V(D)J Recombination: Mechanisms of Initiation 720/2018 Group 3 Paper.pdfOF V(D)J RECOMBINATION Based...

GE45CH09-Schatz ARI 5 October 2011 15:53

V(D)J Recombination:Mechanisms of InitiationDavid G. Schatz1,∗ and Patrick C. Swanson2

1Department of Immunobiology and Howard Hughes Medical Institute, Yale UniversitySchool of Medicine, New Haven, Connecticut 06520-8011; email: [email protected] of Medical Microbiology and Immunology, Creighton University MedicalCenter, Omaha, Nebraska 68178; email: [email protected]

Annu. Rev. Genet. 2011. 45:167–202

First published online as a Review in Advance onAugust 19, 2011

The Annual Review of Genetics is online atgenet.annualreviews.org

This article’s doi:10.1146/annurev-genet-110410-132552

Copyright c© 2011 by Annual Reviews.All rights reserved

0066-4197/11/1201-0167$20.00

∗Corresponding author.

Keywords

site-specific recombination, recombination signal sequence, RAG1,RAG2, HMGB1, protein-DNA complex

Abstract

V(D)J recombination assembles immunoglobulin and T cell receptorgenes during lymphocyte development through a series of carefully or-chestrated DNA breakage and rejoining events. DNA cleavage requiresa series of protein-DNA complexes containing the RAG1 and RAG2proteins and recombination signals that flank the recombining genesegments. In this review, we discuss recent advances in our understand-ing of the function and domain organization of the RAG proteins, thecomposition and structure of RAG-DNA complexes, and the pathwaysthat lead to the formation of these complexes. We also consider thefunctional significance of RAG-mediated histone recognition and ubiq-uitin ligase activities, and the role played by RAG in ensuring properrepair of DNA breaks made during V(D)J recombination. Finally, wepropose a model for the formation of RAG-DNA complexes that in-volves anchoring of RAG1 at the recombination signal nonamer andRAG2-dependent surveillance of adjoining DNA for suitable spacerand heptamer sequences.

167

Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


INTRODUCTIONThe discovery that immunoglobulin genesundergo somatic DNA rearrangement was awatershed moment in the field of immunology,stimulating research that continues to this day

c

b

a

AP-1 site

12-RSS

23-RSS

Dβ2

3’Dβ1 23-RSS

Dβ1Vβ (30) Jβ1 Jβ2Cβ1 Cβ2

B12/23

III

Methylation interference

Ethylation interference

Permanganate interference

Vβ14

Consensus 12-RSS

C1b

Top strand

Bottom strand

Coding Heptamer Spacer Nonamer

CACGGTGattcaattctatgggaagcctttACAAAAACC

GTGCCACtaagttaagatacccttcggaaaTGTTTTTGG

NNNNNNNNNNNNNNACAAAAACC

NNNNNNNNNNNNNNTGTTTTTGG

NNNCACAGTG

NNNGTGTCAC

1 12

AP-1 site

1 2 3 4 5 6 7 1 2 3 4 5 6 7 8 9

Figure 1RSS and Tcrb locus organization. (a) The sequence, location, and numbering scheme of the heptamer andnonamer in a consensus 12RSS are shown overlaid on a triangle that represents the RSS orientation insubsequent figures. Most endogenous RSSs differ in sequence at one or more positions from the consensus,although positions 1–3 of the heptamer are almost perfectly conserved. The 23RSS contains 23 bp betweenthe heptamer and nonamer. Spacer sequence is less well conserved than the heptamer or nonamer. Alsoindicated is the first nucleotide of the coding flank on the bottom strand (C1b), which is thought to besubject to base flipping prior to hairpin formation. (b) Organization of the Tcrb locus. The 12 and 23RSSs areidentified by purple and yellow triangles, respectively, V, D, and J gene segments as green, blue, and grayrectangles, respectively, and constant regions as red rectangles. An AP1 site embedded in the 3′-Dβ 23RSSsis indicated by a small red circle. The RSS composition and orientation for a representative gene segmentwithin a given cluster are shown (e.g., all Jβ1 segments have a 5′ 12RSS); remaining RSSs are omitted forsimplicity. Dβ-to-Jβ recombination (I) precedes Vβ-to-Dβ recombination (II). Despite being formallypermissible by the 12/23 rule, direct Vβ-to-Jβ recombination is restricted by the “beyond 12/23 rule”(B12/23) (c) Sequence of the 3′-Dβ1 23RSS. The location of an embedded AP1 site is indicated by anoverline; RAG-RSS heptamer-spacer contacts revealed by interference footprinting (130), which areconserved in the 3′-Dβ1 23-RSS and overlap with the putative AP1 binding site, are identified above orbelow the nucleotide sequence.

to identify and characterize the mechanismsthat mediate and regulate this process. Asthe organization of the antigen receptor lociwas revealed (see example in Figure 1b),it was recognized that functional V(D)J

168 Schatz · Swanson

Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


Recombinationsignal sequence(RSS): a short DNAsequence containing aconserved heptamerand nonamerseparated by either 12or 23 base pair

Heptamer: relativelywell-conserved 7-bpportion of the RSSimportant for directingthe site of DNAcleavage by RAG

RAG: recombinationactivating gene

Nicking: first step ofDNA cleavage byRAG in which oneDNA strand is broken5′ of the heptamer

Paired (synaptic)complex (PC):protein-DNA complexin which the two RSSsare held in closejuxtaposition by theRAG proteins

Signal end: afterDNA cleavage by theRAG proteins, theDNA end thatterminates in the RSS

Coding end: afterDNA cleavage by theRAG proteins, theDNA end thatterminates in thecoding segment

CSC: cleaved signalcomplex

SEC: signal endcomplex

Nonhomologous endjoining (NHEJ):a DNA repair processthat joins broken DNAends (double-strandbreaks) without usinghomologous DNA as atemplate

rearrangements only occur between two genesegments with flanking recombination signalsequences (RSSs) that differ in the spacer length(either 12 or 23 bp) between the heptamerand nonamer elements (Figure 1a). This ledto speculation that V(D)J recombination mustinvolve the formation of one or more protein-DNA complexes that bridge the two RSSsto initiate the rearrangement process. Theidentification of the recombination activatinggenes, RAG1 and RAG2, and the eventual con-firmation of their direct role in initiating V(D)Jrecombination by introducing site-specificDNA double-strand breaks (DSBs) spurredmuch additional work to dissect the molecularmechanism of the reaction. This reviewdiscusses recent advances in our mechanisticunderstanding of how V(D)J recombinationis initiated and how the RAG (recombinationactivating gene) proteins direct the repairof the DNA breaks they create. We do notprovide detailed consideration of DNA repairmechanisms operative in V(D)J recombination,the evolutionary origins of the RAG proteins,or the mechanisms regulating the accessibilityand positioning of antigen receptor loci duringV(D)J recombination, but instead refer readersto other reviews on these topics (12, 46, 65, 86,114). Likewise, to enable us to focus on morerecent developments within each topic area, wewill briefly overview earlier work in each of themajor sections, referring readers to previousreviews for more in-depth coverage (31, 47, 51,131).

OVERVIEW OF THEBIOCHEMICAL STEPSOF V(D)J RECOMBINATION

Based on extensive biochemical studies ofthe RAG proteins, characterization of V(D)Jrecombination intermediates, and identifica-tion of factors involved in the repair of theseintermediates, one can conceptually divideV(D)J recombination into two distinct phases:a cleavage phase and a joining phase (47,51) (Figure 2). During the cleavage phase, a

12RSS and a 23RSS are brought into closeproximity by the RAG proteins to form a stablemulti-subunit synaptic complex. Accumulatingevidence suggests synaptic complex formationoccurs through a stepwise capture modelof assembly (28, 71, 98) that involves initialRAG binding and perhaps nicking of an RSSwithin a single RSS complex (SC), followed bysubsequent capture of an appropriate partnerRSS to form a synaptic or paired complex (PC)(Figure 2). Next, the RAG proteins nick thepartner RSS and catalyze the coupled cleavageof both RSSs by direct transesterification usingthe 3′-OH exposed by nicking. This reactionyields a pair of blunt, 5′-phosphorylated signalends and a pair of coding ends that terminate incovalently sealed DNA hairpin structures (92).After cleavage, the RAG proteins transientlyretain the DNA ends in a cleaved signalcomplex (CSC) (61). Coding ends are thoughtto dissociate first, while the signal ends remaintightly bound to the RAG proteins in a signalend complex (SEC). Within these postcleavagecomplexes, the broken DNA transitions tothe joining phase, during which the DNAends are reorganized, processed, and repaired.Genetic and biochemical evidence suggeststhe catalytic subunit of DNA-dependentprotein kinase (DNA-PKcs) and Artemisform a structure-specific endonuclease thatresolves the DNA hairpin structures (89).Signal and coding ends are then processed andjoined through the classical nonhomologousend joining (NHEJ) DNA repair pathway,involving at least five proteins, including Ku70,Ku80, XRCC4, DNA Ligase IV, and theCernunnos/XLF protein (86). Members ofthe Pol X family of polymerases (TdT, polμ,and polλ) contribute to the diversificationand repair of coding joints through template-independent nucleotide addition and fill-inDNA synthesis (145). This process generallyproduces two types of joined products: a codingjoint formed from joining the two gene codingsegments together (often with small junctionalnucleotide additions or deletions); and a signaljoint formed by signal end ligation (oftenprecisely) (Figure 2).

www.annualreviews.org • V(D)J Recombination 169

Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


RAG1 AND RAG2: DOMAINORGANIZATION ANDPURIFICATION

Protein Structure Overview

The RAG proteins are highly conserved amongjawed vertebrate organisms. Because much of

12-SC

23-SCsignal complex

PCpaired complex

CSCcleaved signal complex

SECsignal end complex

Bind and nick single RSS

Capture partner RSS

Nick partner RSSCleave both RSSs

(hairpin formation)

Release coding ends

Hairpin openingCoding end processing

End-joining

Artemis/DNA-PKcsKu70/Ku80

XRCC4/DNA Ligase IV/XLF/ATMPolX family (TdT, pol μ, pol λ)

RAG1

HMGB1 (?)RAG2

12-RSS 23-RSS

Clea

vage

Join

ing

Codingjoint

Signaljoint

HO

OH

Or

HO

OH

NHEJfactors

Or

+

our understanding of RAG protein biochem-istry has been gained using the murine RAGproteins, this review uses murine RAG1 andRAG2 amino acid numbering, unless otherwisenoted. Murine RAG1 and RAG2 contain 1,040and 527 amino acids, respectively (Figure 3).Substantial portions of both proteins canbe removed without severely compromisingV(D)J recombination activity on episomal orintegrated substrates in cell culture (131 andreferences therein). The smallest functionaltruncation mutants, referred to as core RAG1(384–1008) and RAG2 (1–387) (Figure 3),are more easily purified than their full-lengthcounterparts, and hence have been (and con-tinue to be) used in most studies of RAG proteinbiochemistry. However, the noncore regionsof RAG1 and RAG2 clearly play critical roles inpromoting physiological V(D)J recombination,because mice expressing only core forms ofRAG1 or RAG2 exhibit defects in the efficiency,fidelity, and ordering of V(D)J rearrangementat endogenous loci (5, 29, 38, 85, 136). Success-ful purification of full-length RAG proteins hasbeen reported more recently, providing impor-tant insights into the function of the noncoreregions (43, 135, 140). Through extensivemutagenesis of both RAG proteins, a fairly

←−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−Figure 2Overview of the cleavage and joining phases ofV(D)J recombination. Evidence suggests thecleavage phase starts with the stepwise binding of a12RSS or 23RSS by the RAG and HMGB1 proteinsto form a single recombination signal sequence(RSS) complex (SC), followed by the capture of anappropriate partner RSS to form a synaptic or pairedcomplex (PC). The RAG proteins may or may notnick the bound RSS before synapsis. Upon synapsis,the RAG proteins complete nicking and hairpinformation steps at both RSSs, yielding blunt signalends and hairpin coding ends that are retained in atransient cleaved signal complex (CSC). The codingends then dissociate, but the signal ends remaintightly bound by the RAG proteins in a signal endcomplex (SEC). Coding and signal ends aresubsequently processed and repaired asynchronouslyby the classical nonhomologous end joining repairpathway to generate signal ends that are typicallyprecise and coding ends that are often imprecise.


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


PHD: planthomeodomain

detailed picture of RAG1 and RAG2 structure-function relationships has emerged (Figure 3;see sidebar, RAG Mutants and Disease).

Full-length RAG1 contains distinct regionswith different functional roles. The N-terminalnoncore portion of RAG1 contains elementsthat regulate cellular protein levels (55, 93,109), mediate interactions with several proteinfactors (25, 52, 90, 109), coordinate zinc ions,and enhance cellular V(D)J recombinationactivity (73, 93, 112). This region also includesa structurally characterized zinc-bindingdimerization domain (ZDD) spanning residues265–380 (110). The core portion of RAG1can be subdivided into three distinct regions(31): (a) A structurally characterized nonamerbinding domain (NBD) (residues 389–442) thatforms a dimer and interacts primarily with theRSS nonamer (36, 126); (b) a central domain(residues 528–760) that exhibits nicking activ-ity on ssDNA, which is enhanced by RAG2(8), mediates heptamer contact (preferringsingle-stranded and nicked substrates) (103),and interacts with RAG2 partly though aC2H2 zinc finger domain (ZnF-B) (4); and (c) aC-terminal domain (CTD) (residues 761–979)that binds double-stranded DNA nonspecif-ically and cooperatively (8), mediates contactwith the coding sequence flanking the RSS(97), and contains two zinc binding sites (57).In mixing experiments, the CTD is reportedto suppress central domain ssDNA nickingactivity, but not when substrates contain anssDNA-dsDNA transition (153). However,this finding could be alternatively explainedby competitive inhibition of DNA binding bythe CTD on ssDNA substrates, whereas thelack of CTD inhibition using ssDNA-dsDNAsubstrates may reflect the stronger dsDNAbinding preference of the CTD relative to thecentral domain. The far C terminus of RAG1(residues 1009–1040) is reported to collaboratewith the RAG2 C terminus to modulate RAGbinding and cleavage activity (discussed below)(55).

Within RAG2, three key regions havebeen defined: (a) an N-terminal region com-prising core RAG2 (1–387), containing six

RAG MUTANTS AND DISEASE

The RAG proteins have been subjected to extensive mutagenesisand RAG mutants have been implicated in multiple disease en-tities. Early mutagenesis studies defined the catalytic core of theRAG1 and RAG2 proteins (51), and recent efforts have reducedthe minimal region of RAG2 required to support levels of coreRAG2 activity to residues 1–360 (24) (K. Dhar and P.C. Swanson,unpublished data). In addition, studies have defined the aminoacid requirements for RAG enzymatic activity [including catalytic(49, 76, 82), step-arrest (75, 105), joining-defective (62, 113), andgain-of-function mutants (78, 102)], DNA binding activity (35,48, 126, 132), RAG1-RAG2 interactions (4, 77, 94), and RAG in-teractions with other factors (3, 91, 109). Importantly, naturallyoccurring human RAG mutations have been identified that causea complete or partial loss of physiological V(D)J recombinationactivity (see Related Resources and 103, 107) and are associatedwith a spectrum of severe immune deficiency disorders rangingfrom classical T-B-SCID to slightly milder variants, such as com-plete or incomplete Omenn syndrome, and RAG deficiency withγδ T cell expansion, granuloma formation, or maternofetal en-graftment (103, 107). Representative RAG mutations are pro-vided in Figure 3.

Kelch-like motifs (residues 1∼350) proposedto adopt a six-bladed beta-propeller-like struc-ture (21), the sixth of which reportedly medi-ates contacts with RAG1 (4); (b) a flexible acidichinge region (residues ∼360–408); and (c) aC-terminal noncanonical plant homeodomain(PHD) (residues 414–487) shown to bind phos-phoinositides (42) and histone H3 trimethy-lated at lysine 4 (hereafter termed H3K4me3)(87, 91). The far C terminus contains a con-served threonine residue (T490) that mediatesphosphorylation-dependent and cell-regulateddegradation of RAG2 via Skp2-SCF-mediatedubiquitination (68).

CORE AND FULL-LENGTH RAGPROTEINS: EXPRESSION,PURIFICATION, ANDOLIGOMERIZATION STATE

Native RAG proteins are generally recal-citrant to purification in yields suitable for


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


biochemical characterization without being ap-pended to a fusion partner to enhance solubility(131). Bacterial, insect, and mammalian cellexpression systems all yield RAG1 that is activein RSS binding and cleavage assays, whereasRAG2 is only recovered in highly activeform from mammalian cells (131). However,coexpressed RAG1 and RAG2 recovered frominsect or mammalian cells are active; the yieldand activity of coexpressed RAG1 and RAG2are reportedly higher than the individuallyexpressed proteins (129). Early proceduresfor purifying core RAG1 or RAG2 relied on

buffers containing high levels of NaCl (>0.5 M)(92). These conditions also permitted isolationof active full-length RAG1 coexpressed withcore RAG2 and vice-versa; however, RAGpreparations containing full-length RAG1 orRAG2 did not bind substrate DNA uniformly(135). Milder purification conditions weresubsequently identified that enable recoveryof full-length RAG proteins with improvedDNA binding characteristics and maintenanceof protein-protein interactions (109).

Core RAG1 alone forms a stable dimer insolution under most in vitro conditions tested

Core

1008 1040

RAG1265 380 384 458 528 760 761 979

ZF-B

D600D708

E962

E649HA3

RAG2

1 387 527

Core

360 408 414 487

PHDHinge

T490

Kelch motifs

hR394W (OS)hS401P (OS)hR449K (OS)

hI956T (OS)hR778W (T-B-)

hR561H (OS)hC602W (T-B-)hG709D (OS)

hE770K (A-S)

hN855I (S-MFT)

K980A#

K608A#H609L#

K596A

R401A/R402A

R440A

R855A/K856AK890AW893A#

R621A/H R713A

R734A R795A

R748A/H750AR773A/R775A

K405A/H406A/R407AW956A

W760A#

E547Q E423Q S723A/C#

E719K

CC HH

H937A/K938AH942AR969A/R970A

Pathologic (disorder)

Nicking(+), hairpin(-)

Joining (-/defective)

Active siteGain-of-function

RSS binding(+),nicking(-), hairpin(-)

RAG2(+), RSS binding(-),nicking(-), hairpin(-)

• Phosphoinositide binding• Histone H3K4me3 binding• Stabilizes RAG1/2 tetramer

Cell cycle–dependentphosphorylationand ubiquitin-dependentdegradation

Pathologic (disorder)

Joining (-/defective)

RSS binding(+),nicking(-), hairpin(-)

RAG1(+), RSS binding(-),nicking(-), hairpin(-)

a

b

hC325Y (OS)

• E3 Ub ligase• Dimerization

• Nonamer binding• Dimerization

• Heptamer binding• ssDNA binding• RAG2 interaction (esp a.a. 692-758)

• Coding end binding• Cooperative binding• Dimerization• Zn binding

• Zn binding• Protein stability• Enhanced V(D)J recombination• Protein interaction*

384

ZF-ARING

1 C1 C2C3 CHBI BIIa/bBIII

ZDD NBD Central C-term

• Enhances RSS binding• Required for RSS cleavage• RAG1 interaction (esp. a.a. 314-371)

Acidic

hG35V(T-B-)

hT215I(T-B-)hR229Q/W(OS)

hW416L(OS)hM443I(OS)hM453R(OS)

hN474S(S-MFT)hC478Y(T-B-)hH481P(T-B-)

K38A/R39A

K119R K56A/K58A K34A

R73A R167A(CJ)

H94A/K97A 361fs(A-NHEJ)

K119A K283A/R


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


EMSA:electrophoreticmobility shift assay

(10, 23, 133); higher-order oligomeric forms(tetramer and octamer) may be favored whenunder conditions of low ionic strength (0.2 MNaCl) and temperature (10◦C) (32). RAG2 isreportedly purified as a mixture of monomericand higher-order oligomeric forms, with themonomeric form predominating (10). Unlikecore RAG1, core RAG2 by itself exhibits lit-tle, if any DNA binding activity (134). CoreRAG1 and RAG2 associate with one anotherin solution without DNA, forming a 1:1 mixedtetramer complex (134). The tetramer config-uration is further stabilized when the RAG2 Cterminus is present (55). Based on this evidence,RAG1 and RAG2 most likely assemble a het-erotetramer complex in vivo, but whether phys-iological synapsis and cleavage of RSS pairs invivo are mediated by one or more (perhaps two)of these complexes remains unclear at present(discussed below).

RAG-RSS SINGLE SITE ANDSYNAPTIC COMPLEX ASSEMBLYAND ACTIVITY

In this section, features of RAG-RSS sin-gle site and synaptic complexes are discussed,deferring for later sections a discussion of whichRSS is targeted first by the RAG proteins andwhat mechanisms enforce the 12/23 rule and

“beyond 12/23 rule” (B12/23) of V(D)J re-combination. Much of what is known aboutRAG-RSS complex assembly and activity hasbeen gained from studies using core RAG pro-teins and short oligonucleotide RSS substrates.These studies are not discussed in depth here,as most of this material has been reviewedpreviously (47, 51, 131, 134).

SINGLE RAG-RSS COMPLEXES

RAG1-RSS Complexes

Core RAG1 alone binds an isolated RSS asa dimer (134) with an apparent Kd rangingfrom 29–41 nM, as measured by fluorescenceanisotropy (23, 153), to 92–114 nM, as mea-sured by electrophoretic mobility shift assay(EMSA) (23, 111, 153). RAG1 subunits donot spontaneously reassort during the processof DNA binding (133). Minor higher-orderoligomeric complexes of uncertain biologicalsignificance are also variably detected by EMSAbut have not been extensively analyzed. RAG1binding to the RSS induces conformationalchanges in RAG1 (23) and promotes DNAbending (3). DNA binding and footprintingstudies of dimeric core RAG1-RSS complexessuggest that RSS binding is primarily mediatedby the RAG1 NBD; the DNA interactions are

←−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−Figure 3Structure-function relationships of RAG1 and RAG2. (a) Regions of RAG1 and (b) RAG2 required ordispensable for supporting basic recombination activity are shaded blue or red, respectively. Key domainswithin these proteins and their location and putative function(s) are shown below the full-length RAG1 andRAG2. Conserved cysteine and histidine residues and basic motifs in RAG1 are identified by lines and bars,respectively, and are located as follows: C1(102–114); C2(168–180); C3(203–213); CH(266–272);CC(902–907); HH(937–942); BI(141–146); BIIa(218–224); BIIb(233–236); and BIII(243–254). Proteinsreported to interact with the RAG1 N terminus include KPNA1, histone H3, and Ku70/Ku80. Selectedhuman pathologic RAG mutations and murine RAG mutations that impair the DNA binding, nicking,hairpin formation, or joining steps of V(D)J recombination without affecting RAG1:RAG2 association[RAG1(+) or RAG2(+)] are summarized. For simplicity, pathologic mutations only include a subset ofhomozygous RAG mutations identified in human patients with T-B-SCID (T-B-), atypical SCID (A-SCID),Omenn syndrome (OS), or SCID with maternofetal engraftment (S-MFT). Murine RAG mutationsidentified by the # symbol exhibit defects that are conditional on features such as coding flank composition(K608A, K609L, and W893A) or DNA substrate length (K980A), or exhibit unusual or restrictedabnormalities, such as impaired SC2 but not SC1 formation (W760A), a joining defect on episomalsubstrates but a DNA cleavage defect on endogenous antigen receptor loci (S723A/C), a selective defect incoding joint formation (CJ), or enhanced usage of alternative NHEJ mechanisms (A-NHEJ). Abbreviations:C-term, C-terminal domain; NBD, nonamer binding domain; PHD, plant homeodomain; ZDD,zinc-binding dimerization domain.


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


Hairpin formation:second step of DNAcleavage by RAG inwhich the 3′-hydroxylof the nicked strandattacks the otherstrand

mostly centered on the nonamer and flankingspacer sequence and involve extensive contactswith the phosphodiester backbone (131). Thesefindings are further substantiated and extendedby recent structural analysis of the RAG1 NBD(residues 389–464) bound to a DNA duplexcontaining the 3′ end of the spacer (4 basepairs) and a consensus nonamer (146). In thiscomplex, the NBD forms a symmetrical ho-modimer that is bound to two DNA molecules,with each of the two subunits contacting bothDNA molecules. The AT-hook-like GGRPRmotif at the N terminus of the NBD binds inthe minor groove of the nonamer at positions5–7 (see Figure 1a), with R391 mediatingbase-specific contacts (T5 and T6; bottomstrand), and R393 contacting the phosphodi-ester backbone at T5. The polypeptide extendsover the phosphodiester backbone betweenT4 and T5, with the remaining polypeptidearranged into three helices. The first helix iskinked, with the N-terminal portion (residues400–407) occupying the major groove over ahighly conserved guanine at nonamer position2 (bottom strand). This residue is contacted byR402, which is consistent with the sensitivityof G2 to methylation interference (131). Thephosphate backbone is contacted by R407,which also plays a structural role in stabilizingthe loop that connects the GGRPR motifto helix H1. The first half of the C-terminalportion of helix 1 (residues 408–422) containsbasic residues that either directly contact(K405 and K412) or approach (R401 andR409) the phosphate backbone in the 3′ endof the spacer (S8–S12 on bottom strand; seeFigure 1a). Helix 2 (residues 426–441) andhelix 3 (residues 444–454) form a four-helixbundle with their counterparts on the otherNBD subunit to create an extensive dimerinterface (burying 1900 A2 of each monomer).Interestingly, helix 2 crosses over to approachthe DNA bound by the other subunit at its Cterminus, with residues Asn443 and His445positioned near the backbone phosphates atS8 and S9 (bottom strand). Taken together,the extensive phosphate contacts involvedin stabilizing RAG1-RSS complexes likely

explain why core RAG1 binding exhibits onlymoderate discrimination between specific andnonspecific DNA (∼tenfold) (23) and somesensitivity to ionic strength in DNA bindingassays (153).

RAG1-RAG2 Single RSS Complexes

Purified RAG1 mixed or coexpressed withRAG2 forms one or two discrete RAG-RSS complexes, generally called SC1 and SC2,which are detectable by EMSA and exhibit sim-ilar catalytic activity in vitro (134). In reactionscontaining Mg2+, this activity is limited mostlyto RSS nicking. However, in Mn2+, the RAGproteins robustly catalyze both cleavage steps(nicking and hairpin formation) in the absenceof synapsis. In the SC complexes, RAG1 likelyretains the dimeric configuration observed inRAG1-RSS complexes, although there is notcomplete agreement on this issue (134). Bycontrast, RAG2 stoichiometry differs betweenSC1 and SC2, with the former and lattercontaining one or two RAG2 subunits, respec-tively. The abundance and distribution of thetwo complexes depend in part on fusion partnercomposition (GST-RAG2 preferentially formsSC2, perhaps due to the intrinsic dimerizationpotential of GST) (32), RAG expressionand purification strategies (coexpression andpurification under mild conditions favors SC2)(109, 129), and the absence or presence of theRAG2 C terminus (SC2 is favored when theRAG2 C terminus is present) (109).

Compared to RAG1 alone, RAG1 andRAG2 together (hereafter called RAG) bind asingle RSS with greater affinity (apparent Kdof 25 nM by EMSA) (153) and specificity (60,96, 132). Enhanced binding can be attributedto additional base-specific and phosphate con-tacts that overlap with and extend from thoseobserved in RAG1-RSS complexes to includethe 5′-end of the spacer, the heptamer, andthe heptamer-coding junction, and are biasedtoward one side of the DNA helix (131).RAG-RSS complex formation also promotesDNA conformational changes in the spacerand heptamer regions that are evidenced by


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


HMGB: highmobility group box

hypersensitivity toward chemical and enzy-matic probes (131). Most of these contactsare probably mediated by the central and C-terminal domains of RAG1, which containthree residues critical for RAG cleavage ac-tivity, but are otherwise dispensable for DNAbinding (D600, D708, and E962, see Figure 3)(49, 76, 82). Evidence suggests these activesite residues, and by extension the protein-DNA contacts mediated by the central and C-terminal domains that contain them, are con-tributed in trans to those mediated by theGGRPR motif of the nonamer-bound NBD inthe RAG1 dimer (128). This model is substan-tiated by structural studies of a RAG1 NBD-DNA complex discussed above and places thelikely cis-trans crossover point in helix 2 ofthe NBD. Two lines of evidence suggest thatdirect DNA contact by RAG2 in RAG-RSScomplexes is a strong possibility, despite itslack of intrinsic DNA binding activity. First,photo-crosslinking studies show that RAG2 lo-calizes proximal to the site of DNA cleavage (41,133). Second, RAG2 mutations in basic residueshave been identified that specifically impairRAG-RSS complex formation without disrupt-ing RAG1-RAG2 interactions (see Figure 3)(48). Nevertheless, any RAG2 interactions withDNA are likely quite restricted, and on balanceRAG2 probably serves more as a cofactor inpromoting RAG1-RSS interactions than exten-sively contributing to RSS recognition.

HMGB1 and HMGB2

Although the RAG proteins alone are suffi-cient to mediate RSS binding and cleavage, thehigh mobility group box proteins HMGB1 andHMGB2 were found to stimulate RAG bind-ing and cleavage activity on DNA substratescontaining a 23RSS, but not a 12RSS, and topromote coupled cleavage of 12RSS and 23RSSpaired in cis (141). HMGB1 and HMGB2 arestructurally very similar and functionally redun-dant in RAG biochemical assays; for simplicity,we will only refer to HMGB1 unless otherwisespecified.

HMGB1 contains tandem HMGB domains,called A and B, followed by a basic linker

and an acidic tail (137). Each ∼80 amino acidHMGB domain contains three α-helixes thatform an L-shaped structure; box A favors bind-ing to distorted DNA structures, whereas boxB can bind and bend linear DNA. AlthoughHMGB1 exhibits both RAG-independent RSSbinding (130) and RSS-independent associationwith the RAG1 NBD (3), these interactionsare not independently robust, but are likelyenhanced by ternary complex formation (A.J.Little and D.G. Schatz, unpublished results).DNA footprinting and photo-crosslinkingstudies, together with more recent structure-function analyses of HMGB1 (see 131, 134, andreferences therein), suggest a model in whichthe HMGB domains play separable roles tostimulate RAG-mediated 23-RSS binding andcleavage: One HMGB domain may be posi-tioned proximal to the nonamer to promoteRAG-mediated bending of the spacer, and theother may be positioned in or near the heptamerto bind distorted DNA structures induced byRAG binding and/or RSS nicking. Based ontheir known DNA binding and bending prop-erties (137), boxes A and B are speculated tooccupy sites in or near the heptamer and non-amer, respectively. The tandem arrangementof the HMGB domains may provide bridgingneeded between the two sites in the SC to stim-ulate binding, but once synapsis with a part-ner RSS has been achieved, a conformationalchange may occur to stabilize the complex andobviate the function of one of the two HMGBdomains (79). The HMGB1 stoichiometry inprecleavage RAG-RSS complexes remains to beformally defined, but recent biochemical char-acterization of the SEC suggests the ratio ofRAG1:RAG2:HMGB1 is approximately 2:2:1in this complex (54).

SYNAPTIC RAG-RSS COMPLEXES

RAG1/RAG2 Precleavage PairedRSS Complexes Assembledon Short DNA Fragments

As discussed above, single RSS-RAG com-plexes assembled in Mg2+ support mainly RSS


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


nicking. By adding appropriate partner RSS inthe presence of HMGB1, however, the RAGproteins catalyze hairpin formation at muchhigher levels in solution, and under these con-ditions form a PC detectable by EMSA withslightly slower electrophoretic mobility andhigher intrinsic cleavage activity (∼fivefold)than its counterpart lacking partner RSS (131).Time-course studies show that under condi-tions favoring synapsis with consensus oligonu-cleotide RSS substrates, nicking occurs rapidlyand largely independently of synapsis (149)with similar unireactant kinetics and catalyticrate constants for both 12- and 23-RSS sub-strates (0.5–0.6 min−1) (149), whereas hairpinformation occurs at an ∼150-fold slower rate(∼0.004 min−1) (147). On long DNA frag-ments, however, the nicking rate may be stim-ulated slightly (∼twofold) by synapsis (40, 80).Using EMSA-based approaches, subsequentstudies found that SC2 is likely converted tothe PC through the capture of an appropriatepartner RSS without altering the complementof RAG proteins in the complex (98). Evidencefavors a RAG1:RAG2 heterotetramer configu-ration in the PC (54, 129), with RAG2, but notRAG1, capable of reassorting during assembly(98, 129). One important caveat to consider isthat the PCs analyzed in these studies were as-sembled on short DNA fragments. As discussedbelow, DNA fragment length may influenceRAG protein stoichiometry and RSS recog-nition in the precleavage synaptic complex.Nevertheless, characterization of the PC inthese experimental systems has provided valu-able insights into the mechanisms that regulatePC formation and activity, and the major find-ings from these studies are reviewed here.

DNA footprinting studies indicate the RAGproteins afford greater protection from DNAseI cleavage to the heptamer and coding sequencewhen they are in the PC than when in the SC(100). In the PC, the RAG proteins protectabout 12 bp of the coding flank on the bottomstrand from DNAse I cleavage (99), but conferDNAse I hypersensitivity on the top strandabout 12 bp upstream of the heptamer (100).Interestingly, when both RSSs are nicked,

bottom strand protection from DNAse I isextended another ∼4 nt into the coding region(99), and protection of the first and third Gresidues on the bottom strand in the heptamerfrom methylation by dimethyl sulfate is ob-served (102). This observation suggests thatsynapsis is accompanied by additional proteinand/or DNA conformational changes thatpromote more intimate RAG interactions withthe cleavage site, which are further enhancedupon RSS nicking. Because the central domainof RAG1 exhibits sequence-specific recog-nition of the heptamer and ssDNA bindingactivity (103), the most plausible scenario isthat RAG binding imparts local melting of theDNA strands in and near the heptamer that ismediated and stabilized by the RAG1 centraldomain, possibly with assistance by HMGB1.Although DNA footprinting data suggestmelting may be partly induced before substratenicking, biochemical studies showing that base-pair mismatches or abasic sites incorporated atthe coding flank stimulate hairpin formationmore than nicking (27, 53, 107) suggest thatmore pronounced unpairing likely occurs afterthe RSS is nicked, presumably to orient theDNA strands in a configuration to catalyzehairpin formation.

More recently, direct evidence of base un-pairing has been obtained using permanganateoxidation to detect extrahelical thymidineresidues. Two independent studies showedthat the terminal nucleotide on the bottomstrand of the coding flank (C1b) exhibits hy-persensitivity to permanganate oxidation whenthe RAG proteins are incubated with a 12/23pair of nicked oligonucleotide RSS substratesunder conditions favoring synapsis; no hyper-sensitivity is detected if one or both substratesare intact (16, 102). This pattern of sensitivityis reminiscent of what is observed with the Tn5and Tn10 transposases, which, like the RAGproteins, generate DNA hairpin intermediatesduring the recombination process and create ahypersensitive thymidine in the strand oppositethe nicking site (14, 15). Structural studiesof the Tn5 postcleavage synaptic complexrevealed that the extrahelical thymidine is


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


Base flipping: themovement of a basefrom its normalintrahelical position indouble-stranded DNAto a position outside ofthe helix

stabilized by stacking interactions with anaromatic tryptophan residue (trp-298) (30).Searches for a functionally equivalent aromaticresidue in RAG1 by three different laboratorieshave led to somewhat conflicting conclusions(16, 53, 88). Four candidate residues wereidentified on the basis of having a nicking pro-ficient but hairpin defective phenotype (W893,W956, Y935, and F971). Of these, W893 wasfavored by Lu et al. (88) because a cleavagedefect observed with a W893A mutant could berescued by replacing alanine with tyrosine orby introducing base-pair mismatches near thecleavage site. By contrast, Grundy et al. favoredW956 because cleavage defects observed witha W956A mutant could be rescued by incorpo-rating an abasic site at C1b (53). However, bothmutants exhibit activity that is inconsistent witha predicted role in base stacking: (a) Cleavagedefects observed with the W893A mutant arerescued by Mn2+ or good coding flanks (53);(b) the W956A mutant shows significant defectsin both nicking and hairpin formation on anisolated RSS but supports hairpin formationunder conditions favoring synapsis, providedboth RSSs are nicked (53, 88); and (c) RAGphoto-crosslinking to substrates containingiodouridine in place of thymidine at C1b ismodestly reduced, but not lost, by W893A orW956A substitutions (16). This outcome isunlike the loss of photo-crosslinking observedin similar experiments performed with theTn5 transposase (14) and occurs despite thefact that all four RAG1 mutants lose the abilityto promote C1b sensitivity to permanganateoxidation (16). These inconsistencies led tospeculation that the four aromatic residuestested do not directly contact C1b through basestacking but rather play an indirect role in thebase-flipping mechanism operative in V(D)Jrecombination (16), perhaps by mediatingprotein-protein interactions that coordinatethe sensing of synapsis with the induction ofbase flipping. This interpretation gains supportfrom the identification of a gain-of-functionRAG1 mutant in the region (HA3) that exhibitsenhanced permanganate sensitivity of C1b inthe absence of synapsis (102).

The Signal End Complex

Once the RAG proteins complete the RSScleavage steps, the coding ends dissociate invitro, but the signal ends remain tightly boundto the RAG proteins in the SEC (2). This com-plex is more resistant than precleavage com-plexes to dissociation, and the RAG proteinsprotect the ends from access by nucleases orjoining factors (70). These features are reflectedby strong protection of the entire RSS to chem-ical and enzymatic probes by DNA footprint-ing (131). Together, these findings suggest theSEC remains intact until it is physically disas-sembled, perhaps involving the targeted degra-dation of RAG2 (35, 95). The SEC is reportedto contain a RAG1:RAG2 heterotetramer (54,129) and one HMGB1 molecule (54). Charac-terization of the SEC by electron microscopy(EM) and atomic force microscopy (AFM) byGrundy et al. (54) suggests the RAG proteincore forms a pseudosymmetric anchor-shapedcomplex containing two fluke-like extensionsspanning about 150–160 A tip-to-tip and cen-tral shank ∼80–90 A wide and ∼125 A long.Selective immunolabeling experiments suggestRAG2 and the RAG1 NBD localize to theflukes and shank, respectively, of the anchor-shaped complex. SECs formed with DNA frag-ments containing long nonamer end extensionsshow that the DNA adopts a parallel config-uration in ∼80% of complexes, and exits theSEC from adjacent points at the end oppositeof where RAG2 is localized (see Figure 4).

DNA Length-Dependent Effects onRAG Stoichiometry and DNAOrganization in Synaptic Complexes

Although a large body of evidence suggestsa RAG1/2 heterotetramer core complex iscapable of binding and cleaving pairs ofoligonucleotide RSS substrates in vitro, andforming an SEC with short and long DNAfragments, there remains some questionabout how closely models drawn from thesestudies reflect RAG-mediated synapsis oflonger, more physiologically relevant DNA


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


substrates. The Roth laboratory provided thefirst direct evidence showing the RAG proteinsdifferentially bind and cleave short and longDNA fragments. Specifically, Huye & Roth(63) found that the RAG proteins exhibit a

substantial bias against cleaving a 12/23 pairof RSSs in trans (on separate DNA molecules)when the RSSs are embedded in long DNAfragments but not when they are incorporatedin short oligonucleotide substrates.

OHOHHHO

HO OHOHH

OHOHHO

Nonamer bindingheptamer surveillance

Stable complex formation(+/– nicking)

Partner RSS capture (nonamer binding)partner RSS heptamer surveillance

Paired complex formation(+/– nicking at one RSS)

Or

Nicking at both RSSs Activation (base flipping) Double hairpin formation

RAG1/2 association(tetramer core)

RAG1 RAG1

12RSS

23RSS

NonamerHeptamerCoding

CTD

NBD

Cen

tral

CTD

RAG2

Cen

tral

HO

RAG2


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


Additional evidence that DNA fragmentlength influences how the RAG proteins me-diate synapsis comes from recent methyla-tion interference and direct 1,10 copper-phenanthroline (Cu-OP) footprinting studiesof RAG-RSS complexes assembled on longDNA fragments containing RSSs paired in cis(80). This study found evidence that the RAGproteins (a) mediate contact with the top strandof the coding flank ∼14 nt from the cleavagesite [close to a DNAse I hypersensitive site de-tected in the PC by Nagawa et al. (100)] andinduce a conformational change evidenced bya shift in CuOP sensitivity that is propagateddistally through another turn of the DNA he-lix, and which is partially alleviated by substratenicking; (b) induce conformational changesin the intersignal sequence that are coupledto substrate nicking and promoted by full-length RAG1; and (c) exhibit a DNA length-dependency in cleavage (hairpin) efficiency thatinvolves putative coding end contacts with theC terminus of RAG1. The last finding is evi-denced by the observation that a K980A RAG1mutant exhibits selectively impaired hairpinformation, but not nicking, on long DNA frag-ments associated with an inability to induceconformational changes in the intersignal se-quence. This outcome was interpreted to sug-gest that after nicking, but before hairpin for-mation, there is an activation step that involvesa significant conformational change to orientthe DNA strands in a configuration suitablefor catalyzing hairpin formation. The fact thatthe DNA structural distortion occurs in the in-tersignal sequence distal to the nonamer that

is contacted by the RAG1 NBD, yet appar-ently triggered by C-terminal RAG1 interac-tions with the coding flank, suggests that DNAstrand reorientation involves both the N- andC-terminal portions of RAG1. It is tempting tospeculate that the activation step involves baseflipping at C1b, as this step occurs after nicking,and is negatively affected by C-terminal RAG1mutations proximal to K980 (16, 102).

What could account for differences inhow short and long DNA fragments arerecognized and cleaved by the RAG proteins?One possibility is that synapsis of long DNAfragments requires a larger complement ofRAG proteins than is found in PCs assembledon short oligonucleotide substrates. In supportof this idea, Landree et al. compared thekinetics of RAG-mediated cleavage betweenwild-type and catalytically inactive RAG1 het-erodimers on long DNA fragments containingRSSs paired in cis and concluded the synapticcomplex formed on this substrate contains apair of RAG1 dimers (81). Consistent withthis possibility, recent studies show that whenprecleavage synaptic complexes are assembledwith long DNA fragments containing embed-ded RSSs in cis or in trans, discrete RAG-RSScomplexes are visualized by EMSA that sup-port 12/23-regulated cleavage activity butexhibit much slower electrophoretic mobilitythan comparable complexes assembled usingoligonucleotide substrates (80, 119).

Additional support for this possibility comesfrom a recent study using AFM to directlyvisualize and characterize RAG precleavagesynaptic complexes assembled on long DNA

←−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−−Figure 4Integrated model of RAG synaptic complex organization, assembly, and RSS cleavage mechanisms. ARAG1:RAG2 heterotetramer core initially binds an RSS (e.g., the 23RSS here) via the NBD of RAG1, andRAG1 and RAG2 collaborate to survey the sequence spanning the spacer and heptamer region for a suitablebinding site. If identified, the RAG proteins and DNA undergo conformational changes to enhanceprotein-DNA interactions and to assume geometries that facilitate nicking. This process is repeated tocapture an appropriate partner RSS. Nicks may or may not be introduced prior to synapsis, but uponsynapsis, RSS nicking is completed, and the complex undergoes a further activation step associated with baseflipping at C1b (see inset). Hairpin formation proceeds in a concerted manner at both RSSs to complete thecleavage reaction. Based on structural and biochemical studies on a given RSS, one subunit of the RAG1dimer interacts with the nonamer, whereas the other RAG1 subunit interacts with the spacer and heptamerand catalyzes both steps of the cleavage reaction (cleavage in trans).


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


fragments in trans in the presence of HMGB1and Mg2+ (119). In these experiments, the esti-mated mass of the RAG synaptic complex, basedon volume measurements of these complexes,was found to be close to the mass predicted for aRAG1/2 heterooctamer (estimated: ∼760 kDa;predicted: 816 kDa). In these complexes, anal-ysis of DNA path lengths and the distance fromthe center of the synaptic complex to the endsof the protruding DNA further suggests that(a) the RAG proteins specifically bind the RSSon each DNA fragment, (b) the two DNA frag-ments are preferentially bound in a side-by-sideconfiguration, rather than crossing over in thecomplex, and (c) the two DNA fragments favora parallel over an antiparallel arrangement ofRSSs by a factor of ∼2:1. This outcome isconsistent with the greater efficiency of in vitrocoupled cleavage and in vivo V(D)J recombina-tion observed with DNA substrates that morereadily support parallel synapsis than antipar-allel synapsis (39, 116). Why might a largerRAG complex be necessary to support synapsisof long DNA fragments? One possibility isthat a higher-order (heterooctameric) RAGcomplex is required to overcome electrostaticor steric repulsion between long DNA strandsthat otherwise pose a significant barrier tosynapsis in vitro. These repulsive forces wouldbe less severe for short DNA fragments, andin this case, a RAG1/2 heterotetramer mightprovide sufficient shielding to overcome thisbarrier and permit synapsis to occur.

The RAG protein content estimated insynaptic complexes studied by Shlyakhtenkoet al. (119) contrasts with results obtained byGrundy et al. (54), who found that RAG pre-cleavage synaptic complexes, although moreheterogeneous in size than SECs, containedno more than a RAG1/2 tetramer on average.One important difference between these twostudies is that Grundy et al. assembled RAGprecleavage complexes in Ca2+ on DNAfragments that contained a short coding flank(16–26 bp). Shlyakhtenko et al. reported thatRAG-RSS synaptic complex size depends onthe composition of the metal ion cofactor andthe length of the coding flank (the transition

to a higher-order complex requires ∼100 bp offlanking sequence) (119). Thus, it is plausiblethe conditions used by Grundy et al. favoredassembling the lower-order RAG1/2 tetramercomplex. Alternatively, it is possible that theRAG protein stoichiometry changes duringthe course of the cleavage reaction from aRAG1/2 heterooctamer in the precleavagesynaptic complex to a RAG1/2 heterotetramerin the postcleavage SEC. This model has an at-tractive feature of enabling a RAG1/2 tetramerto accompany both sets of DNA ends aftercleavage and provides a potential mechanismto explain how the RAG proteins help directend joining of both signal and coding ends.

The observation that synaptic complexes as-sembled on long DNA fragments favor a side-by-side configuration of DNA strands with theRSSs arranged in a parallel orientation con-trasts with an earlier study by Ciubotaru et al.(22) that used FRET to interrogate the distancebetween the ends of oligonucleotide 12RSS and23RSS substrates bound in a PC. In that study,the data favored a model in which the RSSsadopt a bent and crossed configuration in thePC. Although this apparent discrepancy maypartly be attributed to DNA length-dependentdifferences in how the RAG proteins bind shortoligonucleotides and long DNA fragments, theDNA strands in synaptic complexes analyzed byEM and AFM were not sufficiently resolved todetermine the precise alignment of the DNAstrands. Some degree of DNA bending andstrand crossover can easily be accommodatedin structural models of the RAG1/2 heterote-tramer complex developed from EM analysis ofSECs, especially along the shank of the anchor-like projection, yet still remain consistent withdata obtained by FRET.

RULES GOVERNING RSSRECOGNITION AND PAIRING

V(D)J recombination is guided by the 12/23rule and existing evidence favors a capturemodel for synaptic complex assembly. How-ever, the extent to which the capture modelapplies in vivo remains an open issue, and there


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


Recombinationcenter: region ofantigen receptor locuswith strong RAGbinding and high levelsof RNA polymerase II,histone acetylation,and H3K4me3

is also uncertainty concerning which RSS (12or 23) might be recognized first, how the choiceof a partner RSS is made, and when during PCassembly the nicking reactions occur. In thissection, we consider current evidence bearingon these and related questions.

THE 12/23 RULE AND THECAPTURE MODEL

V(D)J recombination in vivo exhibits a strong(>30-fold) preference for a 12/23 RSS pair over12/12 and 23/23 pairs, and this preference canbe recapitulated in DNA cleavage assays usingthe core RAG proteins and HMGB1 under cer-tain conditions (reviewed in 47, 51, 84). Hence,it is likely that the 12/23 rule is imposed at orprior to the generation of DSBs. Indeed, thereis evidence that both RSS synapsis and hairpinformation are more efficient with 12/23 thanwith 12/12 or 23/23 RSS pairs (reviewed in47, 51). The mechanisms underlying preferred12/23 synapsis and cleavage remain unknown.

Two studies have provided evidence for a12RSS-first model in which PC formation oc-curs by initial RAG binding to the 12RSS fol-lowed by capture of a 23RSS. Biochemical ex-periments by Jones & Gellert (71) using coreRAG proteins and a truncated form of HMGB1lacking the C-terminal acidic tail, found thatadherence to the 12/23 rule was influenced bythe order in which the proteins were allowedto bind to the RSSs. When RAG and HMGB1were bound first to the 12RSS, a very strongpreference for synapsis and cleavage of a 23RSSversus a 12RSS partner was observed, but ifthe order of RSS addition was reversed, thepreference for a 12RSS versus a 23RSS part-ner was much smaller (5–6 fold). In the secondstudy, Curry et al. (28) examined RSS nickingat Igh and Igk in pre-B cell lines and at Tcrain primary thymocytes and observed nicks at12RSSs (flanking Dh, Vκ, and Jα gene seg-ments) but not 23RSSs (flanking Vh, Jκ, and Vα

gene segments). These data are consistent witha model in which RAG binds and nicks a 12RSSprior to 23RSS capture, and argue against initial

binding to a 23RSS because, if so, 23RSS nick-ing should have been detected.

There are several considerations to take intoaccount when interpreting the findings of thesetwo studies. First, when HMGB1 is present,the core RAG proteins bind equally well to12RSS and 23RSS substrates (reviewed in 51,131). Hence, if 12RSSs are to be bound firstin vivo, a mechanism would have to exist tomake them the preferred binding sites for RAG.Second, there is no evidence that two 23RSSsare better able to support V(D)J recombinationor DNA cleavage than two 12RSSs (59, 127), asmight have been expected based on biochemi-cal studies by Jones & Gellert (71). As discussedpreviously (131), it is conceivable that use oftruncated HMGB1 in this study allowed forenhanced synapsis and cleavage between two23RSSs. And third, nicking probably does notprovide a reliable measure of RAG binding be-cause the rate of nicking is strongly influencedby the DNA sequence of the RSS and the flank-ing coding nucleotides (37, 148).

Two recent studies have cast doubt onthe general applicability of the 12RSS-firstmodel for PC formation. In the first, nickingwas assessed at Tcrb gene segments in earlydeveloping thymocytes and in biochemicalexperiments; in both assays, nicks were moreabundant at the 23RSS 3′ of Dβ1 than at Jβ12RSSs (see Figure 1b for diagram of Tcrb),suggesting that during Dβ1-to-Jβ recombina-tion, RAG binds first to the 3′ Dβ1 23RSS andthen captures a Jβ 12RSS (45). In the second,direct assessment of RAG1 and RAG2 bindingto antigen receptor loci in primary developinglymphocytes by chromatin immunoprecip-itation (ChIP) revealed no preference forRAG binding to 12RSSs over 23RSSs (67).Both RAG1 and RAG2 were found to bindselectively to small regions of highly activechromatin in the Igh, Igk, Tcrb, and Tcra loci.These regions, referred to as recombinationcenters, span some or all of the J gene segmentsin all four loci as well as J-proximal D genesegments in Igh and Tcrb. Importantly, whereasJ gene segments in Tcrb and Tcra are flankedby 12RSSs, those in Igh and Igk are flanked


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


by 23RSSs. RAG binding correlated well withelevated levels of histone 3 acetylation andH3K4me3 (both markers of transcriptionallyactive chromatin), and there was no evidencethat 12RSSs were selectively located in open,accessible chromatin. It remains unclear whyCurry et al. (28) observed RSS nicking atVκ but not Jκ gene segments, whereas ChIPrevealed RAG binding at Jκ but not Vκ genesegments (67). One possibility is that Igkrecombination is initiated when RAG binds toa small number of Vκ gene segments in eachpre-B cell, but evidence of this went undetectedin the ChIP analysis because of the large num-ber of Vκ gene segments (>100). Alternatively,RAG binding might occur first at Jκ 23RSSsbut is not accompanied by efficient nicking, asproposed in a recent, speculative model (114).

An appealing model arising from the RAGChIP data is that RAG1 and RAG2 bindto RSSs in accessible chromatin regardlessof whether they are 12RSSs or 23RSSs (67,114) and hence that initial RAG binding isdictated by RSS accessibility. It seems likely,however, that other mechanisms in additionto accessibility influence the in vivo pattern ofRAG-RSS binding. For example, endogenousRSSs exhibit considerable sequence variationand some of these variations, particularlyin the spacer and nonamer, are expected tosubstantially affect RAG binding affinity (44,47, 131). In addition, particular RSSs mighthave special mechanisms designed to facilitateRAG binding. For example, in the Tcrb locus(Figure 1b), the 23RSSs 3′ of Dβ1 and Dβ2contain an embedded binding site for the AP1transcription factor (Figure 1b,c), and thec-Fos protein, which is a component of AP1, isable to interact with RAG and facilitate RAGloading onto the 23RSSs of Dβ1 and Dβ2(143). The AP1 and RAG binding sites in Dβ

23RSSs overlap (Figure 1c), raising interestingquestions about the mechanism by which AP1facilitates RAG binding (discussed in 134).RAG binding to Dβ 23RSSs would selectivelysupport Dβ-to-Jβ recombination and couldinterfere sterically with RAG binding to Dβ

12RSSs (Dβ1 and Dβ2 are only 12 bp or 14

bp long, respectively). This might contributeto developmentally ordered recombination ofTcrb (Dβ-to-Jβ before Vβ-to-Dβ). This modelis consistent with the finding that mutation ofthe 3′ Dβ 23RSS allows for readily detectabledirect Vβ-to-Dβ recombination (123). Ad-ditional evidence for this model was recentlyobtained by two groups, who showed that c-Fosand the Dβ1 23RSS inhibit DNA cleavageand recombination at the 5′ Dβ1 12RSS (45,143), and that direct Vβ-to-Dβ recombinationoccurs at an increased frequency in Fos−/−as compared to wild-type thymocytes (143).AP1-directed binding of RAG to 3′ Dβ 23RSSsmight help explain the observation, discussedabove, that nicks are detected at higher levelsat the Dβ1 23RSS than Jβ RSSs (45).

A similar mechanism has been proposed tofacilitate RAG binding to Vh 23RSSs. Thetranscription factor PAX5 has binding sites inthe body of many Vh gene segments, interactswith RAG, and can increase RAG-mediatedcleavage at Vh RSSs in biochemical assays andrecombination of Vh gene segments on epi-somal recombination substrates in vivo (152).This is intriguing given the importance ofPAX5 for recombination of DJh-distal Vh genesegments and for Igh locus contraction, which isthought to bring distal Vh gene segments intoclose proximity of Dh-Jh and allow for synap-sis and Vh-to-DJh recombination (50, 65). It ispossible that PAX5 binding to these sites helpsdeliver RAG to Vh RSSs prior to synapsis witha Dh 12RSS. Alternatively, since Vh gene seg-ments lie outside of the Igh recombination cen-ter (67), PAX5 may help to deliver Vh RSSs toRAG within the recombination center (134).

In summary, initial RSS binding by RAGis likely determined by a superposition of sev-eral mechanisms, including accessibility withinchromatin and the action of RAG-interactingtargeting factors, but a consistent preference forthe 12RSS is not supported by current evidence.

PATHWAYS OF RAGRECRUITMENT IN VIVO

DNA binding, protein-protein interaction, andother types of biochemical experiments had led


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


to the expectation that RSS recognition in vivowould be mediated by a complex containingboth RAG1 and RAG2 (reviewed in 47, 51,131). It was surprising, therefore, when ChIPexperiments revealed that RAG1 and RAG2each exhibited highly specific patterns of chro-matin binding in vivo, including recruitmentinto recombination centers, in the absence ofthe other (67). This led to the proposal thatthere are three distinct pathways for RAG re-cruitment to specific sites in chromatin (dis-cussed in 67, 114): (a) RAG1 binds to the RSSand RAG2 is recruited later by virtue of itsability to interact with RAG1 and H3K4me3;(b) RAG2 uses its PHD finger to bind to nu-cleosomes containing H3K4me3, and RAG1is recruited afterwards by binding to RAG2,perhaps without involving a direct interac-tion with DNA; and (c) a preformed RAG1-RAG2 complex binds to an accessible RSS, toH3K4me3, or to both. The relative contribu-tions of these three pathways are not known,nor is it known what fraction of the RAG1 andRAG2 molecules in the nucleus are associatedwith one another or how dynamic this associa-tion is (114).

Nonspecific DNA binding by RAG1 or bythe RAG1-RAG2 complex might also influ-ence the distribution of the RAG proteins inthe genome. As noted above, RAG1 can rec-ognize the RSS in a sequence-specific manner,whereas inclusion of RAG2 increases the affin-ity and specificity of the interaction (131). Zhaoet al. (153) recently reported that core RAG1binds strongly to nonspecific DNA and thatthis interaction can be dramatically reduced byRAG2. They also observed a relatively stableRAG1–12RSS complex (half-life of 558 s), lead-ing to the proposal that there are two distinctmodes of DNA binding by RAG1: a nonspe-cific mode that relies heavily on electrostaticinteractions and a sequence-specific mode thatis less dependent on electrostatic interactionsand might involve conformational changes inthe protein and the DNA (153). Previous stud-ies had not observed such strong interactionsof RAG1 with nonspecific DNA (6, 23, 126);the reasons for this discrepancy are not clear.

Strong nonspecific DNA binding by RAG1, asobserved by Zhao et al., predicts the possibil-ity of RAG1 binding to many open sites in thegenome where DNA is relatively exposed (suchas promoters and other nuclease hypersensitivesites; such binding should not only be RAG2-independent, but might actually be inhibited byRAG2. In addition, the ability of RAG1 to formrelatively stable, sequence-specific complexescould help explain the RAG2-independent ac-cumulation of RAG1 in recombination cen-ters, which contain multiple, accessible RSSs.It is plausible that HMGB1 further enhancesRSS binding in vivo by RAG1 because RAG1and HMGB1 form a specific complex with the12RSS even in the presence of a large excessof nonspecific DNA (66) and mediate synap-sis of two RSSs in solution in a manner thatis enhanced only modestly by RAG2 (146).Genome-wide analyses of RAG1 binding pat-terns should provide substantial insight intosome of these issues.

THE BEYOND 12/23RESTRICTION

The 12/23 rule dictates that recombination willbe inefficient between two 12RSSs or betweentwo 23RSSs, but does it also mean that recombi-nation will occur efficiently between any 12/23RSS pair? Analysis of the Tcrb locus has re-vealed that the answer to this question is no.Vβ gene segments invariably recombine with aDβ 12RSS rather than with one of the multi-ple Jβ 12RSSs that typically remain in the locus(Figure 1b). Why is this? Elegant in vivo exper-iments involving a transgenic Tcrb minilocus ormodified endogenous Tcrb loci (reviewed in 18,139) have demonstrated that Vβ gene segmentsdisplay a very strong preference to recombinewith the 12RSS 5′ of Dβ1 over any of the Jβ112RSSs, and this is not due to the positioningof the various RSSs or to competition betweenDβ and Jβ 12RSSs (11, 123). Importantly, thisB12/23 restriction can be reconstituted withepisomal recombination substrates in nonlym-phoid cell lines, demonstrating that neitherlymphoid-specific factors nor residence within


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


the genome is essential for enforcement of therestriction (74, 103, 142). Furthermore, usingthe core RAG proteins, HMGB1 or HMGB2,and long DNA substrates, Vβ 23RSSs exhibita five- to tenfold preference for cleavage withthe Dβ1 12RSS over a Jβ 12RSS (74, 103). No-tably, this preference can be as high as 50- to500-fold in vivo with episomal recombinationsubstrates (74) and is likely much greater than100-fold during endogenous Tcrb recombina-tion (11, 123). Hence, although the B12/23 re-striction is substantially determined by the in-trinsic properties of interactions between RAGand Tcrb DNA substrates, additional mecha-nisms might also contribute.

One way to explain preferential recombina-tion of Vβ with DJβ as opposed to Jβ wouldbe if RAG binds to Dβ 12RSSs better thanto Jβ 12RSSs. EMSA experiments utilizing thecore RAG proteins reveal no such preference,and suggest that Dβ and Jβ 12RSSs are boundmuch less efficiently by RAG than the consen-sus 12RSS (37, 103). Nor is there any indica-tion from ChIP experiments that RAG bindingis stronger at Dβ as opposed to Jβ gene seg-ments (67). It is not yet clear, however, how wellRAG ChIP patterns reflect RSS occupancy, orwhether the pattern of RAG binding to Tcrbor Igh changes after D-to-J recombination hasoccurred. It is therefore possible that, afterDβ-to-Jβ recombination, RAG binding occurspreferentially to the 12RSS 5′ of the DJβ junc-tion as opposed to the remaining Jβ 12RSSs.This would be consistent with the detection ofnicks 5′ of Dβ1 but not at Jβ gene segmentsin vivo (45). Perhaps the close proximity of the12RSS 5′ of DJβ to the recent D-to-J joiningevent creates a favorable context for RSS en-gagement by RAG.

Alternatively, a plausible interpretation ofexisting data is that DJβ and Jβ 12RSSs are oc-cupied by RAG with roughly equal efficiency:If this is the case, how is the B12/23 restric-tion enforced? One possibility, suggested by abiochemical analysis using core RAG proteins,truncated HMGB2, and oligonucleotide RSSsubstrates, is that Jβ gene segments are crip-pled for recombination with Vβ gene segments

(37). Jβ substrates were shown to nick slowlyand/or to synapse weakly with Vβ substrates,whereas the Dβ1 12RSS exhibited more ef-ficient synapsis and more rapid nicking thanthe Jβ 12RSSs tested. This leads to a model inwhich initial RAG binding to a Jβ 12RSS fails tosupport recombination with Vβ either becauseslow nicking allows RAG to dissociate from Jβprior to synapsis or hairpin formation, or be-cause, in the case where Jβ nicking does occur,the synaptic complex does not form or is unsta-ble. The Dβ1 12RSS exhibits neither defect andhence, once bound by RAG, nicks rapidly (help-ing to retain the bound RAG) and synapses withVβ efficiently (37). This in turn might increasethe efficiency of Vβ nicking (45), allowing forbase flipping at both RSSs and subsequent hair-pin formation. It will be important to confirmthe crippled nature of Jβ gene segments in ex-periments using full-length HMGB1 proteinand highly active preparations of the RAG pro-teins because in the published analysis, coupledcleavage between Vβ14 and either of the Dβ

12RSSs was inefficient (37).If indeed Jβ gene segments are crippled for

recombination, how does Dβ-to-Jβ recombi-nation occur with sufficient efficiency? As dis-cussed above, AP1 may help promote efficientloading of RAG onto the Dβ 23RSSs. In ad-dition, the Dβ1 23RSS forms a stable synap-tic complex with Jβ gene segments 2- to 3.5-fold more efficiently than does the Vβ14 23RSS(37). Recent data also raise the possibility thatnicking of Jβ is enhanced by synapsis with theDβ1 23RSS (45). It might also be relevantthat Dβ and Jβ gene segments coexist in rel-atively close proximity to one another withinrecombination centers, where most or all of theRAG protein associated with Tcrb appears to belocated (67).

Overall, the B12/23 restriction appears to becontrolled in large part by intrinsic features ofRAG interactions with Tcrb gene segment sub-strates, perhaps complemented by mechanismsthat have yet to be recapitulated biochemically,such as nucleosome positioning, histone mod-ifications, and the consequences of DNA re-pair of a prior recombination event. It will be


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


important to determine the extent and stabilityof RSS occupancy and synapsis by RAG bothbefore and after Dβ-to-Jβ recombination.

CAPTURE AND RELEASE:SAMPLING MULTIPLEPARTNERS PRIOR TORECOMBINATION?

The half life of the PC has not been determinedin solution; even more uncertain is how sta-ble the SC or PC is in vivo, where the RAGproteins must compete with histones and otherDNA binding proteins for RSSs. It is unknownwhether SC formation involves multiple cyclesof RSS binding and release by RAG, or howmany times the SC samples partner RSSs priorto formation of a PC that is sufficiently sta-ble to support hairpin formation. Studies of theTcrb locus have suggested the idea that stablePC formation and recombination do indeed in-volve cycles of capture and release of a partnerRSS. The initial evidence for this came fromthe analysis of mice harboring a Tcrb allele inwhich the 23RSS flanking Vβ14 was replacedwith the strong Dβ1 23RSS (144). This re-placement resulted in a large (sevenfold) in-crease in the frequency of Vβ14 recombina-tion events and of Vβ14-positive thymocytesand peripheral T cells, formally demonstratingthat RSS sequence can influence recombina-tion efficiency. On the assumption that the RSSswap did not increase the frequency with whichan accessible Vβ14 comes into close proximityof DJβ, these data support the conclusion thatin normal thymocytes, Vβ14 has multiple en-counters with DJβ, most of which do not leadto recombination—that is, recombination typi-cally occurs after multiple cycles of partner RSScapture and release (144).

When the Vβ14 RSS was replaced with anunrecombined Dβ1-Jβ1 cassette, the cassettewas observed to undergo recombination in al-most 90% of developing thymocytes, far higherthan the frequency of primary Vβ14 recom-bination events (7%) (108). The simplest in-terpretation is that the unmanipulated Vβ14gene segment resides in accessible chromatinin a high percentage of developing thymocytes

and that recombination of this gene segment islimited by its ability to couple productively withDJβ rather than by its accessibility. This wouldargue against the possibility that replacing theVβ14 RSS with the Dβ 23RSS increased Vβ14usage primarily by increasing Vβ14 accessibil-ity, and provides support for the interpreta-tion that Vβ14 recombination involves cyclesof capture and release.

It is plausible that Vβ14 is not unique andthat many or all of the Vβ gene segmentsengage in cycles of capture and release with theDJβ region. If correct, the different Vβ genesegments might compete with one anotherfor stable PC formation and recombinationwith DJβ. Competition of this sort predictsthat elimination of some Vβ gene segmentswould increase the recombination frequency ofthe remaining ones. Analysis of heterozygousthymocytes containing a Tcrb allele lacking 18of the 21 functional Vβ gene segments yieldedfindings consistent with this prediction: Twoof the three Vβ gene segments on the deletedallele showed significantly elevated recombina-tion frequencies (13). Importantly, the deletedallele recombined somewhat less efficientlythan the intact allele, ruling out the possibilitythat the outcome could be attributed to agenerally elevated recombination efficiency onthe Vβ-deleted allele. These results led theauthors to propose that Vβ gene segmentscompete with one another for productivesynapsis with DJβ (13).

Together with the finding of focal RAGbinding in recombination centers, these stud-ies suggest a model (described in 67, 114) inwhich V gene segments in looped/contractedchromatin cluster around a recombination cen-ter and compete with one another to be cap-tured by RAG bound to J or DJ. This competi-tion would typically involve multiple cycles of Vsegment capture and release, during which theV RSS is sampled by RAG within the recom-bination center. A better and/or more accessi-ble RSS increases the probability that any givencapture event would lead to a stable PC andto recombination. This is consistent with datasuggesting that RSS sequence variations help


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


explain primary recombination frequencies (re-viewed in 44). Because RAG1-RSS interactionsare largely centered on the nonamer, whereasRAG1/2-RSS interactions extend through thespacer and heptamer, it is tempting to specu-late that under normal circumstances, the sam-pling process involves primary identification ofa nonamer-like sequence by RAG1, followedby surveillance of the heptamer and spacer se-quence for suitable sites of interaction in col-laboration with RAG2. Stable complex forma-tion occurs when an appropriate sequence isidentified and is expected to be accompaniedby significant RAG and DNA conformationalchanges that provide the required geometry tocatalyze the biochemical steps of the cleavagereaction. We envision this sampling processcould operate at both the initial RSS bindingand subsequent RSS capture steps of synapticcomplex assembly. A simplified model of thisprocess that is based on the RAG1/2 organiza-tion and subunit stoichiometry suggested fromEM and AFM studies of the SEC is shown inFigure 4.

There is, however, a caveat associated withthe data that support the concept of cap-ture and release. Specifically, the sequencesused to replace the Vβ14 RSS included theDβ1 23RSS and its embedded AP1 bindingsite (108, 144). This arrangement offers thepossibility of cFos-mediated RAG recruitmentat Vβ14, which might in turn alter proper-ties of Vβ14, including its accessibility andpropensity to become juxtaposed with the bonafide recombination center at DJβ, therebyincreasing Vβ14 recombination. It will beimportant to determine whether RAG bindingcan be detected at Vβ14 in these mutant alle-les because no binding at Vβ14 is detected inwild-type alleles (67). A more definitive resolu-tion of this issue will likely require the creationof additional RSS replacement alleles.

RAG NONCORE REGIONS

RAG1 Noncore Regions

As discussed above, the N-terminal third ofRAG1 is dispensable for the catalytic activity of

the RAG complex but required for efficient andhigh-fidelity V(D)J recombination. This regioncontains several conserved cysteine-containingelements (C1, C2, C3, CH), four basic motifs(BI, BIIa, BIIb, BIII), and a C3HC4 RINGdomain flanked by a classical C2H2 zinc fingerdomain (Figure 3). Mutation of the cysteine-containing elements causes modest (C3, CH) tosevere (C1, C2) reduction in the V(D)J recom-bination activity of RAG1 (112). Full-lengthRAG1 is expressed at lower levels than coreRAG1; mapping studies (55, 93, 109) showlittle or no improvement in expression bytruncating the first 134–150 residues, whereasfurther truncation to residue 211 results ina modest increase in expression. Additionaltruncation between residues 238–265 resultsin expression levels at or near core RAG1levels. Thus, the RING and zinc finger do-mains between residues 265 and 383 are notdirectly responsible for the diminished RAG1expression levels associated with the full-lengthprotein. The region between residues 216 and238 contains the BIIa basic motif that mediatesinteractions with KPNA1 (see below) (125)and is necessary to maintain the high recombi-nation levels of full-length RAG1 (93), as wellas a lysine residue (K233) that undergoes auto-ubiquitination by an isolated RAG1218−389

fragment (72). However, alanine or argininesubstitution at K233 alone or in combinationwith K234 and K236 in the context of full-length RAG1 does not fully restore expressionto core RAG1 levels (55), suggesting thatubiquitination of K233, if it occurs at this sitein vivo, plays at most a partial role in regulatingfull-length RAG1 expression.

The RING domain defines a family of E3ubiquitin ligases known for their ability tofacilitate the attachment of ubiquitin (Ub),a small (76-amino-acid) protein modifier, toother proteins to alter their function or stability(for review, see 34). E3 Ub ligases promotetransfer of Ub from an E2 Ub carrier proteinto an internal lysine residue on the targetsubstrate and dictate target specificity, eitheralone or in combination with the E2 protein.Through this process, targets proteins may


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


become monoubiquitinated at one or severalsites, or through multiple iterations becomepolyubiquitinated. The Ub chain length andcomposition largely determine the biologicaloutcome of ubiquitination, with monoubiqui-tination and some types of polyubiquitin chainsgenerally serving to alter protein function,whereas other polyubiquitin chains (e.g., thoseassembled on Lys48) typically direct the targetto the 26S proteasome for degradation.

The presence of a RING domain in RAG1suggests that it functions as an E3 Ub lig-ase, and its importance in regulating V(D)J re-combination is suggested by the identificationof RING domain mutations associated withOmenn syndrome (hC328Y; see Figure 3). Insupport of this contention, two reports showedthat isolated RAG1 fragments containing theRING domain and all (RAG11−383) or a por-tion (RAG1218−389) of the upstream sequencesupport ubiquitination of a test substrate(S protein) (150), and RAG1 autoubiquitinationprimarily at K233 (72). A fragment containingonly the RING domain (GST-RAG1250−350)reportedly promotes polyubiquitin chain for-mation (150). Subsequently, RAG1218−389 wasfound to mediate ubiquitination of karyopherinalpha 1 (KPNA1; also called importin alpha5, SRP1, and RAG cohort protein 2) (122).KPNA1 is a protein implicated in nuclear traf-ficking that was originally identified as a RAG1-interacting factor using a yeast two-hybridscreen (25) and subsequently shown to bindthe BIIa motif upstream of the RING domain(125). RAG1 autoubiquitination and RAG1-mediated ubiquitination of KPNA1 have yet tobe reconstituted in vitro in the context of full-length RAG1.

Recently, Grazini et al. (52) reported thatfull-length, but not core, RAG1, expressed in293 cells, associates with histone H3 (vari-ant unidentified), and that full-length RAG1and RAG2 support monoubiquitination of H3in vitro (using RAG immunoprecipitates) andin 293 cells. However, in these assays, RAG-mediated ubiquitination was not particularlyrobust and was only mildly impaired by muta-tions in the RING domain (H307A or C325G).

This result raised some question about whetherRAG1 directly mediates H3 ubiquitination andwhether RAG1 selectively ubiquitinates dif-ferent variants or posttranslationally modi-fied forms of H3. Very recently, Jones et al.reported that a bacterially expressed RAG1fragment containing the entire RAG1 N ter-minus (RAG11−389), but not a RAG1218−389

fragment, supports mainly monoubiquitinationof full-length histone H3 contained within apreparation of mixed histones purified froma RAG1−/− pro-B cell line (69). By analyz-ing the monoubiquitinated H3 protein by massspectrometry, it was defined as histone vari-ant H3.3 that was posttranslationally modi-fied by acetylation (K14, K18, and K23) andSer31 phosphorylation. Consistent with thisfinding, recombinant histone H3.1 and H3.3proteins were poor substrates for in vitro ubi-quitination, and antibodies specific for H3K23aand H3.3S31p inhibited in vitro ubiquitina-tion (anti-H3, H3K9a, and H3K18a antibodieshad little or no effect). Whether these findingsare recapitulated in the context of full-lengthRAG1 remains untested.

To establish a relationship betweenRAG-mediated ubiquitination and V(D)Jrecombination, several groups have analyzedthe V(D)J recombination activity of RAG1RING mutants. Grazini et al. (52) showed thatRAG1 H307A and C325G single or doublemutants exhibit impaired V(D)J recombinationactivity on extrachromosomal and endogenoussubstrates, with the defect being milder forcoding joint formation (40% to 60% reductionversus wild type) than signal joint formation(∼80% reduction versus wild type), but in thisstudy, RAG-mediated DNA break inductionon extrachromosomal substrates was unalteredby RING domain mutations. Another RAG1RING domain mutant, P326G, characterizedby Simkus et al. (120), showed similar defectsin signal and coding joint formation to thoseinvestigated by Grazini et al. Interestingly,however, these authors found that a C325Ymutant supports much less recombinationactivity (>50-fold reduction versus wild type)than reported for the C325G mutant, which


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


they attributed in part to increased proteininstability (no folding defects were observedwith the P326G mutant). Whether a RAG1C325Y mutant initiates cleavage as well aswild-type RAG1 was not tested. However,both the C325Y and P326G mutants fail tosupport autoubiquitination in an in vitro assay.Analysis of an additional panel of RAG1 RINGdomain mutants (N265A, H270A, I292A,E294A, S327F) suggests there is a positivecorrelation between ubiquitin ligase activityin vitro and cellular V(D)J recombinationactivity in vivo (121). However, a formallink between these processes remains to beestablished because it requires showing thatV(D)J recombination is impaired by the lossof the substrate targeted for RAG1-mediatedubiquitination and more specifically mutationof the target ubiquitination site(s).

Studies of RAG1-mediated ubiquitinationconducted to date must be interpreted withsome caution because although some RINGdomain–containing proteins function as anE3 Ub ligase as a single subunit, other RINGdomain proteins mediate ubiquitination inthe context of multisubunit assemblies (34).For example, in the cullin family of RINGUb ligases (CRL), the RING domain subunitassociates with a cullin scaffold protein anduses different adaptor and substrate receptorsto assemble unique E3 Ub ligase complexes. Itis notable in this regard that the CRL complexCul4-DDB1-Roc1 has been shown to mediatehistone H3 and H4 ubiquitination in responseto UV irradiation (142). Whether the E3 Ubligase activity and target specificity of RAG1are modulated through its association withother factors remains unclear. It is also possiblethat full-length RAG1 directly or indirectly as-sociates with other E3 Ub ligases, a possibilityoffered by Jones et al. (69) to explain the rela-tive insensitivity of histone H3 ubiquitinationto RAG1 RING domain mutations observedby Grazini et al. In addition, the E2 pairingswith RAG1 profoundly influence its E3 ligaseactivity in vitro: UbcH3 (CDC34) supportsRAG1 autoubiquitination (72), UbcH5b sup-ports polyubiquitin chain formation (150), and

histone H3 ubiquitination (52), and UbcH2supports ubiquitination of KPNA1 and acety-lated/phosphorylated histone H3.3 (69, 122).The relevant E2 carrier proteins supportingphysiological RAG1-dependent ubiquitina-tion in vivo remain to be established. Thus,the E3 Ub ligase activity of RAG1 is likelyto be complicated and context dependent.Nevertheless, there is good reason to suspectubiquitination plays an important role in V(D)Jrecombination, as it is implicated in regulatingDNA repair processes in response to otherforms of DNA damage (7, 138). How might itdo so? Grazini et al. (52) have speculated thathistone H3 ubiquitination by RAG1 enhancesthe efficiency of the repair steps of V(D)J re-combination by serving to mark nucleosomesat RAG-mediated DNA breaks to facilitatedetection and joining of these intermediates.Jones et al. (69) refine this model by suggestingthat modified histones associated with recom-binationally active loci are selectively targetedfor RAG1-mediated ubiquitination. Whetherthere is a mechanistic link between DNA breakinduction by the RAG proteins and the acti-vation of RAG1 E3 Ub ligase activity remainsunclear, but it is tempting to speculate that thetwo events are coupled in some manner, pos-sibly triggered by remodeling or modificationof the RAG complex following DNA cleavage.

RAG2 NONCORE REGION

Like the noncore regions of RAG1, thenoncore region of RAG2, which constitutesthe last 140 aa of the protein, plays importantroles in V(D)J recombination efficiency andfidelity (discussed below under End Joiningand Postcleavage Complex Stability). Thisregion includes a portion of a flexible acidichinge, a PHD finger, and the T490 residuethat is important for directing degradationof RAG2 in the S, G2, and M phases of thecell cycle (Figure 3). The function of thevery C-terminal residues of RAG2 (residues500–527) is not well understood, although thelast seven residues of the protein contributeto phosphotidylinositol phosphate binding


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


(42) and decrease protein stability (42, 55).The PHD finger is discussed here; readers aredirected to a recent review for a more completediscussion of the RAG2 noncore region (73).

The RAG2 PHD finger (residues 414–487)has been shown to interact with two quite dis-tinct entities: phosphotidylinositol phosphate(42) and covalently modified forms of theN-terminal tail of histone H3, particularlyH3K4me3 (87, 91). Interaction with phospho-inositides was suggested to modulate an equilib-rium between permissive and inhibitory statesof the PHD finger (42), but the functional sig-nificance of phosphoinositide binding is notwell understood at this point.

The discovery that the RAG2 PHD fingerallows RAG to read the histone code repre-sented a paradigm shift for the field of V(D)Jrecombination by suggesting new pathways forcontrolling the targeting and activity of RAG(87, 91). The RAG2 PHD finger interactswith the N-terminal tail of histone 3 when ly-sine 4 is methylated, with the binding prefer-ence H3K4me3 � H3K4me2 > H3K4me1 >

H3K4me0. Binding to H3K4me3 (Kd ≈ 4 μM)is at least an order of magnitude stronger thanto H3K4me2 (Kd ≈ 60 μM) (91). Structuralanalysis revealed that the PHD finger forms anaromatic channel into which the K4me3 sidechain fits, with residues Y415, M443, and W453constituting critical components of the channel(91, 106). Mutation of any one of these residueseliminates RAG2 binding to a histone 3 N-terminal peptide containing K4me3 (hereafter,H3K4me3 peptide) and significantly reducesV(D)J recombination activity in cell based as-says (87, 91). An unusual feature of the RAG2PHD finger is that it can accommodate methy-lation of H3 arginine 2 (R2), and asymmetricdimethylation of R2 on the H3K4me3 peptideincreases the binding affinity somewhat (106).Curiously, asymmetric dimethylation of H3R2actually inhibits trimethylation of H3K4 (56,64). The relevance of H3R2 dimethylation forRAG2 binding in vivo or for V(D)J recombi-nation remains unknown. Interestingly, muta-tions of several PHD finger residues, includingW453, have been implicated in Omenn syn-

drome (Figure 3) (124). Overall, current dataindicate that the structural integrity of the PHDfinger as well as its ability to bind H3K4me3 arecritical for efficient V(D)J recombination (73).

How does the RAG2 PHD finger facilitatethe recombination process? Given the tight linkbetween H3K4me3 and active or poised RNApolymerase II (113), and between transcriptionand RSS accessibility (114), an appealing ideais that H3K4me3 recognition by RAG2 helpsrecruit RAG to the vicinity of RSSs, therebyfacilitating stable RSS binding (87, 91). Sev-eral pieces of evidence are consistent with thisidea. First, the RAG2 PHD finger can bind tochromatin containing methylated H3K4 (87).Second, ChIP sequencing shows a tight corre-lation between RAG2 binding and H3K4me3,both within antigen receptor loci and elsewherein the genome, arguing that the PHD finger-H3K4me3 interaction plays a dominant role indetermining RAG2 association with chromatin.Third, tethering of H3K4me3 peptide to anRSS oligonucleotide stimulated RSS binding byRAG only if RAG2 contained the PHD finger,demonstrating that the PHD finger-H3K4me3interaction can enhance RSS recognition in adefined biochemical system (118).

Another nonmutually exclusive mechanismis that H3K4me3 binding by RAG2 enhancesthe intrinsic activity of the RAG complex. Insupport of this contention, PHD finger pointmutations that eliminate H3K4me3 binding(RAG2-PHDmut) reduce V(D)J recombinationactivity more than deletion of the entire RAG2C-terminal region (RAG2 core) (87, 91, 106).These findings suggest that the PHD finger-H3K4me3 peptide interaction relieves an in-hibitory effect of the RAG2 C terminus, per-haps by triggering a conformational change inRAG2. A clue to how this might occur was re-vealed by structural analysis of the PHD fin-ger without bound H3 peptide: Under theseconditions, the binding channel contained ashort peptide derived from expression vectorsequences, with a proline residue occupying theposition of the trimethyl-ammonium group ofK4me3 (106). This led to the proposal that,in the absence of H3K4me3, RAG2 is in a


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


repressed configuration caused by occupancyof the PHD finger binding channel by a self-peptide of RAG2 (perhaps a proline-containingflexible loop from the RAG2 core); bindingof H3K4me3 peptide would displace the self-peptide and relieve this inhibition (106). How-ever, for this model to explain the poor activityof RAG2-PHDmut proteins, one must assumethat the PHD finger mutations prevent bind-ing of H3K4me peptide but still allow bindingof the putative RAG2 self-peptide.

Two recent biochemical studies lend fur-ther support for the direct stimulation modelby demonstrating that binding of H3K4me3peptide enhances RAG catalytic activity in amanner that is dependent on an intact PHDfinger (55, 118). The peptide need not beattached to the RSS DNA to exert its effect,arguing that stimulation occurs by a mecha-nism other than simple recruitment of RAGto the DNA. H3K4me3 peptide was found tostimulate RAG-RSS complex formation [theSC (118) and the PC (55)], nicking (118), andhairpin formation (55, 118). Grundy et al. (55)also demonstrated that the RAG2 C-terminalregion (excluding only residues 521–527) leadsto a suppression of hairpin formation that isstrongly dependent on the presence of thelast 32 residues of RAG1 (residues 1009–1040). Furthermore, they found that addingH3K4me3 peptide relieves this inhibitionin a manner that is again far stronger whenthe RAG1 C-terminal residues are present.Grundy et al. (55) proposed that the RAG1and RAG2 C-terminal regions collaborate tolimit RSS substrate flexibility to inhibit a DNAconformational change required for hairpinformation and that H3K4me3 peptide bindingby the PHD finger relieves this inhibition.Shimazaki et al. (118), however, observed RAGstimulation by H3K4me3 peptide using coreRAG1, which lacks the C-terminal residues.The discrepancy might be explained by theuse of different RAG2 proteins (1–527 byShimazaki et al. and 1–520 by Grundy et al.)and by the fact that Grundy et al. did notinclude HMGB1 in their DNA binding orcleavage experiments, noting that the effects

they observed were reduced in the presenceof HMGB1. There is also some disagreementas to whether the RAG2 C-terminal region isinhibitory: Some (55, 80, 109, 135), but notall (118), studies find that full-length or nearlyfull-length RAG2 is less active than core RAG2in supporting DNA cleavage in vitro. Interest-ingly, the former studies used RAG1 containingresidues 1009–1040, whereas the latter did not.

Overall, current evidence indicates that theRAG2 C-terminal domain and an intact PHDfinger are important for proper RAG functionin vivo, that the C-terminal domain can exertan inhibitory effect on DNA cleavage in vitro,that this is manifest at both the DNA bind-ing and cleavage steps, and that inhibition bythe RAG2 C-terminal region is most apparentwhen activity is assayed using RAG1 that con-tains its extreme C-terminal residues. Bindingof the RAG2 PHD finger to H3K4me3 appearsto stimulate V(D)J recombination by at leasttwo mechanisms: enhanced RAG recruitmentto the RSS and stimulation of DNA cleavage.The stimulation of DNA cleavage is mediatedby effects on both RSS binding and catalysis,and is likely due at least in part to a reversal ofthe repressive effects of the RAG2 C-terminaldomain. Important challenges for the future in-clude determining the full range of functionsexerted by the RAG2 PHD finger and whetherconformational changes underlie the stimula-tory effects of engagement of H3K4me3 by theRAG2 PHD finger.

CLEAVAGE-TO-JOININGTRANSITION

DNA DSBs are dangerous lesions with the ca-pacity to lead to genome rearrangements suchas chromosomal translocations. Lymphoid tu-mors frequently harbor chromosomal translo-cations involving antigen receptor loci, and asubset of these translocations are due to er-rors made during the process of V(D)J recom-bination (12). Given the very large number ofV(D)J recombination events that occur eachday in a normal human, it is exceptionally im-portant that RAG-generated DSBs be repaired


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


efficiently and with high fidelity and that thereexist surveillance systems to eliminate the rarecells in which repair goes awry. This is accom-plished through the combined action of numer-ous DNA repair and damage response proteinsand pathways. Here, we consider this only fromthe perspective of the RAG proteins and theirrole in orchestrating proper repair of the DSBsthey generate. The reader is also directed to anexcellent recent review of this topic (19).

RAG AND THE CHOICEOF A REPAIR PATHWAY

As noted above, repair during V(D)J recombi-nation is mediated by the classical NHEJ path-way. A defect in any classical NHEJ repair fac-tor virtually eliminates V(D)J recombination,demonstrating that no other repair pathwaycan compensate for the absence of NHEJ dur-ing V(D)J recombination. This is strikingly dif-ferent from class switch recombination, wherean alternative NHEJ pathway is able to gener-ate products relatively efficiently in cells lack-ing one or more classical NHEJ factors (17and references therein). What prevents RAG-generated ends from efficiently accessing repairpathways other than classical NHEJ?

Accumulating evidence strongly arguesthat the answer lies in the RAG proteinsthemselves and the postcleavage complex(es)they form with the coding and signal ends.Initial evidence for this came from the iden-tification of RAG mutants, many by the Rothlaboratory, that were able to cleave reasonablyefficiently but were defective in V(D)J recom-bination, presumably due to a defect in joining(Figure 3) (19). Subsequent analysis revealedthat two of these mutants (RAG1 K980Aand RAG1 R838A/K839A/R840A) yieldedelevated levels of repair of signal ends byhomologous recombination and coding jointswith an increased frequency of short sequencehomologies and excessive deletions, hall-marks of alternative NHEJ (62, 83). Hence,mutations in RAG1 can allow other repairpathways to gain access to signal and codingends, supporting the idea that RAG guides the

cleaved ends it generates specifically into theclassical NHEJ pathway (83). Strong supportfor this idea came from subsequent studiesfrom the Roth and Meek groups (24, 26). Inone, it was revealed that a frameshift mutationof RAG2 at residue 361 (RAG2–361fs) allowedfor increased access of ends to both the homol-ogous repair and alternative NHEJ pathwayscompared to wild-type RAG2 (24). In addition,RAG2–361fs allowed for remarkably efficientsignal and coding joint formation in cellslacking classical NHEJ repair factors; sequenceanalysis of these products showed an increasedfrequency of short sequence homologies andlarge deletions. The alternative NHEJ pathwayis therefore quite robust and able to processand join RAG-generated ends, but is preventedfrom doing so in the context of the wild-typeRAG proteins (24).

In the second study, joining of ends gener-ated by the I-Sce1 endonuclease was assessedon transiently transfected plasmid substratescontaining two I-Sce1 sites (26). It was firstdemonstrated that I-Sce1 ends were rejoinedefficiently in classical NHEJ-deficient cells,demonstrating ready access of these ends to thealternative NHEJ pathway. However, whencleavage at the I-Sce1 sites was performed bya RAG2core-I-Sce1 fusion protein in the pres-ence of core, full-length, or catalytically inactiveRAG1, rejoining became strongly dependenton an intact classical NHEJ repair pathway (theeffect was significantly weaker in the absenceof coexpressed RAG1) (26). Interestingly, thisstrong dependency on classical NHEJ in thepresence of RAG was not seen on a substratecontaining only a single I-Sce1 site. Hence,tethering of the RAG complex to a heterolo-gous endonuclease can prevent efficient end re-pair by any pathway other than classical NHEJ,and this effect is seen only at a pair of breaks(with four DNA ends), not at a single break (twoends) (26). Together, the data argue stronglythat it is the RAG proteins themselves thatprevent alternative NHEJ from participating inthe repair phase of V(D)J recombination. Giventhe apparent propensity of alternative NHEJto yield aberrant joining products, including


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


chromosomal translocations, this might play asignificant role in ensuring the high fidelity ofthe V(D)J recombination reaction (19).

END JOINING ANDPOSTCLEAVAGE COMPLEXSTABILITY

It is appealing to think that the ability ofRAG to channel ends into the classical NHEJrepair pathway depends on the stability ofpostcleavage complexes (CSC, SEC, and thehypothetical complex containing the twocoding ends) (Figure 2) and that prematureend-release could result in the use of alternaterepair pathways (19). The RAG1 K980Aand RAG1 R838A/K839A/R840A mutantsdiscussed above provide evidence for this:Their defect in constraining repair to theNHEJ pathway correlates with a severe defectin the stability of the SEC (83). Unfortunately,the weak association of coding ends with theCSC in vitro (2, 61) makes it difficult to assessthe effect of RAG mutations on their retentionin the complex. In addition, a recent study byDeriano et al. (33) reports that core RAG2(residues 1–383) exhibits a defect in SECstability and that core RAG2 knockin micedevelop aggressive thymic lymphomas on ap53−/− background. The lymphomas that arisein core RAG2 × p53−/− mice are lethal muchmore quickly than those from p53−/− mice(mean survival approximately 12 weeks versusapproximately 23 weeks) and contain recurrenttranslocations involving the Tcra/Tcrd andIgh loci, which are not observed in thymiclymphomas from p53−/− mice. Given thatcore RAG2 allows for low level bypass of therequirement for classical NHEJ (24), it is plau-sible that the translocations in core RAG2 ×p53−/− thymic lymphomas are due to a destabi-lized postcleavage complex that allows releasedends to engage alternative repair pathways (33).

Further support for this idea comes fromthe analysis of the ATM (ataxia telangiectasiamutated) protein, an important DNA damageresponse factor and tumor suppressor. ATMis important for the stability of one or more

RAG postcleavage complexes in vivo (20), andATM-deficiency predisposes to the develop-ment of thymic lymphomas in both mice andhumans (117). Interestingly, the translocationsand lymphomas seen in RAG2 × p53−/− miceclosely resemble those of Atm−/− mice (33).Hence, it is plausible that RAG-dependentgenome instability arises from similar mech-anisms in core RAG2 × p53−/− and Atm−/−

mice, and that the common denominatoris an unstable postcleavage complex. Thisidea is strengthened further by the recentcharacterization of a postcleavage defectduring V(D)J recombination in cells lackingeither histone H2AX or MDC1 (mediator ofDNA damage checkpoint 1), two proteins inthe ATM DNA damage response pathway.Artemis is the only nuclease normally ableto process RAG-generated hairpin codingends, but in the absence of H2AX or MDC1,a CtIP-dependent end resection pathway isable to gain access to coding ends (58). As aresult, the ends suffer significant deletions andengage the alternative NHEJ repair pathway,leading to the model that ATM, H2AX, andMDC1 are important for genome stabilityduring V(D)J recombination by restricting theaccess of aberrant repair pathways to the codingends, particularly in cases where Artemis failsto open quickly the coding end hairpins (58).

It is important to recognize, however, thatdeletion of the entire RAG2 C-terminal regionremoves multiple important regulatory func-tions and perturbs more than just the stabil-ity of RAG postcleavage complexes. Hence,the phenotypes observed by Deriano et al. incore RAG2 × p53−/− mice might be influ-enced by altered RAG cell cycle regulation, tar-geting, activity, or cleavage site selection (29,73, 151, and references therein) in addition todecreased postcleavage complex stability (33).Zhang et al. (151) recently revealed the impor-tance of proper cell cycle regulation of RAGfor the fidelity of V(D)J recombination throughthe creation and analysis of mice expressingRAG2 T490A, which cannot be destabilizedin S-G2-M phase cells. As expected, thesemice express RAG2 and accumulate signal end


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


intermediates in S-G2-M phase thymocytes,unlike thymocytes expressing wild-type RAG2.Importantly, RAG2 T490A × p53−/− and coreRAG2 × p53−/− mice exhibit overlapping butdistinct tumor phenotypes. Both genotypes de-velop thymic lymphomas containing translo-cations involving antigen receptor genes, butRAG2 T490A × p53−/− mice do not diemore rapidly than p53−/− mice, whereas coreRAG2 × p53−/− mice do (33, 151). Hence, theaccelerated development of thymic lymphomain core RAG2 × p53−/− mice cannot be ex-plained by the loss of RAG cell cycle regula-tion, and can plausibly be attributed, at leastin part, to destabilization of RAG postcleavagecomplexes.

A further link between postcleavage com-plex stability and strict channeling of V(D)Jends into the classical NHEJ repair pathwaycomes from the finding that certain alterationsin the sequence of heptamer positions 4 or 5destabilize the SEC in vitro and allow signaland coding ends to participate in alternative re-pair pathways in vivo (9). This is consistent withan earlier study that found that sequence vari-ation in the RSS can influence the efficiency ofcoding joint formation at a postcleavage stepof V(D)J recombination (1). Together, the twostudies suggest that cleavage at cryptic RSSs oreven certain bona fide RSSs can lead to inef-ficient coding end processing and breakdownof the postcleavage complex, leading either toapoptotic cell death or resolution by an alter-nate DNA repair pathway, which perhaps leadsto a chromosomal translocation (1, 9).

Taken together, current findings suggestthat stabilization of postcleavage complexes byRAG is important for proper channeling of endsinto the classical NHEJ repair pathway and forgenome stability, and that other factors (e.g.,ATM, H2AX, MDC1) collaborate with RAGto achieve stability of postcleavage complexesand proper protection of the DNA ends. In ad-dition, exclusion of RAG-generated ends fromthe S-G2 phases of the cell cycle, when homolo-gous recombination is most active, is importantfor the prevention of RAG-induced genomeinstability.

MAKING THE CONNECTION TOCLASSICAL NONHOMOLOGOUSEND JOINING

We currently have little insight into themechanisms by which the RAG proteinsdirect coding and signal ends into the classicalNHEJ repair pathway, and this is likely tobe an area of active investigation in comingyears. Several distinct types of nonmutuallyexclusive models can be envisioned. (a) TheRAG proteins (or proteins they recruit) mightexclude or inhibit the action of other repairpathways; (b) RAG might direct recombinationto a compartment of the nucleus or a phaseof the cell cycle in which classical NHEJ isfavored; and (c) perhaps most appealingly,RAG might recruit classical NHEJ factorsto the postcleavage complex. A recent studyprovides evidence for this third possibility byrevealing an interaction between full-lengthRAG1 and Ku70/Ku80 (109). Under gentleconditions, the Ku complex copurifies withfull-length RAG1/core RAG2 in a DNA-independent manner, and using such a proteinpreparation, a novel Ku-RAG-RSS complexcan be detected by EMSA. Detection of thiscomplex minimally requires residues 211–383of the RAG1 N-terminal region. Ku does notdetectably alter the DNA cleavage propertiesof RAG, suggesting a postcleavage functionfor the interaction. The authors suggest thatby binding Ku, RAG1 helps coordinate thedelivery of Ku to signal ends when the SECis disassembled during signal joint formation(109). It is conceivable that the presence ofKu might also help stabilize postcleavagecomplexes. An interaction between RAG1and MDC1 that involves determinants inthe RAG1 N-terminal region as well as theextreme C terminus has also been detected(G. Coster, D. Chen, M. Goldberg, and D.G.Schatz, unpublished results). Although it is verytempting to think that RAG-mediated recruit-ment of repair factors after, or even before,DNA cleavage is important for the efficiencyand fidelity of V(D)J recombination, definitiveevidence for this has yet to be obtained.


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


SUMMARY POINTS

1. V(D)J recombination is thought to be initiated through the stepwise binding of one RSSby the RAG and HMGB1 proteins, followed by capture of an appropriate partner RSS.Both binding steps likely involve iterative sampling of DNAs to enable the RAG complexto bind, survey, and release suboptimal heptamer, spacer, and nonamer sequences.

2. The RAG proteins mediate RSS cleavage through a nick-hairpin mechanism. The RAGproteins may or may not nick the first RSS upon binding it, but hairpin formation gen-erally occurs only upon synapsis, and is likely triggered through a mechanism involvingbase flipping at C1b (Figure 1a and Figure 4).

3. The order in which the RSSs are assembled into a synaptic complex, the choice of whichRSSs are ultimately paired together, and the likelihood that an RSS is nicked prior tosynapsis are probably determined by a combination of factors, including the sequenceof the RSS itself, the accessibility of the locus in which it resides, and the potential forRAG-interacting proteins to recruit or stabilize RAG complexes at particular RSSs.

4. RAG1 plays a primary role in RSS recognition and catalyzing DNA cleavage. In addi-tion, its noncore N terminus might promote the cleavage-to-joining transition throughassociation with Ku70/Ku80, or by mediating ubiquitylation of KPNA1 or histone H3to influence RAG1 subcellular localization or to facilitate detection or repair of DNAbreaks, respectively.

5. The RAG1 noncore C terminus might help suppress RAG complex activity throughinteractions with the RAG2 PHD finger. This suppression might be partly relievedupon binding of full-length RAG2 to H3K4me3.

6. RAG2 collaborates with RAG1 to mediate DNA contact near the cleavage site. De-terminants around RAG2 residue 360 help stabilize postcleavage complexes to preventalternative repair mechanisms from accessing RAG-mediated DNA breaks. The RAG2PHD finger mediates recognition of H3K4me3, which might promote RAG recruitmentto active loci and stimulate RAG DNA binding and cleavage activity.

7. The RAG proteins act as guardians of the genome and as tumor suppressors, probablyby multiple mechanisms. These include the stabilization of postcleavage complexes, thechanneling of coding and signal ends into the classical NHEJ repair pathway, and therestriction of RAG-generated broken ends to the G1/G0 phases of the cell cycle.

FUTURE ISSUES

1. How does RAG2 contribute to DNA binding and cleavage by the RAG complex? Doesit cause a conformation change in RAG1?

2. What is the mechanistic basis of the 12/23 rule?

3. How does nicking at one RSS enable hairpin formation at the other, and how do the twoactive sites communicate with one another? How often does nicking occur outside of thecontext of the PC in vivo?


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


4. What are the stoichiometry and structure of the SC, and how does this structure changewhen a second RSS is captured to form the PC? How structurally dynamic are the SCand PC? How does the PC change configuration after cleavage?

5. How does binding of the RAG2 PHD finger to H3K4me3 increase RAG catalytic ac-tivity? What is the mechanistic basis of the inhibition mediated by the RAG2 noncoreregion?

6. What are the functionally important targets of RAG1 E3 ubiquitin ligase activity andhow does ubiquitylation contribute to V(D)J recombination?

7. By what mechanisms does RAG ensure that coding and signal ends are joined by theclassical NHEJ repair pathway? How does RAG collaborate with other factors to stabilizepostcleavage complexes during V(D)J recombination?

8. What are the critical parameters that determine whether a RAG-RSS interaction willlead to a stable SC or whether partner capture will lead to a stable PC? How stable arethe SC and PC in vivo?

DISCLOSURE STATEMENT

The authors are not aware of any affiliations, memberships, funding, or financial holdings thatmight be perceived as affecting the objectivity of this review.

ACKNOWLEDGMENTS

The authors apologize that not all of the relevant literature could be cited due to limitations onthe number of references. This work was supported in part by NIH grants R37 AI32524 to D.G.S.and 3R56 AI055599-06A1S1 to P.C.S. D.G.S. is an investigator of the Howard Hughes MedicalInstitute.

LITERATURE CITED

1. Agard EA, Lewis SM. 2000. Postcleavage sequence specificity in V(D)J recombination. Mol. Cell. Biol.20:5032–40

2. Agrawal A, Schatz DG. 1997. RAG1 and RAG2 form a stable postcleavage synaptic complex with DNAcontaining signal ends in V(D)J recombination. Cell 89:43–53

3. Aidinis V, Bonaldi T, Beltrame M, Santagata S, Bianchi ME, Spanopoulou E. 1999. The RAG1 homeo-domain recruits HMG1 and HMG2 to facilitate recombination signal sequence binding and to enhancethe intrinsic DNA-bending activity of RAG1-RAG2. Mol. Cell. Biol. 19:6532–42

4. Aidinis V, Dias DC, Gomez CA, Bhattacharyya D, Spanopoulou E, Santagata S. 2000. Definition ofminimal domains of interaction within the recombination-activating genes 1 and 2 recombinase complex.J. Immunol. 164:5826–32

5. Akamatsu Y, Monroe R, Dudley DD, Elkin SK, Gartner F, et al. 2003. Deletion of the RAG2 C terminusleads to impaired lymphoid development in mice. Proc. Natl. Acad. Sci. USA 100:1209–14

6. Akamatsu Y, Oettinger MA. 1998. Distinct roles of RAG1 and RAG2 in binding the V(D)J recombinationsignal sequences. Mol. Cell. Biol. 18:4670–78

7. Al-Hakim A, Escribano-Diaz C, Landry MC, O’Donnell L, Panier S, et al. 2010. The ubiquitous roleof ubiquitin in the DNA damage response. DNA Repair (Amst.) 9:1229–40

8. Arbuckle JL, Fauss LJ, Simpson R, Ptaszek LM, Rodgers KK. 2001. Identification of two topologi-cally independent domains in RAG1 and their role in macromolecular interactions relevant to V(D)Jrecombination. J. Biol. Chem. 276:37093–101


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


9. Arnal SM, Holub AJ, Salus SS, Roth DB. 2010. Non-consensus heptamer sequences destabilize theRAG post-cleavage complex, making ends available to alternative DNA repair pathways. Nucleic AcidsRes. 38:2944–54

10. Bailin T, Mo X, Sadofsky MJ. 1999. A RAG1 and RAG2 tetramer complex is active in cleavage in V(D)Jrecombination. Mol. Cell. Biol. 19:4664–71

11. Bassing CH, Alt FW, Hughes MM, D’Auteuil M, Wehrly TD, et al. 2000. Recombination signal se-quences restrict chromosomal V(D)J recombination beyond the 12/23 rule. Nature 405:583–86

12. Bassing CH, Swat W, Alt FW. 2002. The mechanism and regulation of chromosomal V(D)J recombi-nation. Cell 109:S45–S55

13. Bassing CH, Whitlow S, Mostoslovasky R, Yang-Iott K, Ranganath S, Alt FW. 2008. Vβ cluster se-quences reduce the frequency of primary Vβ2 and Vβ14 rearrangements. Eur. J. Immunol. 38:2564–72

14. Bischerour J, Chalmers R. 2007. Base-flipping dynamics in a DNA hairpin processing reaction. NucleicAcids Res. 35:2584–95

15. Bischerour J, Chalmers R. 2009. Base flipping in tn10 transposition: an active flip and capture mechanism.PLoS ONE 4:e6201

16. Bischerour J, Lu C, Roth DB, Chalmers R. 2009. Base flipping in V(D)J recombination: insights intothe mechanism of hairpin formation, the 12/23 rule, and the coordination of double-strand breaks. Mol.Cell. Biol. 29:5889–99

17. Boboila C, Yan C, Wesemann DR, Jankovic M, Wang JH, et al. 2010. Alternative end-joining catalyzesclass switch recombination in the absence of both Ku70 and DNA ligase 4. J. Exp. Med. 207:417–27

18. Bonnet M, Ferrier P, Spicuglia S. 2009. Molecular genetics at the T-cell receptor beta locus: insightsinto the regulation of V(D)J recombination. Adv. Exp. Med. Biol. 650:116–32

19. Brandt VL, Roth DB. 2009. Recent insights into the formation of RAG-induced chromosomal translo-cations. Adv. Exp. Med. Biol. 650:32–45

20. Bredemeyer AL, Sharma GG, Huang CY, Helmink BA, Walker LM, et al. 2006. ATM stabilizes DNAdouble-strand-break complexes during V(D)J recombination. Nature 442:466–70

21. Callebaut I, Mornon JP. 1998. The V(D)J recombination activating protein RAG2 consists of a six-bladedpropeller and a PHD fingerlike domain, as revealed by sequence analysis. Cell. Mol. Life Sci. 54:880–91

22. Ciubotaru M, Kriatchko AN, Swanson PC, Bright FV, Schatz DG. 2007. Fluorescence resonance energytransfer analysis of recombination signal sequence configuration in the RAG1/2 synaptic complex. Mol.Cell. Biol. 27:4745–58

23. Ciubotaru M, Ptaszek LM, Baker GA, Baker SN, Bright FV, Schatz DG. 2003. RAG1-DNA binding inV(D)J recombination. Specificity and DNA-induced conformational changes revealed by fluorescenceand CD spectroscopy. J. Biol. Chem. 278:5584–96

24. Corneo B, Wendland RL, Deriano L, Cui X, Klein IA, et al. 2007. Rag mutations reveal robust alternativeend joining. Nature 449:483–86

25. Cortes P, Ye ZS, Baltimore D. 1994. RAG-1 interacts with the repeated amino acid motif of the humanhomologue of the yeast protein SRP1. Proc. Natl. Acad. Sci. USA 91:7633–37

26. Cui X, Meek K. 2007. Linking double-stranded DNA breaks to the recombination activating genecomplex directs repair to the nonhomologous end-joining pathway. Proc. Natl. Acad. Sci. USA 104:17046–51

27. Cuomo CA, Mundy CL, Oettinger MA. 1996. DNA sequence and structure requirements for cleavageof V(D)J recombination signal sequences. Mol. Cell. Biol. 16:5683–90

28. Curry JD, Geier JK, Schlissel MS. 2005. Single-strand recombination signal sequence nicks in vivo:evidence for a capture model of synapsis. Nat. Immunol. 6:1272–79

29. Curry JD, Schlissel MS. 2008. RAG2’s non-core domain contributes to the ordered regulation of V(D)Jrecombination. Nucleic Acids Res. 36:5750–62

30. Davies DR, Goryshin IY, Reznikoff WS, Rayment I. 2000. Three-dimensional structure of the Tn5synaptic complex transposition intermediate. Science 289:77–85

31. De P, Rodgers KK. 2004. Putting the pieces together: identification and characterization of structuraldomains in the V(D)J recombination protein RAG1. Immunol. Rev. 200:70–82

32. De P, Zhao S, Gwyn LM, Godderz LJ, Peak MM, Rodgers KK. 2008. Thermal dependency of RAG1self-association properties. BMC Biochem. 9:5


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


33. Deriano L, Chaumeil J, Coussens M, Multani A, Chou Y, et al. 2011. The RAG2 C terminus suppressesgenomic instability and lymphomagenesis. Nature 471:119–23

34. Deshaies RJ, Joazeiro CA. 2009. RING domain E3 ubiquitin ligases. Annu. Rev. Biochem. 78:399–43435. Desiderio S, Lee J. 2000. Signaling pathways that control V(D)J recombination. Curr. Top. Microbiol.

Immun. 245:31–5636. Difilippantonio MJ, McMahan CJ, Eastman QM, Spanopoulou E, Schatz DG. 1996. RAG1 mediates

signal sequence recognition and recruitment of RAG2 in V(D)J recombination. Cell 87:253–6237. Drejer-Teel AH, Fugmann SD, Schatz DG. 2007. The beyond 12/23 restriction is imposed at the nicking

and pairing steps of DNA cleavage during V(D)J recombination. Mol. Cell. Biol. 27:6288–9938. Dudley DD, Sekiguchi J, Zhu CM, Sadofsky MJ, Whitlow S, et al. 2003. Impaired V(D)J recombination

and lymphocyte development in core RAG1-expressing mice. J. Exp. Med. 198:1439–5039. Eastman QM, Leu TM, Schatz DG. 1996. Initiation of V(D)J recombination in vitro obeying the 12/23

rule. Nature 380:85–8840. Eastman QM, Schatz DG. 1997. Nicking is asynchronous and stimulated by synapsis in 12/23 rule-

regulated V(D)J cleavage. Nucl. Acids Res. 25:4370–7841. Eastman QM, Villey IJ, Schatz DG. 1999. Detection of RAG protein-V(D)J recombination signal in-

teractions near the site of DNA cleavage by UV cross-linking. Mol. Cell. Biol. 19:3788–9742. Elkin SK, Ivanov D, Ewalt M, Ferguson CG, Hyberts SG, et al. 2005. A PHD finger motif in the C

terminus of RAG2 modulates recombination activity. J. Biol. Chem. 280:28701–1043. Elkin SK, Matthews AG, Oettinger MA. 2003. The C-terminal portion of RAG2 protects against trans-

position in vitro. EMBO J. 22:1931–3844. Feeney AJ, Goebel P, Espinoza CR. 2004. Many levels of control of V gene rearrangement frequency.

Immunol. Rev. 200:44–5645. Franchini DM, Benoukraf T, Jaeger S, Ferrier P, Payet-Bornet D. 2009. Initiation of V(D)J recombina-

tion by Dβ-associated recombination signal sequences: a critical control point in TCRβgene assembly.PLoS ONE 4:e4575

46. Fugmann SD. 2010. The origins of the Rag genes: from transposition to V(D)J recombination. Semin.Immunol. 22:10–16

47. Fugmann SD, Lee AI, Shockett PE, Villey IJ, Schatz DG. 2000. The RAG proteins and V(D)J recom-bination: complexes, ends, and transposition. Annu. Rev. Immunol. 18:495–527

48. Fugmann SD, Schatz DG. 2001. Identification of basic residues in RAG2 critical for DNA binding bythe RAG1-RAG2 complex. Mol. Cell 8:899–910

49. Fugmann SD, Villey IJ, Ptaszek LM, Schatz DG. 2000. Identification of two catalytic residues in RAG1that define a single active site within the RAG1/RAG2 protein complex. Mol. Cell 5:97–107

50. Fuxa M, Skok J, Souabni A, Salvagiotto G, Roldan E, Busslinger M. 2004. Pax5 induces V-to-DJ rear-rangements and locus contraction of the immunoglobulin heavy-chain gene. Genes Dev. 18:411–22

51. Gellert M. 2002. V(D)J recombination: RAG proteins, repair factors, and regulation. Annu. Rev. Biochem.71:101–32

52. Grazini U, Zanardi F, Citterio E, Casola S, Goding CR, McBlane F. 2010. The RING domain of RAG1ubiquitylates histone H3: a novel activity in chromatin-mediated regulation of V(D)J joining. Mol. Cell37:282–93

53. Grundy GJ, Hesse JE, Gellert M. 2007. Requirements for DNA hairpin formation by RAG1/2. Proc.Natl. Acad. Sci. USA 104:3078–83

54. Grundy GJ, Ramon-Maiques S, Dimitriadis EK, Kotova S, Biertumpfel C, et al. 2009. Initial stages ofV(D)J recombination: the organization of RAG1/2 and RSS DNA in the postcleavage complex. Mol. Cell35:217–27

55. Grundy GJ, Yang W, Gellert M. 2011. Autoinhibition of DNA cleavage mediated by RAG1 and RAG2is overcome by an epigenetic signal in V(D)J recombination. Proc. Natl. Acad. Sci. USA 107:22487–92

56. Guccione E, Bassi C, Casadio F, Martinato F, Cesaroni M, et al. 2007. Methylation of histone H3R2 byPRMT6 and H3K4 by an MLL complex are mutually exclusive. Nature 449:933–37

57. Gwyn LM, Peak MM, De P, Rahman NS, Rodgers KK. 2009. A zinc site in the C-terminal domain ofRAG1 is essential for DNA cleavage activity. J. Mol. Biol. 390:863–78


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


58. Helmink BA, Tubbs AT, Dorsett Y, Bednarski JJ, Walker LM, et al. 2010. H2AX prevents CtIP-mediatedDNA end resection and aberrant repair in G1-phase lymphocytes. Nature 469:245–49

59. Hesse JE, Lieber MR, Mizuuchi K, Gellert M. 1989. V(D)J recombination: a functional definition ofthe joining signals. Genes Dev. 3:1053–61

60. Hiom K, Gellert M. 1997. A stable RAG1-RAG2-DNA complex that is active in V(D)J cleavage. Cell88:65–72

61. Hiom K, Gellert M. 1998. Assembly of a 12/23 paired signal complex: a critical control point in V(D)Jrecombination. Mol. Cell 1:1011–19

62. Huye LE, Purugganan MM, Jiang MM, Roth DB. 2002. Mutational analysis of all conserved basic aminoacids in RAG-1 reveals catalytic, step arrest, and joining-deficient mutants in the V(D)J recombinase.Mol. Cell. Biol. 22:3460–73

63. Huye LE, Roth DB. 2000. Differential requirements for cis and trans V(D)J cleavage: effects of substratelength. Nucleic Acids Res. 28:4903–11

64. Hyllus D, Stein C, Schnabel K, Schiltz E, Imhof A, et al. 2007. PRMT6-mediated methylation of R2 inhistone H3 antagonizes H3 K4 trimethylation. Genes Dev. 21:3369–80

65. Jhunjhunwala S, van Zelm MC, Peak MM, Murre C. 2009. Chromatin architecture and the generationof antigen receptor diversity. Cell 138:435–48

66. Ji Y, Little AJ, Banerjee JK, Hao B, Oltz EM, et al. 2010. Promoters, enhancers, and transcription targetRAG1 binding during V(D)J recombination. J. Exp. Med. 207:2809–16

67. Ji Y, Resch W, Corbett E, Yamane A, Casellas R, Schatz DG. 2010. The in vivo pattern of binding ofRAG1 and RAG2 to antigen receptor loci. Cell 141:419–31

68. Jiang H, Chang FC, Ross AE, Lee JY, Nakayama K, Desiderio S. 2005. Ubiquitylation of RAG-2 bySkp2-SCF links destruction of the V(D)J recombinase to the cell cycle. Mol. Cell 18:699–709

69. Jones JM, Bhattacharyya A, Simkus C, Vallieres B, Veenstra TD, Zhou M. 2011. The RAG1 V(D)J re-combinase/ubiquitin ligase promotes ubiquitylation of acetylated, phosphorylated histone 3.3. Immunol.Lett. 136:156–62

70. Jones JM, Gellert M. 2001. Intermediates in V(D)J recombination: a stable RAG1/2 complex sequesterscleaved RSS ends. Proc. Natl. Acad. Sci. USA 98:12926–31

71. Jones JM, Gellert M. 2002. Ordered assembly of the V(D)J synaptic complex ensures accurate recombi-nation. EMBO J. 21:4162–71

72. Jones JM, Gellert M. 2003. Autoubiquitylation of the V(D)J recombinase protein RAG1. Proc. Natl.Acad. Sci. USA 100:15446–51

73. Jones JM, Simkus C. 2009. The roles of the RAG1 and RAG2 “non-core” regions in V(D)J recombinationand lymphocyte development. Arch. Immunol. Ther. Exp. (Warsz) 57:105–16

74. Jung D, Bassing CH, Fugmann SD, Cheng HL, Schatz DG, Alt FW. 2003. Extrachromosomal re-combination substrates recapitulate beyond 12/23 restricted V(D)J recombination in nonlymphoid cells.Immunity 18:65–74

75. Kale SB, Landree MA, Roth DB. 2001. Conditional RAG-1 mutants block the hairpin formation stepof V(D)J recombination. Mol. Cell. Biol. 21:459–66

76. Kim DR, Dai Y, Mundy CL, Yang W, Oettinger MA. 1999. Mutations of acidic residues in RAG1 definethe active site of the V(D)J recombinase. Genes Dev. 13:3070–80

77. Ko JE, Kim CW, Kim DR. 2004. Amino acid residues in RAG1 responsible for the interaction withRAG2 during the V(D)J recombination process. J. Biol. Chem. 279:7715–20

78. Kriatchko AN, Anderson DK, Swanson PC. 2006. Identification and characterization of a gain-of-function RAG-1 mutant. Mol. Cell. Biol. 26:4712–28

79. Kriatchko AN, Bergeron S, Swanson PC. 2008. HMG-box domain stimulation of RAG1/2 cleavageactivity is metal ion dependent. BMC Mol. Biol. 9:32

80. Kumar S, Swanson PC. 2009. Full-length RAG1 promotes contact with coding and intersignal sequencesin RAG protein complexes bound to recombination signals paired in cis. Nucleic Acids Res. 37:2211–26

81. Landree MA, Kale SB, Roth DB. 2001. Functional organization of single and paired V(D)J cleavagecomplexes. Mol. Cell. Biol. 21:4256–64


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


82. Landree MA, Wibbenmeyer JA, Roth DB. 1999. Mutational analysis of RAG1 and RAG2 identifiesthree catalytic amino acids in RAG1 critical for both cleavage steps of V(D)J recombination. Genes Dev.13:3059–69

83. Lee GS, Neiditch MB, Salus SS, Roth DB. 2004. RAG proteins shepherd double-strand breaks to aspecific pathway, suppressing error-prone repair, but RAG nicking initiates homologous recombination.Cell 117:171–84

84. Lewis SM. 1994. The mechanism of V(D)J joining: lessons from molecular, immunological, and com-parative analyses. Adv. Immunol. 56:27–150

85. Liang HE, Hsu LY, Cado D, Cowell LG, Kelsoe G, Schlissel MS. 2002. The “dispensable” portionof RAG2 is necessary for efficient V-to-DJ rearrangement during B and T cell development. Immunity17:639–51

86. Lieber MR. 2008. The mechanism of human nonhomologous DNA end joining. J. Biol. Chem. 283:1–587. Liu Y, Subrahmanyam R, Chakraborty T, Sen R, Desiderio S. 2007. A plant homeodomain in RAG-

2 that binds hypermethylated lysine 4 of histone H3 is necessary for efficient antigen-receptor-generearrangement. Immunity 27:561–71

88. Lu CP, Sandoval H, Brandt VL, Rice PA, Roth DB. 2006. Amino acid residues in Rag1 crucial for DNAhairpin formation. Nat. Struct. Mol. Biol. 13:1010–15

89. Ma YM, Pannicke U, Schwarz K, Lieber MR. 2002. Hairpin opening and overhang processing by anArtemis/DNA-dependent protein kinase complex in nonhomologous end joining and V(D)J recombi-nation. Cell 108:781–94

90. Maitra R, Sadofsky MJ. 2009. A WW-like module in the RAG1 N-terminal domain contributes topreviously unidentified protein-protein interactions. Nucl. Acids Res. 37:3301–9

91. Matthews AG, Kuo AJ, Ramon-Maiques S, Han S, Champagne KS, et al. 2007. RAG2 PHD fingercouples histone H3 lysine 4 trimethylation with V(D)J recombination. Nature 450:1106–10

92. McBlane JF, van Gent DC, Ramsden DA, Romeo C, Cuomo CA, et al. 1995. Cleavage at a V(D)Jrecombination signal requires only RAG1 and RAG2 proteins and occurs in two steps. Cell 83:387–95

93. McMahan CJ, Difilippantonio MJ, Rao N, Spanopoulou E, Schatz DG. 1997. A basic motif in theN-terminal region of RAG1 enhances V(D)J recombination activity. Mol. Cell. Biol. 17:4544–52

94. McMahan CJ, Sadofsky MJ, Schatz DG. 1997. Definition of a large region of RAG1 that is importantfor coimmunoprecipitation of RAG2. J. Immunol. 158:2202–10

95. Mizuta B, Mizuta M, Araki S, Kitamura D. 2002. RAG2 is down-regulated by cytoplasmic sequestrationand ubiquitin-dependent degradation. J. Biol. Chem. 277:41423–27

96. Mo X, Bailin T, Sadofsky MJ. 1999. RAG1 and RAG2 cooperate in specific binding to the recombinationsignal sequence in vitro. J. Biol. Chem. 274:7025–31

97. Mo XM, Bailin T, Sadofsky MJ. 2001. A C-terminal region of RAG1 contacts the coding DNA duringV(D)J recombination. Mol. Cell. Biol. 21:2038–47

98. Mundy CL, Patenge N, Matthews AGW, Oettinger MA. 2002. Assembly of the RAG1/RAG2 synapticcomplex. Mol. Cell. Biol. 22:69–77

99. Nagawa F, Hirose S, Nishizumi H, Nishihara T, Sakano H. 2004. Joining mutants of RAG1 and RAG2that demonstrate impaired interactions with the coding-end DNA. J. Biol. Chem. 279:38360–68

100. Nagawa F, Kodama M, Nishihara T, Ishiguro K, Sakano H. 2002. Footprint analysis of recombinationsignal sequences in the 12/23 synaptic complex of V(D)J recombination. Mol. Cell. Biol. 22:7217–25

101. Niehues T, Perez-Becker R, Schuetz C. 2010. More than just SCID: the phenotypic range of combinedimmunodeficiencies associated with mutations in the recombinase activating genes (RAG) 1 and 2. Clin.Immunol. 135:183–92

102. Nishihara T, Nagawa F, Imai T, Sakano H. 2007. RAG-heptamer interaction in the synaptic complexis a crucial biochemical checkpoint for the 12/23 recombination rule. J. Biol. Chem. 283:4877–85

103. Peak MM, Arbuckle JL, Rodgers KK. 2003. The central domain of core RAG1 preferentially recognizessingle-stranded recombination signal sequence heptamer. J. Biol. Chem. 278:18235–40

104. Piirila H, Valiaho J, Vihinen M. 2006. Immunodeficiency mutation databases (IDbases). Hum. Mutat.27:1200–8

105. Qiu JX, Kale SB, Schultz HY, Roth DB. 2001. Separation-of-function mutants reveal critical roles forRAG2 in both the cleavage and joining steps of V(D)J recombination. Mol. Cell 7:77–87


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


106. Ramon-Maiques S, Kuo AJ, Carney D, Matthews AG, Oettinger MA, et al. 2007. The plant home-odomain finger of RAG2 recognizes histone H3 methylated at both lysine-4 and arginine-2. Proc. Natl.Acad. Sci. USA 104:18993–98

107. Ramsden DA, McBlane JF, van Gent DC, Gellert M. 1996. Distinct DNA sequence and structurerequirements for the two steps of V(D)J recombination signal cleavage. EMBO J. 15:3197–206

108. Ranganath S, Carpenter AC, Gleason M, Shaw AC, Bassing CH, Alt FW. 2008. Productive couplingof accessible Vβ14 segments and DJβ complexes determines the frequency of Vβ14 rearrangement.J. Immunol. 180:2339–46

109. Raval P, Kriatchko AN, Kumar S, Swanson PC. 2008. Evidence for Ku70/Ku80 association with full-length RAG1. Nucl. Acids Res. 36:2060–72

110. Rodgers KK, Bu Z, Fleming KG, Schatz DG, Engelman DM, Coleman JE. 1996. A zinc-binding domaininvolved in the dimerization of RAG1. J. Mol. Biol. 260:70–84

111. Rodgers KK, Villey IJ, Ptaszek L, Corbett E, Schatz DG, Coleman JE. 1999. A dimer of the lymphoidprotein RAG1 recognizes the recombination signal sequence and the complex stably incorporates thehigh mobility group protein HMG2. Nucleic Acids Res. 27:2938–46

112. Roman CAJ, Cherry SR, Baltimore D. 1997. Complementation of V(D)J recombination deficiency inRAG-1(-/-) B cells reveals a requirement for novel elements in the N-terminus of RAG-1. Immunity7:13–24

113. Ruthenburg AJ, Allis CD, Wysocka J. 2007. Methylation of lysine 4 on histone H3: intricacy of writingand reading a single epigenetic mark. Mol. Cell 25:15–30

114. Schatz DG, Ji Y. 2011. Recombination centers and the orchestration of V(D)J recombination. Nat. Rev.Immunol. 11:251–63

115. Schultz HY, Landree MA, Qiu JX, Kale SB, Roth DB. 2001. Joining-deficient RAG1 mutants blockV(D)J recombination in vivo and hairpin opening in vitro. Mol. Cell 7:65–75

116. Sheehan KM, Lieber MR. 1993. V(D)J recombination: signal and coding joint resolution are uncoupledand depend on parallel synapsis of the sites. Mol. Cell. Biol. 13:1363–70

117. Shiloh Y. 2003. ATM and related protein kinases: safeguarding genome integrity. Nat. Rev. Cancer3:155–68

118. Shimazaki N, Tsai AG, Lieber MR. 2009. H3K4me3 stimulates the V(D)J RAG complex for both nickingand hairpinning in trans in addition to tethering in cis: implications for translocations. Mol. Cell 34:535–44

119. Shlyakhtenko LS, Gilmore J, Kriatchko AN, Kumar S, Swanson PC, Lyubchenko YL. 2009. Molecularmechanism underlying RAG1/RAG2 synaptic complex formation. J. Biol. Chem. 284:20956–65

120. Simkus C, Anand P, Bhattacharyya A, Jones JM. 2007. Biochemical and folding defects in a RAG1 variantassociated with Omenn syndrome. J. Immunol. 179:8332–40

121. Simkus C, Bhattacharyya A, Zhou M, Veenstra TD, Jones JM. 2009. Correlation between recombinaseactivating gene 1 ubiquitin ligase activity and V(D)J recombination. Immunology 128:206–17

122. Simkus C, Makiya M, Jones JM. 2009. Karyopherin alpha 1 is a putative substrate of the RAG1 ubiquitinligase. Mol. Immunol. 46:1319–25

123. Sleckman BP, Bassing CH, Hughes MM, Okada A, D’Auteuil M, et al. 2000. Mechanisms that directordered assembly of T cell receptor beta locus V, D, and J gene segments. Proc. Natl. Acad. Sci. USA97:7975–80

124. Sobacchi C, Marrella V, Rucci F, Vezzoni P, Villa A. 2006. RAG-dependent primary immunodeficiencies.Hum. Mutat. 27:1174–84

125. Spanopoulou E, Cortes P, Shih C, Huang CM, Silver DP, et al. 1995. Localization, interaction, andRNA binding properties of the V(D)J recombination-activating proteins RAG1 and RAG2. Immunity3:715–26

126. Spanopoulou E, Zaitseva F, Wang F-H, Santagata S, Baltimore D, Panayotou G. 1996. The homeo-domain of Rag-1 reveals the parallel mechanisms of bacterial and V(D)J recombination. Cell 87:263–76

127. Steen SB, Gomelsky L, Roth DB. 1996. The 12/23 rule is enforced at the cleavage step of V(D)Jrecombination in vivo. Genes Cell. 1:543–53

128. Swanson PC. 2001. The DDE motif in RAG-1 is contributed in trans to a single active site that catalyzesthe nicking and transesterification steps of V(D)J recombination. Mol. Cell. Biol. 21:449–58


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


129. Swanson PC. 2002. A RAG-1/RAG-2 tetramer supports 12/23-regulated synapsis, cleavage, and trans-position of V(D)J recombination signals. Mol. Cell. Biol. 22:7790–801

130. Swanson PC. 2002. Fine structure and activity of discrete RAG-HMG complexes on V(D)J recombina-tion signals. Mol. Cell. Biol. 22:1340–51

131. Swanson PC. 2004. The bounty of RAGs: recombination signal complexes and reaction outcomes.Immunol. Rev. 200:90–114

132. Swanson PC, Desiderio S. 1998. V(D)J recombination signal recognition: distinct, overlapping DNA-protein contacts in complexes containing RAG1 with and without RAG2. Immunity 9:115–25

133. Swanson PC, Desiderio S. 1999. RAG-2 promotes heptamer occupancy by RAG-1 in the assembly of aV(D)J initiation complex. Mol. Cell. Biol. 19:3674–83

134. Swanson PC, Kumar S, Raval P. 2009. Early steps of V(D)J rearrangement: insights from biochemicalstudies of RAG-RSS complexes. Adv. Exp. Med. Biol. 650:1–15

135. Swanson PC, Volkmer D, Wang L. 2004. Full-length RAG-2, and not full-length RAG-1, specificallysuppresses RAG-mediated transposition but not hybrid joint formation or disintegration. J. Biol. Chem.279:4034–44

136. Talukder SR, Dudley DD, Alt FW, Takahama Y, Akamatsu Y. 2004. Increased frequency of aberrantV(D)J recombination products in core RAG-expressing mice. Nucleic Acids Res. 32:4539–49

137. Thomas JO, Travers AA. 2001. HMG1 and 2, and related “architectural” DNA-binding proteins. TrendsBiochem. Sci. 26:167–74

138. Thomson TM, Guerra-Rebollo M. 2010. Ubiquitin and SUMO signalling in DNA repair. Biochem. Soc.Trans. 38:116–31

139. Tillman RE, Wooley AL, Hughes MM, Khor B, Sleckman BP. 2004. Regulation of T-cell receptorβ-chain gene assembly by recombination signals: the beyond 12/23 restriction. Immunol. Rev. 200:36–43

140. Tsai CL, Schatz DG. 2003. Regulation of RAG1/RAG2-mediated transposition by GTP and the C-terminal region of RAG2. EMBO J. 22:1922–30

141. van Gent DC, Hiom K, Paull TT, Gellert M. 1997. Stimulation of V(D)J cleavage by high mobilitygroup proteins. EMBO J. 16:2665–70

142. Wang H, Zhai L, Xu J, Joo HY, Jackson S, et al. 2006. Histone H3 and H4 ubiquitylation by theCUL4-DDB-ROC1 ubiquitin ligase facilitates cellular response to DNA damage. Mol. Cell 22:383–94

143. Wang X, Xiao G, Zhang Y, Wen X, Gao X, et al. 2008. Regulation of Tcrb recombination ordering byc-Fos-dependent RAG deposition. Nat. Immunol. 9:794–801

144. Wu C, Bassing CH, Jung D, Woodman BB, Foy D, Alt FW. 2003. Dramatically increased rearrangementand peripheral representation of Vβ14 driven by the 3′Dβ1 recombination signal sequence. Immunity18:75–85

145. Yamtich J, Sweasy JB. 2010. DNA polymerase family X: function, structure, and cellular roles. Biochim.Biophys. Acta 1804:1136–50

146. Yin FF, Bailey S, Innis CA, Ciubotaru M, Kamtekar S, et al. 2009. Structure of the RAG1 nonamerbinding domain with DNA reveals a dimer that mediates DNA synapsis. Nat. Struct. Mol. Biol. 16:499–508

147. Yu K, Taghva A, Ma Y, Lieber MR. 2004. Kinetic analysis of the nicking and hairpin formation steps inV(D)J recombination. DNA Repair (Amst.) 3:67–75

148. Yu KF, Lieber MR. 1999. Mechanistic basis for coding end sequence effects in the initiation of V(D)Jrecombination. Mol. Cell. Biol. 19:8094–102

149. Yu KF, Lieber MR. 2000. The nicking step in V(D)J recombination is independent of synapsis: impli-cations for the immune repertoire. Mol. Cell. Biol. 20:7914–21

150. Yurchenko V, Xue Z, Sadofsky M. 2003. The RAG1 N-terminal domain is an E3 ubiquitin ligase. GenesDev. 17:581–85

151. Zhang L, Reynolds TL, Shan X, Desiderio S. 2011. Coupling of V(D)J recombination to the cell cyclesuppresses genomic instability and lymphoid tumorigenesis. Immunity 34:163–74

152. Zhang ZX, Espinoza CR, Yu ZH, Stephan R, He T, et al. 2006. Transcription factor Pax5 (BSAP) trans-activates the RAG-mediated V-H-to-DJ(H) rearrangement of immunoglobulin genes. Nat. Immunol.7:616–24

153. Zhao S, Gwyn LM, De P, Rodgers KK. 2009. A non-sequence-specific DNA binding mode of RAG1 isinhibited by RAG2. J. Mol. Biol. 387:744–58


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.


RELATED RESOURCES

IDbases: Databases for Immunodeficiency-Causing Mutations. 2011. RAG1base: mutation registryfor RAG1 deficiency. http://bioinf.uta.fi/RAG1base/

IDbases: Databases for Immunodeficiency-Causing Mutations. 2011. RAG2base: mutation registryfor RAG2 deficiency. http://bioinf.uta.fi/RAG2base/


Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.

http://bioinf.uta.fi/RAG1base/

http://bioinf.uta.fi/RAG2base/

GE45-FrontMatter ARI 11 October 2011 19:52

Annual Review ofGenetics

Volume 45, 2011 Contents

Comparative Genetics and Genomics of Nematodes: GenomeStructure, Development, and LifestyleRalf J. Sommer and Adrian Streit � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 1

Uncovering the Mystery of Gliding Motility in the MyxobacteriaBeiyan Nan and David R. Zusman � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �21

Genetics and Control of Tomato Fruit Ripening and Quality AttributesHarry J. Klee and James J. Giovannoni � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �41

Toxin-Antitoxin Systems in Bacteria and ArchaeaYoshihiro Yamaguchi, Jung-Ho Park, and Masayori Inouye � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �61

Genetic and Epigenetic Networks in Intellectual DisabilitiesHans van Bokhoven � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �81

Axis Formation in HydraHans Bode � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 105

The Rules of Engagement in the Legume-Rhizobial SymbiosisGiles E.D. Oldroyd, Jeremy D. Murray, Philip S. Poole, and J. Allan Downie � � � � � � � � 119

A Genetic Approach to the Transcriptional Regulationof Hox Gene ClustersPatrick Tschopp and Denis Duboule � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 145

V(D)J Recombination: Mechanisms of InitiationDavid G. Schatz and Patrick C. Swanson � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 167

Human Copy Number Variation and Complex Genetic DiseaseSanthosh Girirajan, Catarina D. Campbell, and Evan E. Eichler � � � � � � � � � � � � � � � � � � � � � � 203

DNA Elimination in Ciliates: Transposon Domesticationand Genome SurveillanceDouglas L. Chalker and Meng-Chao Yao � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 227

Double-Strand Break End Resection and Repair Pathway ChoiceLorraine S. Symington and Jean Gautier � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 247

vi

Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.

GE45-FrontMatter ARI 11 October 2011 19:52

CRISPR-Cas Systems in Bacteria and Archaea: Versatile Small RNAsfor Adaptive Defense and RegulationDevaki Bhaya, Michelle Davison, and Rodolphe Barrangou � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 273

Human Mitochondrial tRNAs: Biogenesis, Function,Structural Aspects, and DiseasesTsutomu Suzuki, Asuteka Nagao, and Takeo Suzuki � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 299

The Genetics of Hybrid IncompatibilitiesShamoni Maheshwari and Daniel A. Barbash � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 331

Maternal and Zygotic Control of Zebrafish Dorsoventral AxialPatterningYvette G. Langdon and Mary C. Mullins � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 357

Genomic Imprinting: A Mammalian Epigenetic Discovery ModelDenise P. Barlow � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 379

Sex in FungiMin Ni, Marianna Feretzaki, Sheng Sun, Xuying Wang, and Joseph Heitman � � � � � � 405

Genomic Analysis at the Single-Cell LevelTomer Kalisky, Paul Blainey, and Stephen R. Quake � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 431

Uniting Germline and Stem Cells: The Function of Piwi Proteinsand the piRNA Pathway in Diverse OrganismsCelina Juliano, Jianquan Wang, and Haifan Lin � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 447

Errata

An online log of corrections to Annual Review of Genetics articles may be found at http://genet.annualreviews.org/errata.shtml

Contents vii

Ann

u. R

ev. G

enet

. 201

1.45

:167

-202

. Dow

nloa

ded

from

ww

w.a

nnua

lrev

iew

s.or

g A

cces

s pr

ovid

ed b

y O

hio

Uni

vers

ity L

ibra

ries

on

11/3

0/16

. For

per

sona

l use

onl

y.

V(D)J Recombination: Mechanisms of Initiation 720/2018 Group 3 Paper.pdfOF V(D)J RECOMBINATION Based...

Documents

Transcript of V(D)J Recombination: Mechanisms of Initiation 720/2018 Group 3 Paper.pdfOF V(D)J RECOMBINATION Based...