- Email: [email protected]
Epigenetics: Monoallelic Expression in the Immune System
Current Biology, Vol. 12, R108–R110, February 5, 2002, ©2002 Elsevier Science Ltd. All rights reserved. PII S0960-9822(02)00674-7
Epigenetics: Monoallelic Expression in the Immune System Cristina Rada1 and Anne C. Ferguson-Smith2
Epigenetic modifications to DNA and chromatin programme important genome functions including gene expression, chromosomal architecture and stability, and the maintenance of developmental states. Recent findings further implicate epigenetic modifications in the control of allelic choice in the immune system.
Pathogenic agents, such as bacteria and viruses, are recognized by the B and T cells of the immune system via a vast repertoire of antigen receptors. The diversity of this repertoire is created from a relatively small number of V, D and J gene segments that become shuffled by DNA rearrangements in somatic cells in a process termed V(D)J recombination (reviewed in [1]). Importantly, only one of the two alleles of the antigen receptor is expressed in B cells and T cells, whereas the other one is silenced by a mechanism known as allelic exclusion. It is often the case that the expressed allele is rearranged completely and productively but the silent one is not. It was therefore believed that rearrangement occurred randomly with respect to the two alleles, but that further rearrangement was inhibited when a functional receptor was expressed on the surface of the cell. Now two new papers [2,3] provide evidence that both the initial selection of which allele is to be rearranged, as well as the maintenance of silencing of one of the alleles, is under the control of epigenetic mechanisms. The RAG recombinase, which is responsible for V(D)J recombination, recognises a short consensus sequence known as a recombination signal sequence or RSS that is present in the flanks of the segments to be rearranged. However, since all antigen receptor gene segments are flanked by this signal, it occurs many times in the genome and therefore the accessibility of these sequences to RAG has to be carefully regulated (Figure 1). In vitro, the relative positioning of the RSSs with respect to nucleosomes can affect the initiation of V(D)J recombination by RAG [4]. In vivo, the accessibility of a locus is usually indicated by active transcription, by features affecting the chromatin structure including histone acetylation, and by the replication time of the DNA during S phase of the cell cycle — transcriptionally silent heterochromatic regions replicate later in S phase than actively transcribed euchromatic loci. Replication asynchrony is therefore considered to be a suitable marker for allelic 1MRC
Laboratory of Molecular Biology, Hills Road, Cambridge CB2 2QH, UK. E-mail: [email protected]. 2Department of Anatomy, University of Cambridge, Downing Street, Cambridge CB2 3DY, UK. E-mail: [email protected]
Dispatch
differences in the epigenetic state of the locus. Chromatin immunoprecipitation experiments have shown that histone acetylation correlates strikingly with V(D)J recombination in the T-cell receptor loci [5], and the histones in the immunoglobulin heavy chain locus are hyperacetylated in a stepwise manner, domain by domain, so that the DJ regions are made accessible first while the distal V segments remain hypoacetylated [6]. The data presented in the recent paper by Mostoslavsky and colleagues [2] address whether epigenetic features differentially mark the two heavy chain alleles before rearrangement actually occurs. One modification known to show a differential association between active and repressed alleles is DNA methylation. Differential DNA demethylation has been observed at κ light chain immunoglobulin loci prior to rearrangement, with the demethylated allele undergoing preferential rearrangement [7]. The mechanistic relationship between demethylation and rearrangement is unclear, however, since demethylation does not necessarily lead to rearrangement [8]. The question thus remains of how the differential accessibility of the two alleles is first established, and furthermore how allelic silencing is maintained following recombination. Mostoslavsky and colleagues [2] provide strong evidence that differential accessibility probably underlies the choice of which allele of the antigen receptor locus is to be rearranged first. They show, using fluorescence in situ hybridization analysis of the immunoglobulin receptor locus that there is asynchrony of DNA replication in S phase with one allele replicating considerably earlier than the other allele. This asynchrony of replication is random with respect to the parental origin of the two alleles. It is evident in germ cells and in very early preimplantation stages in the developing embryo, in embryonic stem cells and in fetal fibroblasts, and therefore long before the lymphocyte lineage is determined. Importantly once established, the replication asynchrony is clonally stable through many cell divisions, at least in the immortalized primary cell lines examined. Mostoslavsky et al. [2] suggest that it is the rearranged allele that replicates earlier than the non-rearranged one. This assay is conducted on cells after rearrangement so does not conclusively correlate differential replication with rearrangement. Definitive proof requires a more difficult experiment in a suitable cell type which allows early replication and subsequent rearrangement to be shown for the same allele. Nonetheless, the authors conjecture that the early replicating allele of the immunoglobulin receptor locus is the one chosen for recombination and that early epigenetic programming is influencing that choice. How is early replication controlled? Unlike methylation, the replication asynchrony was independent of the presence of the two main enhancers that control
Current Biology R109
Figure 1. The orderly rearrangement of the B-cell receptor chains during lymphocyte development. Surrogate Light The B-cell receptor is assembled from two light chain heavy chains and two light chains. Each chain chain is assembled from V, D and J gene V-to-Jκ D-to-JH V-to-DJH (V and J only in the light chain) segments glH glκ VDJH VJκ DJH glκ VDJH glκ by DNA rearrangement. After RAG activaVDJH VJκ glκ glκ DJH glκ DJH glκ DJH glκ tion, both heavy chain alleles initiate D to DJH glH J rearrangement first (DJH), while the light κ) remain unrearranged chain loci (κ κ). One of the heavy chain Mature B Pre-B Immature B (germline, glκ Undifferentiated Pro-B alleles initiates rearrangement from the V precursor segment locus into the DJH. After a sucCurrent Biology cessful rearrangement, the heavy chain expression on the cell surface sends a signal to prevent any further rearrangements in the heavy chain locus and rearrangement of one of the light chain κ loci gets started. The genomic configuration of the two alleles of both loci during development is represented by the letters (V, D, J and H for the heavy chain, κ for the light chain). Red represents the early replicating allele as described in [2]. The centromeric domain association of the silent allele in mature B cells described by Skok et al. [3] is shown as the central black dot. The first undifferentiated stage prior to rearrangement shows one allele early replicating (red) while the hypothetical centromeric localization of the late replicated allele (black) is predicted.
RAG
Heavy chain
Heavy chain
expression of the κ loci. Targeted inactivation of either the intron enhancer or the 3′ enhancer of the κ loci did not affect the differential replication timing. This is perhaps not surprising since the asynchronous pattern of replication is established long before the enhancers become active. Even in the presence of a rearranged transgene for the κ light chain (which prevents rearrangement in the endogenous loci) the two silent germline κ loci replicated asynchronously. Therefore, the epigenetic accessibility of one of the alleles is controlled by a heritable mechanism, which is not directly dependent on the transcriptional control of the locus. In fact, the developmental time at which asynchronous replication of the immunoglobulin loci is established may be similar to that of other monoallelically expressed loci such as those on the X chromosome in females. The idea put forward by Mostoslavsky and colleagues [2] is that the monoallelic marking of antigen receptor loci is part of a more general mechanism that regulates marking of genes on autosomes that are monoallelically expressed later in development. This general mechanism differs from imprinting, in which germline modifications are retained throughout preimplantation development [9], because the epigenetic features responsible for the replication asynchrony at these other autosomal loci do not appear to be retained throughout preimplantation development but are re-established around implantation [2]. Once recombination has taken place, how is the non-rearranged allele kept transcriptionally silent? Skok and colleagues [3] have shown that the rearranged and the silent allele, respectively, of the B cell antigen receptor loci have different intranuclear localizations, with the silent allele being localised close to positions in the nucleus in which centromeric heterochromatin reside. Localisation to a centromeric zone was also observed with isotypically excluded alleles (isotype refers to the light chain of the BCR, either κ or λ; the κ light chain is used first and more frequently than λ in mice, consequently, most B cells have unrearranged λ
B-cell receptor
loci and >90% of the B cells show centromeric localization of both λ loci). This mechanism provides a useful and robust way of maintaining expression patterns in allelic silencing, but clearly also operates more generally in situations where there is silencing without distinction of the alleles. One of the interesting questions raised by these observations is how the localization of silent late-replicating domains is targeted to centromeric zones within the nucleus. The identification of histone variants associated with centric heterochromatin as well as the inactive X [10,11], suggests that the histone composition of the chromatin might be one of the mechanisms for targeting domains to centromeric late-replicating zones. The localization of the silent allele in the study of Skok et al. [3] is mediated by the DNA-binding protein, Ikaros. The relationship, if any, between intranuclear zones of inactive chromatin, the specific unmodified chromatin-associated proteins associated with this localization, and epigenetic modifications involved in the silencing of DNA or chromatin, remains to be determined. What happens when the first rearrangement attempt is not productive? V(D)J recombination is an imprecise recombination process and a successful protein is not always generated. This is in fact a frequent event because potentially two thirds of all rearrangements are non-productive. Critically, the second allele then needs to be rearranged and expressed. While neither study has managed to follow asynchronous replication or centromeric localization during the rearrangement process or indeed in situations in which the second allele undergoes rearrangement, the implication is that the epigenetic control of rearrangement and of silencing has to be reversible to allow the other previously silent allele to rearrange. It will be interesting to determine the allele-specific replication pattern and changes in nuclear localization under those circumstances [12]. The number of genes known to be regulated by epigenetic mechanisms is increasing all the time. In
Dispatch R110
addition to the classical examples of epigenetic gene regulation such as imprinted genes [9] and genes on the inactive X chromosome [13], the list includes interleukin genes [14,15], olfactory receptor genes [16] and now the B cell and T cell receptor genes. There are interesting parallels in these systems in which monoallelic control plays a key role. For example, all are associated with common organizational features such as clustering of the gene members and both long- and short-range local control mechanisms. All exhibit both epigenetic establishment and maintenance properties, most are epigenetically programmed in cell types in which they may not be transcriptionally active and, where tested, exhibit asynchronous replication of DNA. Differences include the extent of involvement of parental-origin-specific parameters in their initiation, the extent to which DNA methylation may be involved and the precise timing of the initiation and establishment of their epigenetic states. Interestingly, it has been proposed that the emergence of DNA methylation in the mammalian genome resulting in epigenetic control of gene expression originally evolved as a host-defence mechanism against parasitic or foreign elements such as viruses and transposable elements [17]. It is tempting to extrapolate this hypothesis to include epigenetic control of the host’s immunological response to pathogens. Whether keeping track of pathogens and odorants, or keeping control of development, epigenetic mechanisms are here to stay. The new papers raise many questions, including how asynchronous DNA replication is initiated and maintained, and which protein and chromatin mechanisms are involved in moving silent or active domains around in the nucleus. New insights contributing to our understanding of these fundamental functional genomic issues are expected to come from further studies in epigenetics. References 1. Hesslein, D.G., and Schatz, D.G. (2001). Factors and forces controlling V(D)J recombination. Adv. Immunol. 78, 169–232. 2. Mostoslavsky, R., Singh, N., Tenzen, T., Goldmit, M., Gabay, C., Elizur, S., Qi, P., Reubinoff, B.E., Chess, A., Cedar, H. et al. (2001). Asynchronous replication and allelic exclusion in the immune system. Nature 414, 221–225. 3. Skok, J.A., Brown, K.E., Azuara, V., Caparros, M.L., Baxter, J., Takacs, K., Dillon, N., Gray, D., Perry, R.P., Merkenschlager, M. et al. (2001). Nonequivalent nuclear location of immunoglobulin alleles in B lymphocytes. Nat. Immunol. 2, 848–854. 4. Kwon, J., Imbalzano, A.N., Matthews, A., and Oettinger, M.A. (1998). Accessibility of nucleosomal DNA to V(D)J cleavage is modulated by RSS positioning and HMG1. Mol. Cell 2, 829–839. 5. McBlane, F., and Boyes, J. (2000). Stimulation of V(D)J recombination by histone acetylation. Curr. Biol. 10, 483–486. 6. Chowdhury, D., and Sen, R. (2001). Stepwise activation of the immunoglobulin µ heavy chain gene locus. EMBO J 20, 6394–6403. 7. Mostoslavsky, R., Singh, N., Kirillov, A., Pelanda, R., Cedar, H., Chess, A., and Bergman, Y. (1998). Kappa chain monoallelic demethylation and the establishment of allelic exclusion. Genes Dev. 12, 1801–1811.
8. Cherry, S.R., Beard, C., Jaenisch, R., and Baltimore, D. (2000). V(D)J recombination is not activated by demethylation of the kappa locus. Proc. Natl. Acad. Sci. U.S.A. 97, 8467–8472. 9. Ferguson-Smith, A.C., and Surani, M.A. (2001). Imprinting and the epigenetic asymmetry between parental genomes. Science 293, 1086–1089. 10. Palmer, D., O’Day, K., Trong, H., Charbonneau, H., and Margolis, R. (1991). Purification of the centromere-specific Protein CENP-A and demonstration that it is a distinctive histone. Proc. Natl. Acad. Sci. U.S.A. 88, 3734–3738. 11. Costanzi, C., and Pehrson, J.R. (1998). Histone macroH2A1 is concentrated in the inactive X chromosome of female mammals. Nature 393, 599–601. 12. Gasser, S.M. (2001). Positions of potential: nuclear organization and gene expression. Cell 104, 639–642. 13. Maxfield Boumil, R., and Lee, J.T. (2001). Forty years of decoding the silence in X chromosome inactivation. Hum. Mol. Genet. 10, 2225–2232. 14. Riviere, I., Sunshine, M.J., and Littman, D.R. (1998). Regulation of IL-4 expression by activation of individual alleles. Immunity 9, 217–228. 15. Kelly, B.L., and Locksley, R.M. (2000). Coordinate regulation of the IL-4, IL-13, and IL-5 cytokine cluster in Th2 clones revealed by allelic expression patterns. J. Immunol. 165, 2982–2986. 16. Serizawa, S., Ishii, T., Nakatani, H., Tsuboi, A., Nagawa, F., Asano, M., Sudo, K., Sakagami, J., Sakano, H., Ijiri, T. et al. (2000). Mutually exclusive expression of odorant receptor transgenes. Nat. Neurosci. 3, 687–693. 17. Bestor, T., and Tycko, B. (1996). Creation of genomic methylation patterns. Nat. Genet. 12, 363–367.
Epigenetics: Monoallelic Expression in the Immune System Cristina Rada1 and Anne C. Ferguson-Smith2
Epigenetic modifications to DNA and chromatin programme important genome functions including gene expression, chromosomal architecture and stability, and the maintenance of developmental states. Recent findings further implicate epigenetic modifications in the control of allelic choice in the immune system.
Pathogenic agents, such as bacteria and viruses, are recognized by the B and T cells of the immune system via a vast repertoire of antigen receptors. The diversity of this repertoire is created from a relatively small number of V, D and J gene segments that become shuffled by DNA rearrangements in somatic cells in a process termed V(D)J recombination (reviewed in [1]). Importantly, only one of the two alleles of the antigen receptor is expressed in B cells and T cells, whereas the other one is silenced by a mechanism known as allelic exclusion. It is often the case that the expressed allele is rearranged completely and productively but the silent one is not. It was therefore believed that rearrangement occurred randomly with respect to the two alleles, but that further rearrangement was inhibited when a functional receptor was expressed on the surface of the cell. Now two new papers [2,3] provide evidence that both the initial selection of which allele is to be rearranged, as well as the maintenance of silencing of one of the alleles, is under the control of epigenetic mechanisms. The RAG recombinase, which is responsible for V(D)J recombination, recognises a short consensus sequence known as a recombination signal sequence or RSS that is present in the flanks of the segments to be rearranged. However, since all antigen receptor gene segments are flanked by this signal, it occurs many times in the genome and therefore the accessibility of these sequences to RAG has to be carefully regulated (Figure 1). In vitro, the relative positioning of the RSSs with respect to nucleosomes can affect the initiation of V(D)J recombination by RAG [4]. In vivo, the accessibility of a locus is usually indicated by active transcription, by features affecting the chromatin structure including histone acetylation, and by the replication time of the DNA during S phase of the cell cycle — transcriptionally silent heterochromatic regions replicate later in S phase than actively transcribed euchromatic loci. Replication asynchrony is therefore considered to be a suitable marker for allelic 1MRC
Laboratory of Molecular Biology, Hills Road, Cambridge CB2 2QH, UK. E-mail: [email protected]. 2Department of Anatomy, University of Cambridge, Downing Street, Cambridge CB2 3DY, UK. E-mail: [email protected]
Dispatch
differences in the epigenetic state of the locus. Chromatin immunoprecipitation experiments have shown that histone acetylation correlates strikingly with V(D)J recombination in the T-cell receptor loci [5], and the histones in the immunoglobulin heavy chain locus are hyperacetylated in a stepwise manner, domain by domain, so that the DJ regions are made accessible first while the distal V segments remain hypoacetylated [6]. The data presented in the recent paper by Mostoslavsky and colleagues [2] address whether epigenetic features differentially mark the two heavy chain alleles before rearrangement actually occurs. One modification known to show a differential association between active and repressed alleles is DNA methylation. Differential DNA demethylation has been observed at κ light chain immunoglobulin loci prior to rearrangement, with the demethylated allele undergoing preferential rearrangement [7]. The mechanistic relationship between demethylation and rearrangement is unclear, however, since demethylation does not necessarily lead to rearrangement [8]. The question thus remains of how the differential accessibility of the two alleles is first established, and furthermore how allelic silencing is maintained following recombination. Mostoslavsky and colleagues [2] provide strong evidence that differential accessibility probably underlies the choice of which allele of the antigen receptor locus is to be rearranged first. They show, using fluorescence in situ hybridization analysis of the immunoglobulin receptor locus that there is asynchrony of DNA replication in S phase with one allele replicating considerably earlier than the other allele. This asynchrony of replication is random with respect to the parental origin of the two alleles. It is evident in germ cells and in very early preimplantation stages in the developing embryo, in embryonic stem cells and in fetal fibroblasts, and therefore long before the lymphocyte lineage is determined. Importantly once established, the replication asynchrony is clonally stable through many cell divisions, at least in the immortalized primary cell lines examined. Mostoslavsky et al. [2] suggest that it is the rearranged allele that replicates earlier than the non-rearranged one. This assay is conducted on cells after rearrangement so does not conclusively correlate differential replication with rearrangement. Definitive proof requires a more difficult experiment in a suitable cell type which allows early replication and subsequent rearrangement to be shown for the same allele. Nonetheless, the authors conjecture that the early replicating allele of the immunoglobulin receptor locus is the one chosen for recombination and that early epigenetic programming is influencing that choice. How is early replication controlled? Unlike methylation, the replication asynchrony was independent of the presence of the two main enhancers that control
Current Biology R109
Figure 1. The orderly rearrangement of the B-cell receptor chains during lymphocyte development. Surrogate Light The B-cell receptor is assembled from two light chain heavy chains and two light chains. Each chain chain is assembled from V, D and J gene V-to-Jκ D-to-JH V-to-DJH (V and J only in the light chain) segments glH glκ VDJH VJκ DJH glκ VDJH glκ by DNA rearrangement. After RAG activaVDJH VJκ glκ glκ DJH glκ DJH glκ DJH glκ tion, both heavy chain alleles initiate D to DJH glH J rearrangement first (DJH), while the light κ) remain unrearranged chain loci (κ κ). One of the heavy chain Mature B Pre-B Immature B (germline, glκ Undifferentiated Pro-B alleles initiates rearrangement from the V precursor segment locus into the DJH. After a sucCurrent Biology cessful rearrangement, the heavy chain expression on the cell surface sends a signal to prevent any further rearrangements in the heavy chain locus and rearrangement of one of the light chain κ loci gets started. The genomic configuration of the two alleles of both loci during development is represented by the letters (V, D, J and H for the heavy chain, κ for the light chain). Red represents the early replicating allele as described in [2]. The centromeric domain association of the silent allele in mature B cells described by Skok et al. [3] is shown as the central black dot. The first undifferentiated stage prior to rearrangement shows one allele early replicating (red) while the hypothetical centromeric localization of the late replicated allele (black) is predicted.
RAG
Heavy chain
Heavy chain
expression of the κ loci. Targeted inactivation of either the intron enhancer or the 3′ enhancer of the κ loci did not affect the differential replication timing. This is perhaps not surprising since the asynchronous pattern of replication is established long before the enhancers become active. Even in the presence of a rearranged transgene for the κ light chain (which prevents rearrangement in the endogenous loci) the two silent germline κ loci replicated asynchronously. Therefore, the epigenetic accessibility of one of the alleles is controlled by a heritable mechanism, which is not directly dependent on the transcriptional control of the locus. In fact, the developmental time at which asynchronous replication of the immunoglobulin loci is established may be similar to that of other monoallelically expressed loci such as those on the X chromosome in females. The idea put forward by Mostoslavsky and colleagues [2] is that the monoallelic marking of antigen receptor loci is part of a more general mechanism that regulates marking of genes on autosomes that are monoallelically expressed later in development. This general mechanism differs from imprinting, in which germline modifications are retained throughout preimplantation development [9], because the epigenetic features responsible for the replication asynchrony at these other autosomal loci do not appear to be retained throughout preimplantation development but are re-established around implantation [2]. Once recombination has taken place, how is the non-rearranged allele kept transcriptionally silent? Skok and colleagues [3] have shown that the rearranged and the silent allele, respectively, of the B cell antigen receptor loci have different intranuclear localizations, with the silent allele being localised close to positions in the nucleus in which centromeric heterochromatin reside. Localisation to a centromeric zone was also observed with isotypically excluded alleles (isotype refers to the light chain of the BCR, either κ or λ; the κ light chain is used first and more frequently than λ in mice, consequently, most B cells have unrearranged λ
B-cell receptor
loci and >90% of the B cells show centromeric localization of both λ loci). This mechanism provides a useful and robust way of maintaining expression patterns in allelic silencing, but clearly also operates more generally in situations where there is silencing without distinction of the alleles. One of the interesting questions raised by these observations is how the localization of silent late-replicating domains is targeted to centromeric zones within the nucleus. The identification of histone variants associated with centric heterochromatin as well as the inactive X [10,11], suggests that the histone composition of the chromatin might be one of the mechanisms for targeting domains to centromeric late-replicating zones. The localization of the silent allele in the study of Skok et al. [3] is mediated by the DNA-binding protein, Ikaros. The relationship, if any, between intranuclear zones of inactive chromatin, the specific unmodified chromatin-associated proteins associated with this localization, and epigenetic modifications involved in the silencing of DNA or chromatin, remains to be determined. What happens when the first rearrangement attempt is not productive? V(D)J recombination is an imprecise recombination process and a successful protein is not always generated. This is in fact a frequent event because potentially two thirds of all rearrangements are non-productive. Critically, the second allele then needs to be rearranged and expressed. While neither study has managed to follow asynchronous replication or centromeric localization during the rearrangement process or indeed in situations in which the second allele undergoes rearrangement, the implication is that the epigenetic control of rearrangement and of silencing has to be reversible to allow the other previously silent allele to rearrange. It will be interesting to determine the allele-specific replication pattern and changes in nuclear localization under those circumstances [12]. The number of genes known to be regulated by epigenetic mechanisms is increasing all the time. In
Dispatch R110
addition to the classical examples of epigenetic gene regulation such as imprinted genes [9] and genes on the inactive X chromosome [13], the list includes interleukin genes [14,15], olfactory receptor genes [16] and now the B cell and T cell receptor genes. There are interesting parallels in these systems in which monoallelic control plays a key role. For example, all are associated with common organizational features such as clustering of the gene members and both long- and short-range local control mechanisms. All exhibit both epigenetic establishment and maintenance properties, most are epigenetically programmed in cell types in which they may not be transcriptionally active and, where tested, exhibit asynchronous replication of DNA. Differences include the extent of involvement of parental-origin-specific parameters in their initiation, the extent to which DNA methylation may be involved and the precise timing of the initiation and establishment of their epigenetic states. Interestingly, it has been proposed that the emergence of DNA methylation in the mammalian genome resulting in epigenetic control of gene expression originally evolved as a host-defence mechanism against parasitic or foreign elements such as viruses and transposable elements [17]. It is tempting to extrapolate this hypothesis to include epigenetic control of the host’s immunological response to pathogens. Whether keeping track of pathogens and odorants, or keeping control of development, epigenetic mechanisms are here to stay. The new papers raise many questions, including how asynchronous DNA replication is initiated and maintained, and which protein and chromatin mechanisms are involved in moving silent or active domains around in the nucleus. New insights contributing to our understanding of these fundamental functional genomic issues are expected to come from further studies in epigenetics. References 1. Hesslein, D.G., and Schatz, D.G. (2001). Factors and forces controlling V(D)J recombination. Adv. Immunol. 78, 169–232. 2. Mostoslavsky, R., Singh, N., Tenzen, T., Goldmit, M., Gabay, C., Elizur, S., Qi, P., Reubinoff, B.E., Chess, A., Cedar, H. et al. (2001). Asynchronous replication and allelic exclusion in the immune system. Nature 414, 221–225. 3. Skok, J.A., Brown, K.E., Azuara, V., Caparros, M.L., Baxter, J., Takacs, K., Dillon, N., Gray, D., Perry, R.P., Merkenschlager, M. et al. (2001). Nonequivalent nuclear location of immunoglobulin alleles in B lymphocytes. Nat. Immunol. 2, 848–854. 4. Kwon, J., Imbalzano, A.N., Matthews, A., and Oettinger, M.A. (1998). Accessibility of nucleosomal DNA to V(D)J cleavage is modulated by RSS positioning and HMG1. Mol. Cell 2, 829–839. 5. McBlane, F., and Boyes, J. (2000). Stimulation of V(D)J recombination by histone acetylation. Curr. Biol. 10, 483–486. 6. Chowdhury, D., and Sen, R. (2001). Stepwise activation of the immunoglobulin µ heavy chain gene locus. EMBO J 20, 6394–6403. 7. Mostoslavsky, R., Singh, N., Kirillov, A., Pelanda, R., Cedar, H., Chess, A., and Bergman, Y. (1998). Kappa chain monoallelic demethylation and the establishment of allelic exclusion. Genes Dev. 12, 1801–1811.
8. Cherry, S.R., Beard, C., Jaenisch, R., and Baltimore, D. (2000). V(D)J recombination is not activated by demethylation of the kappa locus. Proc. Natl. Acad. Sci. U.S.A. 97, 8467–8472. 9. Ferguson-Smith, A.C., and Surani, M.A. (2001). Imprinting and the epigenetic asymmetry between parental genomes. Science 293, 1086–1089. 10. Palmer, D., O’Day, K., Trong, H., Charbonneau, H., and Margolis, R. (1991). Purification of the centromere-specific Protein CENP-A and demonstration that it is a distinctive histone. Proc. Natl. Acad. Sci. U.S.A. 88, 3734–3738. 11. Costanzi, C., and Pehrson, J.R. (1998). Histone macroH2A1 is concentrated in the inactive X chromosome of female mammals. Nature 393, 599–601. 12. Gasser, S.M. (2001). Positions of potential: nuclear organization and gene expression. Cell 104, 639–642. 13. Maxfield Boumil, R., and Lee, J.T. (2001). Forty years of decoding the silence in X chromosome inactivation. Hum. Mol. Genet. 10, 2225–2232. 14. Riviere, I., Sunshine, M.J., and Littman, D.R. (1998). Regulation of IL-4 expression by activation of individual alleles. Immunity 9, 217–228. 15. Kelly, B.L., and Locksley, R.M. (2000). Coordinate regulation of the IL-4, IL-13, and IL-5 cytokine cluster in Th2 clones revealed by allelic expression patterns. J. Immunol. 165, 2982–2986. 16. Serizawa, S., Ishii, T., Nakatani, H., Tsuboi, A., Nagawa, F., Asano, M., Sudo, K., Sakagami, J., Sakano, H., Ijiri, T. et al. (2000). Mutually exclusive expression of odorant receptor transgenes. Nat. Neurosci. 3, 687–693. 17. Bestor, T., and Tycko, B. (1996). Creation of genomic methylation patterns. Nat. Genet. 12, 363–367.