able to make a mRNA and protein chain
`from some part of the unrearranged H chain locus, and we call
`
`/'
`
`l cell death
`
`ll \
`..;-_:-/.J.l”
`
`W-I chains
`
`-1---
`
`a
`
`BiP on
`ribosorne
`
`a
`
`b
`
`I
`-\
`WH chain + BiP:
`H-generase
`
`\ a
`
`-——-5
`
`unbound
`BiP
`
`/,
`
`L chain
`
`ig
`
`h
`
`H chain 9 BiP:
`L-generase
`
`FIG. 1. Schematic illustration of the proposed model.
`
`

`
`6978
`
`Immunology: Wabl and Steinberg
`
`Proc. Natl Acad. Sci. USA 79 (1982)
`
`that it is most unlikely that any such model can be true in every
`detail. The point is that one can make a complete model based
`on our two postulates and that the model is consistent with a
`variety of known facts. We hope that our model will be useful
`in suggesting further experiments.
`
`We are indebted to Dr. Georges Kéhler for many helpful discussions
`in prose and in verse. We are also indebted to Prof. Niels Jeme for
`helping to separate the wheat from the chaff. The Basel Institute for
`Immunology was founded and is supported by F. Hoffmann-La Roche
`and Co., Basel, Switzerland.
`
`6
`
`lmmunol. 114,
`
`this pseudo-H chain (I/I-H). We note that a mRNA transcribed
`from an unrearranged K chain locus has been described (14).
`The nascent BiP combines with ab-H to create the H-generase,
`which ultimately leads to the synthesis of H chain, with the re-
`sult that the ur-B-cell becomes a pre-B-cell. In this model ll;-H
`is part of an enzyme (H-generase) that destroys the gene (un-
`rearranged H chain) that codes for it.
`As more H chain is synthesized, it combines with nascent
`BiP, a process that “regulates” the amount of BiP, so that there
`is little or no free BiP or H chain. This type of regulation has
`its limits, however; when H chain is synthesized fast enough,
`the amount of BiP mRNA, not the freeing of nascent BiP, limits
`BiP production, and free toxic H chain will accumulate. As dis-
`cussed above, this should happen in a cell with two productive
`H chain genes, and such a cell will be eliminated. The amount
`of BiP mRNA must, of course, vary in different types of cells;
`this model requires only that BiP is produced in amounts pro-
`portional to H chain. There is a precedent for this type of pro-
`portionality-the immunoglobulin molecule. Although the ab-
`solute rate of synthesis ofboth H and L chains differs by several
`orders of magnitude in small resting B lymphocytes and plasma
`cells, the ratio of H chain to L chain is essentially the same (15).
`We imagine that the H chain-BiP complex is the L-generase,
`which leads to the synthesis of L chain, with the result that the
`pre-B-cell becomes a B cell. The L chain displaces the BiP from
`the complex, and as a result, there is no generase. This dis-
`placement does not require that the initial binding of L chain
`to H chain be drastically stronger than that of BiP to H chain
`because the BiP—H chain binding is noncovalent (unpublished
`data), while disulfide bridges are usually formed between the
`L and H chains.
`We will refrain from attempting to justify most of the details
`of the above model and will not list alternatives because we feel
`
`1. Weiler, E. (1965) Proc. Natl Acad. Sci. USA 54, 1765-1772.
`2
`Pemis, B., Chiappino, G., Kelus, A. S. & Cell, P. G. H. (1965)
`]. Exp. Med. 122, 853-875.
`3. Lyon, M. (1961) Nature (London) 190, 372-373.
`4. Ohno, S. (1966) In Vitro 2, 46-60.
`5. Wabl, M. R. 8: Tenkhoff, M. (1982) Proc. Natl Acad. Sci. USA 79,
`606-607.
`Perry, R. P., Kelley, D. E., Coleclough, C., Seidman, ]. G.,
`Leder, P., Tonegawa, S., Matthyssens, C. 8: Weigert, M. (1980)
`Proc. Natl Acad. Sci. USA 77, 1937-1941.
`Symposium (1972) Fed. Proc. Fed. Am. Soc. Exp. Biol 31,
`176-209.
`Early, P. & Hood, L. (1981) cell 24, 1-3.
`Morrison, S. L. & Scharff, M. D.
`(1975) J .
`655-659.
`Coffino, P. & Scharff, M. D. (1971) Proc. Natl Acad. Sci. USA 68,
`219-223.
`Wilde, C. D. & Milstein, C.
`462-467.
`Kéhler, G. (1980) Proc. Natl Acad. Sci. USA 77, 2197-2199.
`Burrows, P. D., Le Jeune, M. 8: Kearney, J. F. (1979) Nature
`(London) 280, 838-841.
`Van Ness, B. G., Weigert, M., Coleclough, C., Mather, E. L.,
`Kelley, D. E. 8: Perry, R. P. (1981) Cell 27, 593-602.
`Baumal, R. 8: Scharff, M. D. (1973) ]. Immunol. 111, 448-456.
`
`(1980) Eur. ] .
`
`lmmanol 10,
`
`
`
`
`
`r—->-—»—->—-r—-r—-