An Evolutionary View of the Mechanism for Immune and Genome Diversity Lucia Kato, Andre Stanlie, Nasim A. Begum, Maki Kobayashi, Masatoshi Aida and Tasuku Honjo This information is current as of October 1, 2021. J Immunol 2012; 188:3559-3566; ; doi: 10.4049/jimmunol.1102397 http://www.jimmunol.org/content/188/8/3559 Downloaded from References This article cites 65 articles, 30 of which you can access for free at: http://www.jimmunol.org/content/188/8/3559.full#ref-list-1

Why The JI? Submit online. http://www.jimmunol.org/

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on October 1, 2021 Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2012 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. An Evolutionary View of the Mechanism for Immune and Genome Diversity Lucia Kato, Andre Stanlie, Nasim A. Begum, Maki Kobayashi, Masatoshi Aida, and Tasuku Honjo An ortholog of activation-induced cytidine deaminase Another mechanism for genome diversification in eukar- (AID) was, evolutionarily, the first enzyme to generate yotes is mediated by meiotic recombination, which occurs in acquired immune diversity by catalyzing conversion gametogenesis. Most of the mechanisms and molecules in and probably somatic hypermutation (SHM). AID be- meiotic recombination are highly conserved from yeast to gan to mediate class switch recombination (CSR) only humans (6). Meiotic recombination might have contributed after the evolution of frogs. Recent studies revealed that to enhance the scrambling of mutations and facilitate Dar- winian selection. the mechanisms for generating immune and genetic di- Downloaded from versity share several critical features. Meiotic recombina- Vertebrate immune diversity is mediated by two types of tion, V(D)J recombination, CSR, and SHM all require DNA-alteration mechanisms: V(D)J recombination by RAG1/ H3K4 trimethyl histone modification to specify the tar- RAG2 and somatic hypermutation (SHM)/class switch re- get DNA. Genetic instability related to dinucleotide or combination (CSR)/gene conversion (GC) by activation- induced cytidine deaminase (AID) (7, 8). Acquired immune triplet repeats depends on DNA cleavage by topoisom- diversity appears to have evolved at a very early stage of ver- http://www.jimmunol.org/ erase 1, which also initiates DNA cleavage in both tebrate evolution, because an AID ortholog, Petromyzon SHM and CSR. These similarities suggest that AID marinus cytidine deaminase (PmCDA), was identified in hijacked the basic mechanism for genome instability a lamprey, P. marinus (9). In lampreys, two AID family when AID evolved in jawless fish. Thus, the risk of in- members, PmCDA1 and PmCDA2, are expressed differen- troducing genome instability into nonimmunoglobulin tially in two types of Ag-recognizing cells that correspond to loci is unavoidable but tolerable compared with the ad- T cell and B cell precursors, which express variable lymphocyte vantage conferred on the host of being protected against receptors (VLR) A and VLRB, respectively. RAG1/RAG2 and pathogens by the enormous Ig diversification. The V(D)J recombination appeared later in vertebrate evolution. Journal of Immunology, 2012, 188: 3559–3566. Given their gene structure, which lacks introns, and their by guest on October 1, 2021 recognition of a specific DNA sequence, RAG1/RAG2 may have been introduced as retrotransposons, as the result of an enome stability is critical for the survival of living infection (10). Thus, it is likely that the evolution of AID was organisms. However, complete genome stability would the origin of acquired immune diversity. G have prevented the evolution of organisms, as well as Recent findings suggest that the SHM and CSR mediated by novel biological activities. Thus, an appropriate balance between AID have mechanisms similar to those of meiotic recombi- genome stability and instability has been maintained during nation, TAM, and triplet contraction/expansion (3–5, 11–15). evolution. There are several endogenous mechanisms in addition Some of these processes are shared by V(D)J recombination to exogenous genome-altering mechanisms, such as viral infec- (8). In this review, we attempt to discuss, from an evolu- tion, irradiation, and DNA-damaging chemicals. It is well tionary point of view, the mechanisms for acquired immune established that high-frequency transcription enhances muta- diversity and classical genetic alterations and propose that genesis; this phenomenon is called transcription-associated mu- AID adopted the basic mechanism of genome instability to tation (TAM) (1). In addition, in the many triplet-associated generate immune diversity in lymphocytes. diseases, such as Huntington’s disease (2), the triplets are usually located within the transcribed region of , and the triplet General genome diversity repeats are either expanded or contracted in somatic cells. Recent Genome instability. Mutations form the basis of genetic diver- studies showed that topoisomerase 1 (Top1) is involved in both sity and instability. Living organisms have several intrinsic TAM and the triplet contraction/expansion (3–5). mechanisms for inducing mutations in DNA, in addition to

Department of Immunology and Genomic Medicine, Graduate School of Medicine, Yoshida Sakyo-ku, Kyoto 606-8501, Japan. E-mail address: [email protected]. Kyoto University, Kyoto 606-8501, Japan ac.jp. Received for publication December 27, 2011. Accepted for publication February 9, Abbreviations used in this article: AID, activation-induced cytidine deaminase; CSR, 2012. class switch recombination; GC, gene conversion; LRR, leucine-rich repeat; MRN, Mre11/Rad50/NBS1 complex; PmCDA, Petromyzon marinus cytidine deaminase; SHM, This work was supported by Grant-in-Aid for Specially Promoted Research 17002015 somatic hypermutation; TAM, transcription-associated mutation; Top1, topoisomerase 1; from the Ministry of Education, Culture, Sports, Science and Technology of Japan. UNG, uracil-DNA glycosylase; VLR, variable lymphocyte receptor. Address correspondence and reprint requests to Prof. Tasuku Honjo, Department of Immunology and Genomic Medicine, Graduate School of Medicine, Kyoto University, Copyright Ó 2012 by The American Association of Immunologists, Inc. 0022-1767/12/$16.00 www.jimmunol.org/cgi/doi/10.4049/jimmunol.1102397 3560 BRIEF REVIEWS: SIMILARITIES BETWEEN IMMUNE DIVERSITY AND GENOME INSTABILITY errors that occur during DNA replication. The best-known example is TAM, which has been observed in a wide range of organisms, including Escherichia coli, yeast, and mammals (1). Highly active transcription increases the frequency of various types of mutations: base replacement, deletion, and recombination. Among these, 2–5-bp deletions were recently shown to be catalyzed by Top1 (4, 5). The 2–5-bp deletion type of TAM involves mutation hotspots containing di- or trinucleotide repeats. Efficient transcription often induces a loosening of the helical struc- ture, resulting in the creation of excessively negative super- coiling behind the transcription machinery. Top1 is the enzyme that corrects this excessive DNA superhelix (Fig. 1A, 1B). Top1 nicks the DNA, forms a transient covalent bond with the DNA through its tyrosine residue, rotates the DNA around the helix, and then religates the cleaved ends; this set of actions can result in an increase or decrease of the super- helix (16). Notably, the active transcription of repeat sequences is predicted to form non-B DNA structures (17, Downloaded from 18). When the DNA has aberrant structures like non-B DNA, rotation around the helix may be inhibited; thus, the cleavage by Top1 may be irreversible and trigger genome instability (Fig. 1C) (4, 5, 17). Another well-known example of genome instability is as- sociated with triplet diseases, such as Huntington’s disease, http://www.jimmunol.org/ Fragile X syndrome, and myotonic dystrophy, in which triplet repeats composed of CAG, CGG, CTG, and others increase or decrease in copy number (2). These triplet repeats are FIGURE 1. Molecular mechanism for transcription-induced DNA cleavage usually located within a transcription unit and are probably by Top1. (A) Transcription creates positive and negative supercoils in the prone to form non-B structures similarly to TAM (18). Re- front and rear, respectively, of the transcriptional machinery. This excessive supercoil has to be corrected by Top1. Otherwise, non-B DNA structures cently, triplet repeat contraction was shown to be caused by tend to form under the negative supercoil condition in areas where repetitive Top1 (3). Taken together, these findings suggest that TAM sequences or inverted repeats are abundant. (B) Excessive supercoil is cor- and triplet diseases probably arise by a similar molecular rected by Top1. Top1 introduces a nick and covalently associates with the 39 by guest on October 1, 2021 mechanism, as follows: non-B DNA structures induced by the end of the DNA through its tyrosine residue. Then, Top1 rotates the DNA active transcription of repetitive sequences cause irreversible around the axis of the helix and corrects the excessive supercoil. After rotation, cleavage by Top1, triggering mutations, such as deletions and the transiently formed covalently bound DNA ends are resealed, and Top1 is C 1 duplications (Fig. 1). released from the DNA. ( - ) Non-B DNA is a preferred site for Top1 cleavage. Top1 binds preferentially to non-B DNA structures. When Top1 Meiotic recombination. Because all of the mutations that block binds and introduces a nick-type cleavage, it forms a covalent intermediate. meiotic recombination inhibit or suppress the appropriate However, rotation around the axis is inhibited by the non-B DNA structure. generation of sperm or eggs, meiotic recombination is re- Therefore, Top1 remains associated with the DNA, causing irreversibly quired for spermatogenesis and oogenesis (19). Gametogen- cleaved ends. (C-2) Chromatin modification by H3K4me is also shown (see esis undergoes two successive cell divisions after replication, Fig. 4). The covalently associated Top1 DNA complex is resected from the resulting in the generation of cells with haploid sets of genetic DNA by several enzymes, including the MRN complex and Ctip. These enzymes were shown to be involved in the resection of Spo11 in meiotic material. Meiotic recombination takes place during the first recombination. Tyrosyl-DNA phosphodiesterase 1 is another enzyme that cell division. The direct outcome of meiotic recombination is removes Top1 from the DNA. the scrambling of the genetic material on the paternal and maternal . Meiotic recombination reassembles of histone H3 on the adjacent chromatin (20–22). Because various mutations on homologous chromosomes and facili- PRDM9 is essential for meiotic recombination, the combi- tates the spreading of newly introduced mutations among nation of the cis DNA element and trans H3K4me3 histone a species population by generating new combinations of modification appears to determine the initiation site of the them. DNA cleavage for meiotic recombination. The molecular mechanisms for meiotic recombination have The enzyme that initiates DNA cleavage is well established as been extensively studied using yeast genetics (15). It was re- a type of topoisomerase II, Spo11, which forms a dimer and cently shown that meiotic recombination requires at least introduces nicking cleavage on two strands, four bases apart two marks to determine where the DNA cleavage takes place. (23). Spo11 forms a transient covalent association with DNA, The DNA sequence of a loosely conserved 13mer is the cis- similarly to other topoisomerases. The DNA-bound Spo11 marking element, which is relatively abundant within the needs to be excised to generate double-stranded cleavages by genome and usually linked to a promoter. The other mark the resection step involving the Mre11/Rad50/NBS1 (MRN) for cleavage was recently discovered by studies on PRDM9, complex and Ctip (24). Subsequently, exonucleolytic diges- which recognizes the conserved 13mer DNA sequence by its tion with exonuclease 1 exposes a long single-stranded region. zinc finger motif and carries out the trimethylation of lysine 4 The next event is homologous recombination, which is ini- The Journal of Immunology 3561 tiated by strand invasion of the homologous and mediated by many . The Holliday structure, consist- ing of four strands of DNA, is stabilized by mismatch repair enzyme members, Msh4 and Msh5, which do not function as repair proteins in this situation. Finally, the Holliday junction is cut to segregate the two chromosomes, resulting in seg- mental exchange of the genetic material between the paternal and maternal chromosomes. Notable aspects of meiotic re- combination are the marking of cleavage target chromatin by histone H3K4me3 modification and the involvement of a topoisomerase member in the DNA cleavage.

Evolution of acquired immune diversity PmCDA1 and PmCDA2, primordial orthologs of AID, are probably the oldest enzymes responsible for Ag-specific re- ceptor diversification in vertebrates (Fig. 2) (9, 25). PmCDA1 increases the rate of mutation and recombination when overexpressed in yeast (26). Unlike AID, the PmCDAs are Downloaded from expressed constitutively, because the levels and pattern of PmCDAs expression do not change in response to stimula- tion. Indeed, PmCDA1 is selectively expressed in the thymoid region where assembly of VLRA occurs, whereas PmCDA2 is expressed in VLRB-expressing lymphocytes; PmCDA1 and PmCDA2 are thought to be involved in the genetic di- http://www.jimmunol.org/ versification of these respective lymphocytes (27). The VLR cDNAs contain an invariant signal peptide followed by highly variable leucine-rich repeat (LRR) modules consisting of five FIGURE 2. Evolution of B and T lymphocytes and their AgR diversification units, which each range from 24 to 63 aa in length. The mechanisms. In jawless vertebrates, such as lampreys and hagfish, CDA1 and germline VLR genes only contain sequences coding for the in- CDA2, orthologs of AID, assemble VLRA and VLRB by a GC mechanism. variant N-terminal and C-terminal portions of the VLRs sepa- VLRA and VLRB are specifically expressed in B- and T-lineage cells, re- rated by large noncoding sequences. However, in lymphocytes, spectively. Sometime during the evolution of primordial jawed vertebrates, these noncoding intervening sequences are replaced by multiple transposon insertion generated TCR and BCR genes simultaneously with site- LRR-encoding modules to form a functional VLR gene. The specific recombinases RAG1/2. This enzyme mediates such an efficient and by guest on October 1, 2021 presence of incompletely assembled VLR genes indicates that precise reassembly of AgR genes that the system using CDA1, CDA2, VLRA, and VLRB lost its advantage for generating the AgR repertoire. However, in B- the LRR modules are put together step by step. In lamprey, lineage cells, AID that evolved from CDA2 retained its advantage because of 393 VLRA-related LRR cassettes and 454 VLRB-related its activity for generating Ag-induced Ab memory. It is likely that AID has

LRR cassettes have been characterized within a continuous maintained its GC and SHM activity in jawed vertebrates. When the Ig CH locus of ∼2 megabase pairs surrounding their respective VLR gene was duplicated in amphibians, AID’s GC activity was used for CSR. The genes (26). DNA-cleavage mechanism in CSR is very similar to that in SHM and GC. The VLR gene assembly mechanism has been postulated to The recombination-specific activity in CSR is probably also similar to that in GC, because both depend on the cis pairing of DNA segments in adjacent be of the GC type, because of the nonreciprocal insertion of loci. The genes, enzymes, and genetic events involved in the primordial im- LRR cassettes into the intervening sequences of the germline mune diversity and their descendants are in green. Myrs, million years; Rb, gene (26, 28, 29). The VLRs are highly repetitive sticky pro- recombination. teins, whose specificity is probably determined by their broad surface, which contains the many b sheets of LRR domains. The evolutionary origin of RAGs and the recombination A computational assessment of mature VLRB gene sequences signal sequence is assumed to be transposon insertion (10), predicted a potential repertoire .1014 AgRs (28). These which must have happened in primordial jawed vertebrates, striking observations indicate that the GC-type DNA rear- because no RAG ortholog has been identified in jawless fish, rangement is the most primordial molecular mechanism for such as lampreys. RAG evolution must have been generating Ag-recognition receptors. Although SHM has been coupled with segregation of the Ig ancestor gene into V, D, proposed to occur in VLRA-expressing lymphocytes, there is and J segments by transposon insertion. It is striking that both no evidence that VLRB diversification also occurs by SHM as H3K4me3 mark and recombination signal sequence are re- the result of cytidine deaminase activity of PmCDA2 (27). quired for DNA cleavage by RAG1/RAG2 (8). The creation In vertebrate evolution, CSR first took place in frogs (Fig. 2). of combinational diversity by RAGs probably replaced the Interestingly, fish AID can complement mouse AID for VLRA and VLRB system, which depends on the less efficient CSR, although fish carry out SHM but not CSR (30–32). GC event. CDA1 has a selective disadvantage in T cells, Considering the cytidine deaminase involvement in GC in which should avoid additional mutations after their selection lampreys, as well as that GC is the major mechanism of so- in the thymus. Therefore, VLRA and CDA1 in T cells were matic diversification in chicken Ig genes (33, 34), it is likely probably lost during evolution, after RAG’s insertion into that fish AID can carry out GC; however, this remains to be primitive vertebrates, which occurred sometime after the tested. evolution of jawed fish (Fig. 2). 3562 BRIEF REVIEWS: SIMILARITIES BETWEEN IMMUNE DIVERSITY AND GENOME INSTABILITY

In contrast, in B-lineage cells, even after the evolution of the of this mechanism is that the appropriate Top1 level is RAGs and Ig genes, AID maintained its selective advantage by determined by a balance with the efficiency of transcription. being required for Ag-induced affinity maturation and Ab The augmented transcription in TAM causes a relative memory through its GC mechanism, which probably coexisted reduction in Top1 at the locus. with SHM. Thus, AID carries out GC and SHM in the Ig When Top1 tries to correct the excessive supercoil by locus of chicken and other vertebrates. Mouse and human AID nicking, the non-B form DNA prevents rotation of the DNA can exhibit GC activity in chicken DT40 B cells (35, 36), around the helix, and Top1 is stuck, bound to the DNA (Fig. although it is not clear whether mouse or human AID has GC 1). To generate cleaved DNA ends, the covalently associated capability in vivo. Top1 is probably removed by several resection enzymes, such as the MRN complex and Ctip, which are used in meiotic Similarities between mechanisms for immune and genome diversity recombination. Consistent with this scenario, when resection of the Top1–DNA complex is blocked by the Top1 inter- AID has two functions: 1) the DNA cleavage that is required calating inhibitor Camptothecin (30 nM), the double- for CSR, SHM, and GC and 2) the activity specific to CSR and stranded break formation, SHM, and CSR are inhibited probably GC. The N-terminal portion of AID is required for (11, 12). It is well known that both the S and V regions DNA cleavage, whereas the C-terminal portion is required for contain direct or inverted repeats, which are prone to form CSR, as described later (35, 37–39). Two contrasting hy- non-B structures (17, 41, 42). potheses were put forward to explain how AID introduces Previously, a large number of AID target loci was proposed

DNA cleavage in the target DNA; the initially proposed Downloaded from by the experiments with chromatin immunoprecipitation RNA-editing hypothesis assumed that AID edits RNA to and microarray or genome-wide sequencing (43). These ex- generate a DNA cleaving enzyme, whereas the much straight- periments were based on the assumption that AID-interacting forward DNA-deamination hypothesis postulated that AID molecules are located at the AID-cleavage sites. Unfortu- directly deaminates target DNA to initiate DNA cleavage. nately, however, there were no consensus features among Recent results from our studies revealed striking similarities a large number of proposed AID targets. To search for between mechanisms for immune diversity and genome di- http://www.jimmunol.org/ common properties of highly efficient targets of AID, we versity. carried out a new approach to identify the cleaved loci. We Involvement of Top1 in DNA cleavage. Several lines of evidence used human BL2 cells expressing the C- indicate that Top1 is involved in the DNA cleavage in CSR terminally truncated AID fused with the estrogen receptor and SHM. First, AID activation reduces Top1 protein level; hormone-binding domain, which were briefly activated (3 h) interestingly, this reduction is not seen in a loss-of-function with tamoxifen (44). AID-induced DNA cleaved ends were mutant of AID (12). Second, artificial reduction of Top1 labeled by a biotin oligonucleotide and enriched by strepta- by small interfering RNA augments DNA-cleavage activity, vidin beads. A library of DNA fragments containing the

SHM, and CSR by AID (11, 12). Furthermore, in vivo, the cleaved ends was amplified and analyzed by whole-genome by guest on October 1, 2021 f/+ haplo-insufficiency of Top1 (by the cross of Top1 with sequencing and microarray analyses. The candidates were CAG-cre mice) dramatically augments SHM in B cells further confirmed by quantitative PCR quantitation of their isolated from Peyer’s patches (gut follicular structures in actual cleavage sites in the original cleaved end-enriched li- which germinal centers are continuously induced by brary. Finally, we confirmed that mutations accumulated in bacterial stimulation) (11). Thus, it is clear that a reduction the new loci as frequently as did Ig loci. Thus, we identified in Top1 correlates with enhanced DNA cleavage and a four hypermutating loci: MALAT1, SNHG3, BCL7A, and subsequent increase in SHM or CSR. CUX1 (44). All of these loci are flanked by abundant repet- How can AID reduce the Top1 protein level? We proposed itive sequences, which are prone to form non-B DNA struc- that the N-terminal portion of AID may interact with a pu- tures. tative cofactor N that recognizes a precursor of the microRNA Histone modification H3K4me3 as a mark for DNA cleavage. for editing (12). The edited microRNA is assumed to bind 9 In addition to the non-B DNA structure, chromatin together with the RNA-induced silencing complex to the 3 modification, especially H3K4me3, was shown to be es- untranslated region of Top1 mRNA (Fig. 3A). This binding sential for DNA cleavage of the AID target. Knockdown of would suppress the translation of Top1 mRNA, causing re- the histone chaperones facilitates chromatin transcription duction of the protein. In support of this hypothesis, specific complex or Spt6 or the elongation factor Spt5 blocked the deletion of Dicer from AID-expressing B cells drastically re- generation of H3K4me3 and DNA cleavage at the AID target duced SHM and CSR (40). Importantly, the observed re- regions without inhibition of S region transcription, and yet it duction in switching, SHM, and affinity maturation in Dicer- blocked AID-induced cleavage (13) (A. Stanlie, N.A. Begum, deficient B cells was not due to cell death, because it could be H. Akiyama, and T. Honjo, submitted for publication; N.A. still observed in B cells that were rescued from apoptosis by Begum, A. Stanlie, H. Akiyama, and T. Honjo, submitted the deletion of Bim, a proapoptotic member of the Bcl-2 for publication). Knockdown of PTIP abolished CSR with family (40). simultaneous blockade of H3K4me3 and S region trans- Non-B DNA structures as a mark for cleavage target. As mentioned cription, leaving the possibility that CSR inhibition could above, Top1 is involved in correcting the excessive positive be due to loss of target transcription (45). H3K4me3 and negative supercoils that accumulate before and after the accumulation was shown in both the V and S regions of transcription machinery, respectively. However, when Top1 is the Ig locus. The nucleosomes of the newly identified AID reduced, it cannot appropriately correct the excessive supercoil, targets (MALAT1 and others) also have the H3K4me3 mark and, as a result, more non-B DNA forms. An important aspect (44). Another well-known target of AID, the locus, is The Journal of Immunology 3563

FIGURE 3. The RNA-editing hypothesis for AID. (A) AID’s N-terminal domain is required for DNA cleavage. This activity probably depends on a putative cofactor N that interacts with the N-terminal region of AID and captures a precursor of microRNA. AID edits the microRNA precursor in the nucleus, and the edited microRNA is processed and interacts with the Top1 mRNA 39 untranslated region in a complex with the RISC protein. This interaction inhibits the translation of Top1 RNA, resulting in a reduction in the Top1 pro- tein. The reduction in Top1 protein enhances the amount of non-B form DNA in the S and V regions, which are actively transcribed. The formation of the unusual structure in the S and V regions induces irre- versible cleavage by Top1. (B) The C-terminal domain Downloaded from of AID is involved in editing the mRNA encoding a putative bending factor. The mRNA is captured by a putative cofactor C, which interacts with the C-ter- minal region of AID. The edited mRNA is translated into a new DNA-bending factor. This protein (denoted by a yellow ring) is required to pair two switch regions or two V regions for CSR and GC, respectively, by DNA http://www.jimmunol.org/ bending. The stabilization of the DNA pairing probably requires additional proteins, such as UNG and Msh2/6. by guest on October 1, 2021

flanked by repetitive sequences and marked by H3K4me3 Evolutionary consideration of the DNA-deamination model (46, 47). Therefore, we believe that cis marking by non-B The DNA-deamination model assumes that AID directly structural formation and trans marking by H3K4me3 on binds DNA and deaminates cytosine. The conversion of C to chromatin serve as marks for recruiting Top1 to the cleavage U generates a mismatched U/G , followed by DNA target, probably with the help of other proteins (Fig. 4). cleavage introduced by the base excision repair or mismatch Remarkably, H3K4me3 is involved in marking cleavage repair pathway (48). There is much evidence for the DNA- targets in meiotic recombination, as well as V(D)J recom- deamination model, the most straightforward of which is the bination. in vitro deamination of DNA by recombinant AID (49–51).

FIGURE 4. Common features between acquired im- mune diversity and genetic instability and diversifica- tion. (A) Meiotic recombination, V(D)J recombination, and CSR/SHM all depend on two marks on the targets: H3K4me3 on chromatin (trans) and unique DNA structure (cis). The cleaving enzymes are different. However, both meiotic recombination and CSR use members of the topoisomerase family. The non-B structure formation in CSR is enhanced by the AID- dependent decrease in Top1. (B) Genome instability mechanisms, such as those in TAM and triplet diseases, depend on the strong activation of transcription, which induces a relative Top1 deficiency. Non-B DNA struc- tures might also be enhanced by the relative Top1 defi- ciency. Top1 is also a DNA-cleaving enzyme that prob- ably recognizes non-B structures. Although it has not been assessed whether this target is also marked by H3K4me3, genome instability shares similar mecha- nisms with AID-induced acquired immune diversity. Rb, recombination; Set1/MLLs, histone methyltransferases. 3564 BRIEF REVIEWS: SIMILARITIES BETWEEN IMMUNE DIVERSITY AND GENOME INSTABILITY

In this system, ssDNA is a much better substrate than dsDNA ends on the same chromosome (Fig. 3B). This protein, which for AID (52). In addition, nucleosomal DNA is a less efficient we call “bending factor,” may act with UNG and Msh2/6, target for DNA deamination (53). However, the overex- which are also required for GC (54, 55). Because both GC pression of AID in E. coli or yeast introduces preferential C- and CSR require recombination between cis loci, UNG and to-T or G-to-A mutations. Another piece of evidence is Msh2/6 might be involved in pairing the cis loci required for the requirement for uracil-DNA glycosylase (UNG), which both GC and CSR. Because the evolution of AID originated removes U from the U/G mismatch, for CSR and GC, al- from a role in GC, AID must have had two activities (DNA though UNG inhibits SHM (54, 55). All of these mechanisms cleavage and cis pairing) from the beginning. If so, it is not are unique and are not known to be involved in any types of surprising that fish AID can exhibit CSR activity in mouse genetic alterations. On the contrary, both base excision repair cells. To test this hypothesis, it is important to re-examine and mismatch repair enzymes are required for the prevention whether the C-terminal domain of AID is also required for of genetic alterations rather than genome instability. GC. Although a C-terminal truncation mutant of AID was Although broadly accepted, the DNA-deamination model reported to show the same level of GC and SHM as the wild- has shortcomings that have been extensively discussed in other type AID (35), the data appear to contradict the observation reviews (7, 56). Therefore, we will list only a few points here. of Barreto et al. (35) that this AID mutant reverted a stop First, the in vitro and in E. coli DNA-deamination activities codon mutation in the Vl-chain gene of DT40 cells almost are catalyzed by a bona fide RNA-editing enzyme APOBEC 10 times more frequently than did the wild-type AID. 1, suggesting that these artificial assays do not provide evi- dence for in vivo DNA deamination by AID (57, 58). In fact, Conclusions Downloaded from APOBEC 1 requires a cofactor ACF for its RNA-editing Because nothing in biology makes sense except in the light activity in vivo, as well as in vitro (59). To explain the AID of evolution (69), we compared the mechanisms for various target specificity, the DNA-deamination hypothesis postulates types of genome diversification. Surprisingly, recent research the presence of guiding factors that physically interact with on AID function revealed that immune and genome diversi- AID. A large number of AID-interacting proteins has been fication mechanisms share many features with respect to http://www.jimmunol.org/ found: PKA, PTPB2, RNA exosome complex, RPA, Trim28, cleaving enzymes and target markers. Top1 is used in AID- MDM2, Pol II, Spt6, Spt5, DNA-PK, and 14-3-3 (43). induced DNA alterations (CSR and SHM), as well as in However, it is difficult to explain how such a large number of TAM deletion and triplet contraction/expansion (Fig. 4). proteins can specify DNA targets, because most of them are Meiotic recombination uses a type of topoisomerase, Spo11, rather ubiquitous proteins. for DNA cleavage. The cis markings for CSR, SHM, TAM, and triplet contraction/expansion all appear to depend on Evolutionary consideration of the CSR-specific role of the C-terminal a non-B DNA structure. H3K4me3 on chromatin of the region of AID cleavage target is used as a trans mark for CSR, SHM, meiotic The C-terminal region of AID is highly conserved among recombination, and V(D)J recombination. It is worth exam- by guest on October 1, 2021 vertebrates, suggesting that it has an important function (31, ining whether there is some posttranslational modification of 60). This region of AID has two activities: the nuclear export histone in targets of TAM and triplet contraction/expansion, signal and the CSR-specific activity. This is because C- because transcription is mandatory for both events. We pro- terminal mutants retain the in vitro DNA-deamination ac- pose that AID evolution in the primordial vertebrates took tivity and in vivo SHM activities in both the V and S regions advantage of the pre-existing genome instability mechanism but lose CSR activity (35, 38, 39). Thus, it is likely that the for DNA cleavage and recombination for accomplishing re- CSR-specific function of the C-terminal domain of AID is ceptor diversity. Nonetheless, the risk of introducing genome independent from the DNA-deamination activity. instability into non-Ig loci is tolerable compared with the Several hypotheses have been proposed for the function advantage of the protection against pathogens by the enor- of AID’s C-terminal domain, including stabilization of the mous immune diversification by AID. AID protein or protection from cell death induced by double- strand breakage (61–63). However, these hypotheses cannot Acknowledgments explain that a C-terminal mutation causes a loss of CSR We thank Drs. M. Yamamoto, S. Takeda, H. Nagaoka, K. Kinoshita, M. Mur- but less of a defect in SHM in hyper-IgM syndrome type amatsu, and S. Fagarasan for critical reading and comments on the manu- 2 patients (39). In addition, C-terminally truncated AID script, as well as Y. Shiraki and K. Fukui for manuscript preparation. mutants fused with the estrogen receptor hormone-binding domain, which stabilizes the protein, show a similar loss of Disclosures CSR with intact or hyper-SHM, making the stability hy- The authors have no financial conflicts of interest. pothesis less likely (37). The C-terminal domain of AID is responsible for bind- References ing poly(A) RNA through a putative cofactor protein (64). 1. Aguilera, A. 2002. The connection between transcription and genomic instability. Consistent with this idea, CSR depends on de novo protein EMBO J. 21: 195–201. synthesis (65). It was proposed that synapsis formation may 2. La Spada, A. R., and J. P. Taylor. 2010. Repeat expansion disease: progress and puzzles in disease pathogenesis. Nat. Rev. Genet. 11: 247–258. be required for appropriate pairing of cleaved ends of two 3. Hubert, L., Jr., Y. Lin, V. Dion, and J. H. Wilson. 2011. Topoisomerase 1 and separate S regions (66–68). These observations led us to single-strand break repair modulate transcription-induced CAG repeat contraction speculate that AID edits an mRNA captured by cofactor C, as in human cells. Mol. Cell. Biol. 31: 3105–3112. 4. Lippert, M. J., N. Kim, J. E. Cho, R. P. Larson, N. E. Schoenly, S. H. O’Shea, and well as that translation of the edited mRNA generates a novel S. Jinks-Robertson. 2011. Role for topoisomerase 1 in transcription-associated protein required for the synapsis formation between cleaved mutagenesis in yeast. Proc. Natl. Acad. Sci. USA 108: 698–703. The Journal of Immunology 3565

5. Takahashi, T., G. Burguiere-Slezak, P. A. Van der Kemp, and S. Boiteux. 2011. 35. Barreto, V., B. Reina-San-Martin, A. R. Ramiro, K. M. McBride, and Topoisomerase 1 provokes the formation of short deletions in repeated sequences M. C. Nussenzweig. 2003. C-terminal deletion of AID uncouples class switch re- upon high transcription in Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA 108: combination from somatic hypermutation and gene conversion. Mol. Cell 12: 501– 692–697. 508. 6. Gerton, J. L., and R. S. Hawley. 2005. Homologous chromosome interactions in 36. Harris, R. S., J. E. Sale, S. K. Petersen-Mahrt, and M. S. Neuberger. 2002. AID is meiosis: diversity amidst conservation. Nat. Rev. Genet. 6: 477–487. essential for immunoglobulin V gene conversion in a cultured B cell line. Curr. Biol. 7. Muramatsu, M., H. Nagaoka, R. Shinkura, N. A. Begum, and T. Honjo. 2007. 12: 435–438. Discovery of activation-induced cytidine deaminase, the engraver of antibody 37. Doi, T., L. Kato, S. Ito, R. Shinkura, M. Wei, H. Nagaoka, J. Wang, and T. Honjo. memory. Adv. Immunol. 94: 1–36. 2009. The C-terminal region of activation-induced cytidine deaminase is respon- 8. Schatz, D. G., and P. C. Swanson. 2011. V(D)J recombination: mechanisms of sible for a recombination function other than DNA cleavage in class switch re- initiation. Annu. Rev. Genet. 45: 167–202. combination. Proc. Natl. Acad. Sci. USA 106: 2758–2763. 9. Boehm, T., N. McCurley, Y. Sutoh, M. Schorpp, M. Kasahara, and M. D. Cooper. 38. Ta, V. T., H. Nagaoka, N. Catalan, A. Durandy, A. Fischer, K. Imai, S. Nonoyama, 2012. VLR-Based Adaptive Immunity. Ann. Rev. Immunol. In press. J. Tashiro, M. Ikegawa, S. Ito, et al. 2003. AID mutant analyses indicate require- 10. Thompson, C. B. 1995. New insights into V(D)J recombination and its role in the ment for class-switch-specific cofactors. Nat. Immunol. 4: 843–848. evolution of the immune system. Immunity 3: 531–539. 39. Durandy, A., S. Peron, N. Taubenheim, and A. Fischer. 2006. Activation-induced 11. Kobayashi, M., Z. Sabouri, S. Sabouri, Y. Kitawaki, Y. Pommier, T. Abe, cytidine deaminase: structure-function relationship as based on the study of H. Kiyonari, and T. Honjo. 2011. Decrease in topoisomerase I is responsible for mutants. Hum. Mutat. 27: 1185–1191. activation-induced cytidine deaminase (AID)-dependent somatic hypermutation. 40. Xu, S., K. Guo, Q. Zeng, J. Huo, and K. P. Lam. 2012. The RNase III enzyme Proc. Natl. Acad. Sci. USA 108: 19305–19310. Dicer is essential for germinal center B-cell formation. Blood 119: 767–776. 12. Kobayashi, M., M. Aida, H. Nagaoka, N. A. Begum, Y. Kitawaki, M. Nakata, 41. Tashiro, J., K. Kinoshita, and T. Honjo. 2001. Palindromic but not G-rich A. Stanlie, T. Doi, L. Kato, I. M. Okazaki, et al. 2009. AID-induced decrease in sequences are targets of class switch recombination. Int. Immunol. 13: 495–505. topoisomerase 1 induces DNA structural alteration and DNA cleavage for class 42. Rogozin, I. B., V. V. Solovyov, and N. A. Kolchanov. 1991. Somatic hypermuta- switch recombination. Proc. Natl. Acad. Sci. USA 106: 22375–22380. genesis in immunoglobulin genes. I. Correlation between somatic mutations and 13. Stanlie, A., M. Aida, M. Muramatsu, T. Honjo, and N. A. Begum. 2010. Histone3 repeats. Somatic mutation properties and clonal selection. Biochim. Biophys. Acta lysine4 trimethylation regulated by the facilitates chromatin transcription complex is 1089: 175–182.

critical for DNA cleavage in class switch recombination. Proc. Natl. Acad. Sci. USA 43. Storck, S., S. Aoufouchi, J. C. Weill, and C. A. Reynaud. 2011. AID and partners: Downloaded from 107: 22190–22195. for better and (not) for worse. Curr. Opin. Immunol. 23: 337–344. 14. Lichten, M., and B. de Massy. 2011. The impressionistic landscape of meiotic re- 44. Kato, L., N. A. Begum, A. M. Burroughs, T. Doi, J. Kawai, C. O. Daub, combination. Cell 147: 267–270. T. Kawaguchi, F. Matsuda, Y. Hayashizaki, and T. Honjo. 2012. Non- 15. Phadnis, N., R. W. Hyppa, and G. R. Smith. 2011. New and old ways to control immunoglobulin target loci of activation-induced cytidine deaminase (AID) share meiotic recombination. Trends Genet. 27: 411–421. unique features with immunoglobulin genes. Proc. Natl. Acad. Sci. USA 109: 2479– 16. Pommier, Y. 2006. Topoisomerase I inhibitors: camptothecins and beyond. Nat. 2484. Rev. Cancer 6: 789–802. 45. Daniel, J. A., M. A. Santos, Z. Wang, C. Zang, K. R. Schwab, M. Jankovic, 17. Zhao, J., A. Bacolla, G. Wang, and K. M. Vasquez. 2010. Non-B DNA structure- D. Filsuf, H. T. Chen, A. Gazumyan, A. Yamane, et al. 2010. PTIP promotes induced genetic instability and evolution. Cell. Mol. Life Sci. 67: 43–62. chromatin changes critical for immunoglobulin class switch recombination. Science http://www.jimmunol.org/ 18. Chuzhanova, N., J. M. Chen, A. Bacolla, G. P. Patrinos, C. Fe´rec, R. D. Wells, and 329: 917–923. D. N. Cooper. 2009. Gene conversion causing human inherited disease: evidence 46. Michelotti, G. A., E. F. Michelotti, A. Pullner, R. C. Duncan, D. Eick, and for involvement of non-B-DNA-forming sequences and recombination-promoting D. Levens. 1996. Multiple single-stranded cis elements are associated with activated motifs in DNA breakage and repair. Hum. Mutat. 30: 1189–1198. chromatin of the human c-myc gene in vivo. Mol. Cell. Biol. 16: 2656–2669. 19. Sze´kvo¨lgyi, L., and A. Nicolas. 2010. From meiosis to postmeiotic events: homol- 47. Yamane, A., W. Resch, N. Kuo, S. Kuchen, Z. Li, H. W. Sun, D. F. Robbiani, ogous recombination is obligatory but flexible. FEBS J. 277: 571–589. K. McBride, M. C. Nussenzweig, and R. Casellas. 2011. Deep-sequencing identi- 20. Baudat, F., J. Buard, C. Grey, A. Fledel-Alon, C. Ober, M. Przeworski, G. Coop, fication of the genomic targets of the cytidine deaminase AID and its cofactor RPA and B. de Massy. 2010. PRDM9 is a major determinant of meiotic recombination in B lymphocytes. Nat. Immunol. 12: 62–69. hotspots in humans and mice. Science 327: 836–840. 48. Stavnezer, J. 2011. Complex regulation and function of activation-induced cytidine 21. Myers, S., R. Bowden, A. Tumian, R. E. Bontrop, C. Freeman, T. S. MacFie, deaminase. Trends Immunol. 32: 194–201. G. McVean, and P. Donnelly. 2010. Drive against hotspot motifs in primates 49. Petersen-Mahrt, S. K., R. S. Harris, and M. S. Neuberger. 2002. AID mutates E. implicates the PRDM9 gene in meiotic recombination. Science 327: 876–879. coli suggesting a DNA deamination mechanism for antibody diversification. Nature by guest on October 1, 2021 22. Parvanov, E. D., P. M. Petkov, and K. Paigen. 2010. Prdm9 controls activation of 418: 99–103. mammalian recombination hotspots. Science 327: 835. 50. Bransteitter, R., P. Pham, M. D. Scharff, and M. F. Goodman. 2003. Activation- 23. Keeney, S. 2008. Spo11 and the Formation of DNA Double-Strand Breaks in induced cytidine deaminase deaminates deoxycytidine on single-stranded DNA but Meiosis. Genome Dyn. Stab. 2: 81–123. requires the action of RNase. Proc. Natl. Acad. Sci. USA 100: 4102–4107. 24. Hartsuiker, E., K. Mizuno, M. Molnar, J. Kohli, K. Ohta, and A. M. Carr. 2009. 51. Chaudhuri, J., M. Tian, C. Khuong, K. Chua, E. Pinaud, and F. W. Alt. 2003. Ctp1CtIP and Rad32Mre11 nuclease activity are required for Rec12Spo11 removal, Transcription-targeted DNA deamination by the AID antibody diversification en- but Rec12Spo11 removal is dispensable for other MRN-dependent meiotic func- zyme. Nature 422: 726–730. tions. Mol. Cell. Biol. 29: 1671–1681. 52. Shen, H. M., S. Ratnam, and U. Storb. 2005. Targeting of the activation-induced 25. Pancer, Z., C. T. Amemiya, G. R. Ehrhardt, J. Ceitlin, G. L. Gartland, and cytosine deaminase is strongly influenced by the sequence and structure of the M. D. Cooper. 2004. Somatic diversification of variable lymphocyte receptors in the targeted DNA. Mol. Cell. Biol. 25: 10815–10821. agnathan sea lamprey. Nature 430: 174–180. 53. Shen, H. M., M. G. Poirier, M. J. Allen, J. North, R. Lal, J. Widom, and U. Storb. 26. Rogozin, I. B., L. M. Iyer, L. Liang, G. V. Glazko, V. G. Liston, Y. I. Pavlov, 2009. The activation-induced cytidine deaminase (AID) efficiently targets DNA in L. Aravind, and Z. Pancer. 2007. Evolution and diversification of lamprey antigen nucleosomes but only during transcription. J. Exp. Med. 206: 1057–1071. receptors: evidence for involvement of an AID-APOBEC family cytosine deaminase. 54. Rada, C., J. M. Di Noia, and M. S. Neuberger. 2004. Mismatch recognition and Nat. Immunol. 8: 647–656. uracil excision provide complementary paths to both Ig switching and the A/T- 27. Bajoghli, B., P. Guo, N. Aghaallaei, M. Hirano, C. Strohmeier, N. McCurley, focused phase of somatic mutation. Mol. Cell 16: 163–171. D. E. Bockman, M. Schorpp, M. D. Cooper, and T. Boehm. 2011. A thymus 55. Saribasak, H., N. N. Saribasak, F. M. Ipek, J. W. Ellwart, H. Arakawa, and candidate in lampreys. Nature 470: 90–94. J. M. Buerstedde. 2006. Uracil DNA glycosylase disruption blocks Ig gene con- 28. Alder, M. N., I. B. Rogozin, L. M. Iyer, G. V. Glazko, M. D. Cooper, and version and induces transition mutations. J. Immunol. 176: 365–371. Z. Pancer. 2005. Diversity and function of adaptive immune receptors in a jawless 56. Honjo, T., M. Kobayashi, N. Begum, A. Kotani, S. Sabouri, and H. Nagaoka. vertebrate. Science 310: 1970–1973. 2012. The AID dilemma: infection, or cancer? Adv. Cancer Res. 113: In press. 29. Tasumi, S., C. A. Velikovsky, G. Xu, S. A. Gai, K. D. Wittrup, M. F. Flajnik, 57. Shivarov, V., R. Shinkura, and T. Honjo. 2008. Dissociation of in vitro DNA R. A. Mariuzza, and Z. Pancer. 2009. High-affinity lamprey VLRA and VLRB deamination activity and physiological functions of AID mutants. Proc. Natl. Acad. monoclonal antibodies. Proc. Natl. Acad. Sci. USA 106: 12891–12896. Sci. USA 105: 15866–15871. 30. Barreto, V. M., Q. Pan-Hammarstrom, Y. Zhao, L. Hammarstrom, Z. Misulovin, 58. Petersen-Mahrt, S. K., and M. S. Neuberger. 2003. In vitro deamination of cytosine and M. C. Nussenzweig. 2005. AID from bony fish catalyzes class switch recom- to uracil in single-stranded DNA by apolipoprotein B editing complex catalytic bination. J. Exp. Med. 202: 733–738. subunit 1 (APOBEC1). J. Biol. Chem. 278: 19583–19586. 31. Wakae, K., B. G. Magor, H. Saunders, H. Nagaoka, A. Kawamura, K. Kinoshita, 59. Blanc, V., and N. O. Davidson. 2003. C-to-U RNA editing: mechanisms leading to T. Honjo, and M. Muramatsu. 2006. Evolution of class switch recombination func- genetic diversity. J. Biol. Chem. 278: 1395–1398. tion in fish activation-induced cytidine deaminase, AID. Int. Immunol. 18: 41–47. 60. Barreto, V. M., and B. G. Magor. 2011. Activation-induced cytidine deaminase 32. Ichikawa, H. T., M. P. Sowden, A. T. Torelli, J. Bachl, P. Huang, G. S. Dance, structure and functions: a species comparative view. Dev. Comp. Immunol. 35: 991– S. H. Marr, J. Robert, J. E. Wedekind, H. C. Smith, and A. Bottaro. 2006. 1007. Structural phylogenetic analysis of activation-induced deaminase function. J. 61. Chaudhuri, J., C. Khuong, and F. W. Alt. 2004. Replication protein A interacts Immunol. 177: 355–361. with AID to promote deamination of somatic hypermutation targets. Nature 430: 33. Thompson, C. B., and P. E. Neiman. 1987. Somatic diversification of the chicken 992–998. immunoglobulin light chain gene is limited to the rearranged variable gene segment. 62. Wu, X., P. Geraldes, J. L. Platt, and M. Cascalho. 2005. The double-edged sword of Cell 48: 369–378. activation-induced cytidine deaminase. J. Immunol. 174: 934–941. 34. Reynaud, C. A., V. Anquez, H. Grimal, and J. C. Weill. 1987. A hyperconversion 63. Ha¨sler, J., C. Rada, and M. S. Neuberger. 2011. Cytoplasmic activation-induced mechanism generates the chicken light chain preimmune repertoire. Cell 48: 379– cytidine deaminase (AID) exists in stoichiometric complex with translation elon- 388. gation factor 1a (eEF1A). Proc. Natl. Acad. Sci. USA 108: 18366–18371. 3566 BRIEF REVIEWS: SIMILARITIES BETWEEN IMMUNE DIVERSITY AND GENOME INSTABILITY

64. Nonaka, T., T. Doi, T. Toyoshima, M. Muramatsu, T. Honjo, and K. Kinoshita. 67. Larson, E. D., M. L. Duquette, W. J. Cummings, R. J. Streiff, and N. Maizels. 2009. Carboxy-terminal domain of AID required for its mRNA complex formation 2005. MutSalpha binds to and promotes synapsis of transcriptionally activated in vivo. Proc. Natl. Acad. Sci. USA 106: 2747–2751. immunoglobulin switch regions. Curr. Biol. 15: 470–474. 65. Doi, T., K. Kinoshita, M. Ikegawa, M. Muramatsu, and T. Honjo. 2003. De 68. Wuerffel, R., L. Wang, F. Grigera, J. Manis, E. Selsing, T. Perlot, F. W. Alt, novo protein synthesis is required for the activation-induced cytidine deaminase M. Cogne, E. Pinaud, and A. L. Kenter. 2007. S-S synapsis during class switch function in class-switch recombination. Proc. Natl. Acad. Sci. USA 100: 2634– recombination is promoted by distantly located transcriptional elements and 2638. activation-induced deaminase. Immunity 27: 711–722. 66. Bassing, C. H., and F. W. Alt. 2004. H2AX may function as an anchor to hold 69. Dobzhansky, T. 1973. Nothing in Biology Makes Sense Except in the Light of broken chromosomal DNA ends in close proximity. Cell Cycle 3: 149–153. Evolution. Am. Biol. Teach. 35: 125–129. Downloaded from http://www.jimmunol.org/ by guest on October 1, 2021