- Email: [email protected]
Mechanism of V(D)J recombination
1 75
Mechanism of V(D)J recombination Molly Bogue and David B Roth* V(D)J recombination can be separated into two basic operations: DNA cleavage and joining of broken ends. Our understanding of both reactions has increased substantially in the past year. Major advances include the development of a cell-free system capable of cleavage and the identification of several proteins involved in both V(D)J recombination and double-strand break repair.
Address Department of Microbiology and Immunology,Baylor College of Medicine, ImmunologyM929, 1 Baylor Plaza, Houston, TX 77030, USA; *e-mail:[email protected] Current Opinion in Immunology 1996, 8:175-180 © Current Biology Ltd ISSN 0952-7915
Abbreviations D diversity DNA-PK DNA-dependentprotein kinase DSB double-strandbreak DNA-PKcs catalyticsubunit of DNA-dependent protein kinase J joining RSS recombinationsignal sequences SClD severe combined immunodeficiency TCR T-cell receptor V variable
observation that a certain RAG-1 mutant renders the recombination process hypersensitive to alterations in the coding nucleotides immediately adjacent to the RSS [6]. The joining mechanism was studied by analysis of mutations that interfere with junction formation. T h e mouse scid (severe combined immunodeficiency) mutation blocks the formation of coding joints and is characterized by accumulation of covalently sealed (hairpin) coding ends [3], which are normal intermediates in the reaction (see below). T h e murine scid mutation provided the first link between V(D)J recombination and DSB repair with the discovery that cells from SCID mice exhibit defective DSB repair [7-9]. Additional mutations that interfere with both V(D)J joining and DSB repair were identified subsequently [10,11]. T h e pace of discovery has accelerated in the past year, and much more is now known about both the cleavage and joining components of the V(D)J recombination mechanism. This review will focus on these exciting developments. The cleavage mechanism: breaks and hairpins In vivo studies
Introduction
V(D)J recombination assembles T C R and immunoglobulin genes from germline gene segments during development of precursor lymphocytes (reviewed in [1]). T h e recombination system recognizes conserved sequence elements, termed recombination signal sequences (RSS), that adjoin the V, D, and J coding elements. RSS consist of conserved heptamer and nonamer elements that are separated by 12 or 23 nucleotide spacer sequences. Efficient recombination requires a pair of RSS with different spacer lengths (the '12/23 rule'). Work during the early years of this decade provided several important clues about the recombination mechanism. Information about the cleavage reaction was provided by analysis of presumed in vivo recombination intermediates in precursor lymphocytes. These studies indicated that double-strand breaks (DSB) are introduced precisely between the RSS and the adjacent coding segment [2-5], producing two types of terminus, coding ends and signal ends, that join to form coding joints and signal joints (Fig. 1). A role for the lymphoid-specific recombination activating proteins, RAG-1 and RAG-2, in the cleavage reaction was suggested by the observation that both are required for formation of DSB at RSS in vivo [5]. T h e first evidence suggesting that RAG-1 might be directly involved in recognition of the RSS was provided by the
double-strand
As described above, previous characterization of site-specific DSB in DNA preparations from mouse thymus or bone marrow suggested that blunt signal ends and hairpin coding ends are recombination intermediates. However, the possibility that these molecules might represent dead-end products rather than true intermediates could not be excluded. This issue was recently addressed by studies using a pre-B cell line in which rearrangement of immunoglobulin light chain loci can be induced by growth at high temperature [12"]. Although coding ends are detected, they are rapidly joined. Signal ends accumulate at high temperature and can be chased into joined products upon return to low temperature. These observations provide compelling evidence that signal ends are authentic reaction intermediates. The detection of hairpin coding ends in vivo in the absence of the scid mutation strongly suggests that hairpin coding ends are normal reaction intermediates, a point which is further supported by biochemical experiments (see below). Insights into the mechanism of cleavage and hairpin formation were provided by analysis of the fine structure of signal and coding ends produced during T C R gene rearrangement in thymocytes of scid mice. Cleavage occurs precisely at the border between the RSS and the coding segment, producing a flush DSB at the signal end. T h e other product of cleavage is a coding end in
1 76 Lymphocyte development
Figure 1
(a) Cleavage
(b) Joining
C C ~ ~ " ~ T C : _ _
'~7 Join signal ends
~
OPeT~hairpi.~ns
---GGcc ~7 C C )
F°rm hairpins (TCG~
~
AG~ ~
-Signal joint
Process ends join coding ends
CCggTC GGccAG P nucleotides Coding joint
© 1996CurrentOpinionin Immunology A double-strand cleavage model for V(D)J recombination. The cleavage step is illustrated in (a). Recombination requires the presence of a pair of RSS containing 23 and 12 nucleotide spacer sequences (shown as filled and open triangles, respectively). Cleavage occurs in two steps, both of which are dependent upon the presence of RAG-1 and RAG-2 proteins [15"]. First, a nick is introduced precisely at the border between the RSS and the coding elements. (Coding segments are indicated by two nucleotides shown adjacent to each RSS. These particular nucleotides were chosen for purposes of illustration only.) The second step of the reaction, hairpin formation, involves attack of the 3' end of the coding segment on the opposite strand (small black arrows) [15°°]. Attack occurs precisely at the border between the RSS and the coding element, generating a blunt signal end and a hairpin structure that contains all nucleotides of the coding element [13°,15"]. The joining step is illustrated in (b). The blunt signal ends are joined to form a reciprocal product that contains a signal joint. Hairpin coding ends must be opened prior to joining. Presumably, the hairpin opening nuclease(s) (shown as scissors) recognize some aspect of the hairpin structure, such as distortions in DNA structure in the vicinity of the loop [4?], and introduces a nick near the terminus. Nicking at the terminus would generate a blunt end (shown on the right), whereas introduction of a nick near the terminus would produce a short single-stranded extension (shown on the left). Incorporation of the extension into the junction yields a specific type of extra nucleotides, termed P nucleotides, that are frequently found at coding joints [43,46]. Not shown are additional modifications, such as addition of N nucleotides by terminal deoxynucleotidyl transferase or loss of nucleotides from coding ends prior to joining (reviewed in [1]).
which theterminal nucleotides of each strand are ligated, producing a hairpin that contains all nucleotides of the coding segment [13°]. These results suggest a coupled mechanism in which hairpin coding ends and signal ends are both formed by the same cleavage event (Fig. 1).
Biochemical studies A major breakthrough in deciphering the mechanism of V(D)J recombination occurred in 1995, with the development of a cell-free system capable of performing the initial steps of the reaction, recognition and cleavage [14°°]. RAG-1 and RAG-2 are the only proteins required for cleavage, which produces blunt signal ends and hairpin coding ends of precisely the form detected in vivo [14"',15°']. These data also strongly suggest that hairpin coding ends are normal intermediates in V(D)J recombination. More detailed biochemical studies using oligonucleotide substrates revealed that cleavage can be separated into two steps: introduction of a single-stranded nick immediately 5' to the RSS; and conversion of this nick into a hairpin coding end, which liberates a blunt signal end (Fig. 1). Both steps require an RSS, RAG-l, and RAG-2 - - no additional proteins or high energy cofactors are required [15°°]. T h e possibility that RAG-1 and RAG-2 might func-
tion as a complex is supported by co-immunoprecipitation studies demonstrating an in vivo association between the two proteins [16].
Joining broken ends: connections between V(D)J recombination and DNA repair T h e realization that the mouse scid mutation affects both DSB repair and V(D)J recombination stimulated a search for additional mutations that might affect both reactions. Several mutants (comprising four complementation groups, termed XRCC4 through XRCC7) with defects in both DSB repair and V(D)J recombination have been identified (for review, see [17°,18"]). All mutants are defective for both signal and coding joint formation except murine scid (XRCC7, see below). XRCC5 mutants lack expression of a D N A end binding activity termed Ku [19,20], a heterodimer composed of 70kDa and 80kDa subunits that binds to altered DNA structures such as single-stranded gaps, broken ends, and hairpins [21-23]. In addition to its end-binding properties, Ku is a DNA helicase [24] and serves as the DNA-binding component of a DNA-dependent protein kinase, DNA-PK [21,25,26]. XRCC5 mutants do not produce a functional 80kDa Ku subunit, lack DNA-PK activity [27°], and are complemented by Ku80 eDNA [28"',29°,30°]. The sxi-1 cell line (XRCC6) exhibits defects in DSB repair, DNA end binding, and V(D)J recombination, and the DSB
Mechanism of V(D)J recombinationBogueand Roth 177
repair phenotype can be complemented by introduction of a cDNA encoding Ku70 [17",31]. T h e identification of the DNA-binding component of DNA-PK as a critical factor in both V(D)J recombination and DSB repair suggested that mutations in the catalytic subunit of the kinase, DNA-PKcs, might also affect these processes. This prediction was fulfilled by analysis of XRCC7 mutants (a group that includes the mouse scid mutation), which lack DNA-PK activity [32°°,33°°,34]. T h e murine scid defect has recently been mapped to DNA-PKcs [32°°,33"°]. It is remarkable that three of the four classes of mutation that affect V(D)J recombination and DSB repair also affect DNA-PK. T h e remaining gene, XRCC4, encodes a novel protein (see Note added on proof). When activated by binding to DNA lesions, DNA-PK is capable of phosphorylating a wide range of substrates in vitro, including several transcription factors, Ku, replication protein A, and p53 [34-36]. Possible roles of DNA-PK include signaling the presence of DNA damage, recruiting other components of DNA repair machinery, and inhibiting transcription in the vicinity of DNA lesions [18°,33",37,38]. Another possibility is that this large (465 kDa) protein [39°] might perform a scaffolding function, assembling components of the repair apparatus in the vicinity of DNA damage. Analysis of the nucleotide sequence of the human DNA-PKcs gene provided an unexpected clue about possible function. T h e carboxyl terminus contains a region that is homologous to the catalytic domains of proteins in the phosphatidylinositol-3 kinase superfamily [39°], which includes the AtaxiaTelangiectasia protein (ATM) and several other proteins that regulate cell-cycle progression (reviewed in [40]). Whether DNA-PK plays a similar role remains to be determined, although it is noteworthy that cell lines deficient for DNA-PK activity do not exhibit obvious checkpoint defects [41].
A w o r k i n g model for V(D)J r e c o m b i n a t i o n Although many questions remain, we have attempted to organize the recent wealth of information into a working model (Fig. 2). T h e first step involves recognition of RSS and cleavage by the RAG-1 and RAG-2 proteins, perhaps with the help of accessory (regulatory?) factors. After cleavage RAG-1 and RAG-2 may remain associated with the signal ends, protecting them from degradation and possibly facilitating joining [15°°]. T h e suggestion that all four ends might remain associated after cleavage is supported by the intriguing observation that the processing of coding ends may depend upon whether they are attached to an RSS with a 12 nucleotide or a 23 nucleotide spacer [42]. After cleavage, Ku binds to hairpin coding ends and signal ends and recruits DNA-PKcs, which somehow promotes the joining of hairpin coding ends.
T h e following observations must be kept in mind when considering the role of DNA-PK in coding joint formation. Hairpin coding ends accumulate to high levels in thymocytes of scid mice [3,13°]. However, the ability of scid cell lines to open exogenously introduced hairpin substrates is not detectably impaired [43,44]. One solution to this paradox is to propose that authentic hairpin coding ends are specifically protected by a DNA-protein complex that renders them unavailable for hairpin opening. Perhaps DNA-PKcs regulates accessibility of the hairpin ends in this complex to nucleases [13°,45]. This could be accomplished in several ways. Perhaps phosphorylation of substrate(s) serves to alter the complex, making the hairpins accessible to nuclease activities. Alternatively, phosphorylation events could serve to recruit a specific nuclease into the complex. DNA-PKcs could have functions in addition to kinase activity, particularly in view of its whopping size. The mechanism of hairpin opening remains uncharacterized. Analysis of coding junction sequences suggests that hairpin opening occurs within a few nucleotides of the terminus [1,43,46]. It is thought that hairpins are frequently opened to give ends with short single-stranded extensions, the presumed precursors of a particular class of junctional inserts termed P nucleotides (reviewed in [1]). T h e hairpin-opening nuclease could be a specialized enzyme that specifically recognizes hairpins or a more general enzyme that recognizes a variety of alternative DNA structures. Precedent for the latter possibility is provided by the observation that both mung bean nuclease and P1 nuclease, which recognize distortions in DNA structure, can efficiently open hairpins to leave short single-stranded extensions [47]. Opened coding ends then become accessible for modification, including loss of nucleotides and addition of extra nucleotides by terminal deoxynucleotidyl transferase, prior to joining (reviewed in [1]). Once hairpins are opened, they are joined much more rapidly than signal ends [12°]. Access of signal ends to modification appears to be much more restricted, perhaps because they remain associated with RAG-1 and RAG-2, as suggested above.
Conclusions and future prospects The past year has seen a number of exciting advances in our knowledge of the mechanism of V(D)J recombination. T h e initial step of the reaction, cleavage, is now amenable to detailed biochemical investigation, and the basic mechanism of cleavage should be elucidated in the near future. However, important aspects of the cleavage reaction as it occurs in vivo have not yet been reconstituted in cell-free systems. Several lines of evidence (discussed in [15°°]) suggest that in living cells cleavage occurs at pairs of signals, rather than individual signals. In contrast, in vitro cleavage at one RSS is not affected by the presence of a second RSS in the same substrate [14°°,15°°]. It will be interesting to see whether factors in addition to RAG-1 and RAG-2 are required for regulation of cleavage and
178 Lymphocyte development
Figure 2 A proposed role for DNA-PKcs: regulation of hairpin opening by controlling accessibility of hairpins to nucleases. We propose that RAG-1 and RAG-2 cleave at the RSS, and the resulting hairpin coding ends and blunt signal ends are bound by Ku. At this stage, it is possible that all four broken ends are in proximityto one another. The proposed DNA-protein complex renders the hairpin ends inaccessible to the hairpin-openingmachinery. DNA-PKcs (the murine scid factor) binds to Ku, is activated, and phosphorylates target molecules, resulting in increased accessibility of the coding ends to the hairpin opening machinery. Once accessible, hairpins are rapidly opened and joined.
M II
Cleavage at RSS Ku binds to broken ends
~
G
Hairpincoding end
I Hairpins inaccessible I
Ku 70•80 -1 + RAG-2
Bind and activate DNA-PKcs
~7
DNA-PKcs
Phosphorylationof target(s)
~7
IHairpins accessible I
Nuclease
Open hairpins join coding and signal ends
© 1996CurrentOpinioninImmunolog~ •
RSS
[]
for enforcement of the 12/23 rule. Other important levels of regulation include cell cycle control of recombinase activity (reviewed in [48"]) and substrate accessibility (reviewed in [1]).
Coding segments
0
Ku 70•80
O
RAG-1 + RAG-2
Although substantial progress has been made in identification of factors that participate in the joining reaction, the biochemical mechanisms of the steps following cleavage and hairpin formation remain undefined. Although the
Mechanism of V(D)J recombination Bogue and Roth
genes encoding three factors involved in the joining reaction have been identified, the XRCC4 gene remains to be characterized, and perhaps other complementation groups will be identified. More detailed analysis of some recently identified mutations in other species may be informative. For example, cells homozygous for the equine scid mutation lack detectable DNA-PK activity or immunoreactive DNA-PKcs protein [49"], and, unlike cells from scid mice, both coding and signal joint formation are severely affected [49"]. It will also be interesting to examine the effects of the recently described human cell line bearing an XRCC7 mutation [50] on coding and signal joint formation. Ultimately, biochemical experiments will be necessary to obtain detailed information about the joining mechanism. T h e next few years should prove to be exciting, as investigations on many fronts reveal more about the mechanism of V(D)J recombination and illuminate the connections between this lymphoid-specific reaction and general DNA repair mechanisms.
Note added in proof Since the submission of this review, a paper reporting that the XRCC4 gene encodes a novel protein has been published [51].
Acknowledgements The authors apologize to our colleagues whose work could not be cited due to space limitations. We are grateful to Martin Gellert and Katberyn Meek for communicating unpublished results. We thank Mark Landree, Sharon Roth and Sharri Steen for critical review of the manuscript. Mary Lowe-provided secretarial assistance. Work in the author's laboratory is supported by grants from the NIH (AI-36420) and the American Cancer Society (DB-118). DB Roth is a Charles E Culpeper Medical Scholar.
and a deficiency in DNA double-strand break repair. Proc Natl Acad Sci USA 1991, 88:1394-1397. 8.
FuIop GM, Phillips RA: The $cid mutation in mice causes a general defect in DNA repair. Nature 1990, 347:4?9-482.
9.
Hendrickson EA, Ctin X-Q, Bump EA, Schatz DG, Oettinger M, Weaver DT: A link between double-strand break-related repair and V(D)J recombination: the sc/d mutation. Proc Nat/Acad Sci USA 1991, 88:4061-4065.
10.
Pergola F, Zdzienicka MZ, Lieber MR: V(D)J recombination In mammalian mutants defective in DNA double-strand break repair. Mol Cell Bio/1993, 13:3464-3471.
11.
Taccioli GE, Rathbun G, Oltz E, Stamato T, Jeggo PA, AIt FW: Impairment of V(D)J recombination In double-strand break repair mutants. Science 1993, 260:207-210.
RamsdenDA, Gelled M: Formation and resolution of doublestrand break intermediates in V(D)J rearrangemenL Genes Dev 1995, 9:2409-2420. These elegant studies make use of a cell line with inducible V(D)J recombination activity to demonstrate that broken molecules are recombination intermediates. The first demonstration of hairpin coding ends in vivo in the absence of the scid mutation. 12.
•
13. •
Zhu C, Roth DB: Characterization of coding ends In thymocytes of scld mice: implications for the mechanism of V(D)J recombination. Immunity 1995, 2:101-112. A detailed investigation of the structure of coding and signal ends isolated from scid mouse thymocytes. The first analysis of hairpin coding ends at the nucleotide sequence level. The structures of ends produced in vivo are exactly the same as those produced by cleavage in cell-free systems. A hairpin accessibility model for scid defect is proposed. 14. ••
Van Gent DC, McBlane JF, Remsden DA, Sadofsky MJ, Hesse JE, Gelled M: Initiation of V(D)J recombination in a cell-free system. Cell 1995, 81:925-934. A landmark study, the first to demonstrate specific cleavage at RSS to give blunt signal ends and hairpin coding ends in a cell-frse system. Also showed that cleavage depends on both RAG-1 and RAG-2. 15. •,
McBlane JF, Van Gent DC, Ramsden DA, Romeo C, Cuomo CA, Gelled M, Oettinger MA: Cleavage at a V(D)J recombination signal requires only RAG-1 and RAG-2 proteins end occurs In two steps. Cell 1995, 83:387-395. Purified RAG-t and RAG-2 proteins and oligonucleotide substrates are used to dissect the details of the cleavage reaction. 16.
References and recommended reading Papers of particular interest, published within the annual period of review, have been highlighted as: * of special interest == of outstanding interest 1.
Lewis SM: The mechanism of V(D)J joining: lessons from molecular, immunological and comparative analyses. Adv Immunol 1994, 56:27-150.
2.
Roth DB, Nakajima PB, Menetski JP, Bosma MJ, Gelled M: V(D)J recombination in mouse thymocytes: double-strand breaks near T cell receptor rearrangement signals. Cell 1992, 69:41-53.
3.
Roth DB, Menetski JP, Nakajima PB, Bosma MJ, Gelled M: V(D)J recombination: broken DNA molecules with covalentiy sealed (hairpin) coding ends in scid mouse thymocytes. Cell 1992, 70:983-991.
4.
Roth DB, Zhu C, Gelled M: Characterization of broken DNA molecules associated with V(D)J recombination. Proc Nat/Acad Sci USA 1993, 90:10788-10792.
5.
Schlissel M, Constantinescu A, Morrow T, Baxter M, Pang A: Double-strand signal sequence breaks in V(D)J recombination are blunt, 5'-phosphorylated, RAG-dependent, and cell cycle regulated. Genes Dev 1993, 7:2520-2532.
6.
Sadofsky MJ, Hesse JE, Van Gent DC, Gelled M: RAG1 mutations that affect the target specificity of V(D)J recombination: a possible direct role of RAG-1 in site recognition. Genes Dev 1995, 9:2193-2199.
7.
Biedermann KA, Sun J, Giaccia AJ, Tosto LM, Brown JM: scid mutation in mice confers hypersensitivity to Ionizing radiation
179
Leu TM, Schatz DG: RAG-1 and RAG-2 are components of a high-molecular-weight complex, and association of RAG-2 with this complex is RAG-1 dependent. Mol Cell Biol 1995, 15:5657-5670.
17. An
Weaver DT: What to do at an end: DNA double-strand-break repair. Trends Genet 1995, 11:388-392. excellent recent review of of factors involved in both DSB repair and V(D)J recombination. Summarizes very recent data on XR-1 (XRCC4), sxi-1 (XRCC6), and sxi-2/sxi-3 (XRCCS) mutants. 18. •
JacksonSP, Jeggo PA: DNA double-strand break repair and V(D)J recombination: involvement of DNA-PK. Trends Biochem Sci 1995, 20:412-415. Very up-to-date review, with emphasis on role of DNA-PK. 19.
Getts RC, Stamato TD: Absence of a Ku-Ilke DNA end binding activity in the xrs double-strand DNA repair-deficient mutant, J Biol Chem 1994, 269:15981-15984.
20.
RathmellWK, Chu G: A DNA end-binding factor involved in double-strand break repair and V(D)J recombination. Mol Cell Biol 1994, 14:4741-4?48.
21.
Morozov VE, Falzon M, Anderson CW, Kuff EL: DNA-dependent protein kinase is activated by nicks and larger single-stranded gaps. J Biol Chem 1994, 269:16684-16688.
22.
Blier P, Griffith AJ, Craft J, Hardin JA: Binding of Ku protein to DNA: measurement of affinity for ends and demonstration of binding to nicks. J Biol Chem 1993, 268:7594-7601.
23.
Palllard S, Strauss F: Analysis of the mechanism of interaction of simian Ku protein with DNA. Nucleic Acids Res 1991, 19:5619-5624.
24.
Tuteja N, Tuteja R, Ochem A, Taneja P, Huang NW, Simoncsits A, Susic S, Rahman K, Marusic L, Chen J et el.: Human DNA heiicase II - a novel DNA unwinding enzyme identified as the Ku autoantigen. EMBO J 1994, 13:4991-5001.
25.
Gottlieb TM, Jackson SP: The DNA-dependent protein Idnase: requirement for DNA ends and association with Ku antigen. Cell 1993, 72:131-142.
180
Lymphocyte development
26.
Dvir A, Peterson SR, Knuth MW, Lu H, Dynan WS: Ku autoantlgen Is the regulatory component of a templateassociated protein kinase that phosphorylates RNA polymerase II. Proc Nat/Acad Sci USA 1992, 89:11920-11924.
27. •
Rnnie NJ, Gottlieb TM, Blunt T, Jeggo PA, Jackson SP: DNAdependent protein kinase activity Is absent in xrs-6 cells: implications for site-specific recombination and DNA doublestrand break repair. Proc Nat/Acad Sci USA 1995, 92:320-324. A sensitive new assay for DNA-PK activity is used to show that xrs-6 cells (XRCC5 mutants) which lack Ku end-binding, also lack DNA-PK activity. 28. ••
Taccioli GE, Gottlieb TM, Blunt T, Priestley A, Demengsot J, Mizuta R, Lehmann AR, AIt FW, Jackson SP, Jeggo PA: Ku80: product of the XRCC5 9ene and its role in DNA repair and V(D)J recombination. Science 1994, 265:1442-1445. First report identifying XRCC5 as Ku80. A cDNA encoding human Ku80 complements the V(D)J recombination, DNA end binding, and X-ray sensitivity phenotypes of XRCC5 mutant hamster cells. 29. •
Boubnov NV, Hall KT, Wills Z, Lee SE, He DM, Benjamin DM, Pulaski CR, Band H, Reeves W, Hendrickson EA, Weaver DT: Complementetion of the Ionizing radiation sensitivity, DNA end binding, and V(D)J recombination defects of double-strand break repair mutants by the p86 Ku autoanflgen. Proc Natl Acad Sci USA 1995, 92:890-894. Two new mutants with defects in DSB repair and V(D)J recombination, sxi-2 and sxi-3, are found to be in the XRCC5 group. Complemantation with a cDNA encoding Ku80 restores defects in V(D)J recombination and DNA repair.
15 and 37 in the amino-terminal transactlvation domain of human p53. Mo/Cell Bio11992, 12:5041-5049. 37.
Kuhn A, Gottlieb TM, Jackson SP, Grummt i: DNA-dependent protein kinase: a potent Inhibitor of transcription by RNA polymerase I. Genes Dev 1995, 9:193-203.
38.
Labhart P: DNA-dependent protein klnase specifically represses promoter-dlracted transcription Initiation by RNA polymerase I. Proc Nat/Acad Sci USA 1995, 92:2934-2938.
39. •
Hartley KO, Gell D, Smith GCM, Zhang H, Divecha N, Connelly MA, Admon A, Lees-Miller SP, Anderson CW, Jackson SP: DNAdependent protein kinase catalytic subunit: a relative of phosphatidyllnositol 3-klnase and the ataxla telanglectasla 9ene product. Cell 1995, 82:849-856. The sequence of the gene encoding DNA-PKcs reveals homology to a family of kinases related to phosphatidylinositol-3 kinase. No lipid kinase activity was detected, indicating that the catalytic function of DNA-PKcs may be restricted to protein kinase activity. 40.
Keith CT, Schreiber SL: PlK-ralatad kinases: DNA repelr, recombination, and cell cycle checkpoints. Science 1995, 270:50-51.
41.
Jeggo PA: X-ray sensitive mutants of Chinese hamster ovary cell line: radio-sensitivity of DNA synthesis. Mutation Res 1985, 145:171-176.
42.
Smider V, Rathmell WK, Lieber MR, Chu G: Restoration of X-ray resistance and V(D)J recombination In mutant cells by Ku cDNA. Science 1994, 266:288-291. Complementation with a cDNA encoding Ku80 restores defects in V(D)J recombination and DNA repair in XRCC5 mutant cells.
Ezekiel UR, Engler P, Stern D, Storb U: Asymmetric processing of coding ends and the effect of coding end nucleotlde composition on V(D)J recombination. Immunity 1995, 2:381-389.
43.
Lewis SM: P nucleotide insertions and the resolution of hairpin DNA structures in mammalian calls. Proc Natl Acad Sci USA 1994, 91:1332-1336.
31.
44.
Staunton JE, Weaver DT: scld cells efficiently Integrate hairpin and linear DNA substrates. Mol Cell Bio11994, 14:3876-3883.
45.
Roth DB, Lindahl T, Gellert M: How to make ends meet. Curr Biol 1995, 5:496-499.
46.
Meier JT, Lewis SM: P nucleotldes in V(D)J recombination: a fine-structure analysis. Mol Cell Biol 1993, 13:1078-1092.
47.
Kabotyanski EB, Zhu C, Roth DB: Hairpin opening by single-strand-specific nucleases. Nucleic Acids Res 1995, 23:3872-3881.
30. •
Lee SE, Pulaski CR, He DM, Benjamin DM, Voas M, Um J, Hendrickson EA: Isolation of mamma,an call mutants that are X-ray sensitive, impaired in DNA double-strand break repair and defective for V(D)J recombination. Murat Res 1995, 336:279-291.
32. ••
Kirchgassner CU, Patil CK, Evans JW, Cuomo CA, Fried LM, Carter T, Oettinger MA, Brown JM: DNA-dependent klnase (p350) as a candidate gene for the murine scld defect. Science 1995, 267:1178-1183. Positional cloning and complementation studies identified DNA-PKcs as a candidate for the murine scid gene. Analysis of DNA-PKcs protein levels by western blotting using two monoclonal antibodies revealed low levels of protein, suggesting that the mutation may cause decreased synthesis or decreased stability of the DNA-PKcs protein. 33. ••
48. Lin W-C, Desiderio S: V(D)J recombination and the cell cycle. • Immunol Today 1995, 16:279-289. Comprehensive review of regulation of V(D)J recombination activity, with emphasis on call-cycle regulation.
Blunt T, Finnio NJ, Taccioli GF_,Smith GCM, Demengeot J, Gottlieb TM, Mizuta R, Varghase AJ, AIt FW, Jeggo PA, Jackson SP: Defective DNA-dependent protein klnase activity is linked to V(D)J recombination and DNA repair defects associated with the murlne scld mutation. Cell 1995, 80:813-823. XRCC7 mutants lack DNA-PK activity due to deficiency in the catalytic subunit. Beth DSB repair and V(D)J recombination defects were complemented by yeast artiFmialchromosomes encoding DNA-PKcs.
Wiler R, Leber R, Moore BB, VanDyk LF, Perryman LE, Meek K: Equine severe combined Immunodefidency: a defect In V(D)J recombination and DNA-dependent protein Idnase activity. Proc Nat/Acad Sci USA 1995, 92:11485-11489. Defective DNA-PK activity in Arabian foals with autosomal recessive severe combined immunodeficiency is linked to defects in both coding and signal joint formation.
34.
Boubnov NV, Weaver DT: sc/d cells are deficient in Ku and replication protein A phosphorylation by the DNA-dependent protein kinase. Mol Cell Biol 1995, 15:5700-5706.
50.
Lees-Miller SP, Godbout R, Chan DW, Weinfeld M, Day RS, Barren GM, Allalunis-Tumer J: Absence of p350 subunlt of DNAactivated protein klnase from a radlosensitive human cell line. Science 1995, 267:1183-1185.
35.
Gottlieb TM, Jackson SP: Protein klnases and DNA damage. Trends Biochem Sci 1994, 19:500-503.
51.
36.
Lees-Miller SP, Sakaguchi K, UIIrich SJ, Appella E, Anderson CW: Human DNA-actlvated protein kinase phosphorylates serlnes
Li Z, Otevrel T, Gao Y, Chang H-L Seed B, Stamato TD, Taccioli GE, AIt FW: The XRCC4 9ene encodes a novel protein Involved in DNA double-stranded break repair and V(D)J recombination. Cell 1995, 83:1079-1089.
49. •
Mechanism of V(D)J recombination Molly Bogue and David B Roth* V(D)J recombination can be separated into two basic operations: DNA cleavage and joining of broken ends. Our understanding of both reactions has increased substantially in the past year. Major advances include the development of a cell-free system capable of cleavage and the identification of several proteins involved in both V(D)J recombination and double-strand break repair.
Address Department of Microbiology and Immunology,Baylor College of Medicine, ImmunologyM929, 1 Baylor Plaza, Houston, TX 77030, USA; *e-mail:[email protected] Current Opinion in Immunology 1996, 8:175-180 © Current Biology Ltd ISSN 0952-7915
Abbreviations D diversity DNA-PK DNA-dependentprotein kinase DSB double-strandbreak DNA-PKcs catalyticsubunit of DNA-dependent protein kinase J joining RSS recombinationsignal sequences SClD severe combined immunodeficiency TCR T-cell receptor V variable
observation that a certain RAG-1 mutant renders the recombination process hypersensitive to alterations in the coding nucleotides immediately adjacent to the RSS [6]. The joining mechanism was studied by analysis of mutations that interfere with junction formation. T h e mouse scid (severe combined immunodeficiency) mutation blocks the formation of coding joints and is characterized by accumulation of covalently sealed (hairpin) coding ends [3], which are normal intermediates in the reaction (see below). T h e murine scid mutation provided the first link between V(D)J recombination and DSB repair with the discovery that cells from SCID mice exhibit defective DSB repair [7-9]. Additional mutations that interfere with both V(D)J joining and DSB repair were identified subsequently [10,11]. T h e pace of discovery has accelerated in the past year, and much more is now known about both the cleavage and joining components of the V(D)J recombination mechanism. This review will focus on these exciting developments. The cleavage mechanism: breaks and hairpins In vivo studies
Introduction
V(D)J recombination assembles T C R and immunoglobulin genes from germline gene segments during development of precursor lymphocytes (reviewed in [1]). T h e recombination system recognizes conserved sequence elements, termed recombination signal sequences (RSS), that adjoin the V, D, and J coding elements. RSS consist of conserved heptamer and nonamer elements that are separated by 12 or 23 nucleotide spacer sequences. Efficient recombination requires a pair of RSS with different spacer lengths (the '12/23 rule'). Work during the early years of this decade provided several important clues about the recombination mechanism. Information about the cleavage reaction was provided by analysis of presumed in vivo recombination intermediates in precursor lymphocytes. These studies indicated that double-strand breaks (DSB) are introduced precisely between the RSS and the adjacent coding segment [2-5], producing two types of terminus, coding ends and signal ends, that join to form coding joints and signal joints (Fig. 1). A role for the lymphoid-specific recombination activating proteins, RAG-1 and RAG-2, in the cleavage reaction was suggested by the observation that both are required for formation of DSB at RSS in vivo [5]. T h e first evidence suggesting that RAG-1 might be directly involved in recognition of the RSS was provided by the
double-strand
As described above, previous characterization of site-specific DSB in DNA preparations from mouse thymus or bone marrow suggested that blunt signal ends and hairpin coding ends are recombination intermediates. However, the possibility that these molecules might represent dead-end products rather than true intermediates could not be excluded. This issue was recently addressed by studies using a pre-B cell line in which rearrangement of immunoglobulin light chain loci can be induced by growth at high temperature [12"]. Although coding ends are detected, they are rapidly joined. Signal ends accumulate at high temperature and can be chased into joined products upon return to low temperature. These observations provide compelling evidence that signal ends are authentic reaction intermediates. The detection of hairpin coding ends in vivo in the absence of the scid mutation strongly suggests that hairpin coding ends are normal reaction intermediates, a point which is further supported by biochemical experiments (see below). Insights into the mechanism of cleavage and hairpin formation were provided by analysis of the fine structure of signal and coding ends produced during T C R gene rearrangement in thymocytes of scid mice. Cleavage occurs precisely at the border between the RSS and the coding segment, producing a flush DSB at the signal end. T h e other product of cleavage is a coding end in
1 76 Lymphocyte development
Figure 1
(a) Cleavage
(b) Joining
C C ~ ~ " ~ T C : _ _
'~7 Join signal ends
~
OPeT~hairpi.~ns
---GGcc ~7 C C )
F°rm hairpins (TCG~
~
AG~ ~
-Signal joint
Process ends join coding ends
CCggTC GGccAG P nucleotides Coding joint
© 1996CurrentOpinionin Immunology A double-strand cleavage model for V(D)J recombination. The cleavage step is illustrated in (a). Recombination requires the presence of a pair of RSS containing 23 and 12 nucleotide spacer sequences (shown as filled and open triangles, respectively). Cleavage occurs in two steps, both of which are dependent upon the presence of RAG-1 and RAG-2 proteins [15"]. First, a nick is introduced precisely at the border between the RSS and the coding elements. (Coding segments are indicated by two nucleotides shown adjacent to each RSS. These particular nucleotides were chosen for purposes of illustration only.) The second step of the reaction, hairpin formation, involves attack of the 3' end of the coding segment on the opposite strand (small black arrows) [15°°]. Attack occurs precisely at the border between the RSS and the coding element, generating a blunt signal end and a hairpin structure that contains all nucleotides of the coding element [13°,15"]. The joining step is illustrated in (b). The blunt signal ends are joined to form a reciprocal product that contains a signal joint. Hairpin coding ends must be opened prior to joining. Presumably, the hairpin opening nuclease(s) (shown as scissors) recognize some aspect of the hairpin structure, such as distortions in DNA structure in the vicinity of the loop [4?], and introduces a nick near the terminus. Nicking at the terminus would generate a blunt end (shown on the right), whereas introduction of a nick near the terminus would produce a short single-stranded extension (shown on the left). Incorporation of the extension into the junction yields a specific type of extra nucleotides, termed P nucleotides, that are frequently found at coding joints [43,46]. Not shown are additional modifications, such as addition of N nucleotides by terminal deoxynucleotidyl transferase or loss of nucleotides from coding ends prior to joining (reviewed in [1]).
which theterminal nucleotides of each strand are ligated, producing a hairpin that contains all nucleotides of the coding segment [13°]. These results suggest a coupled mechanism in which hairpin coding ends and signal ends are both formed by the same cleavage event (Fig. 1).
Biochemical studies A major breakthrough in deciphering the mechanism of V(D)J recombination occurred in 1995, with the development of a cell-free system capable of performing the initial steps of the reaction, recognition and cleavage [14°°]. RAG-1 and RAG-2 are the only proteins required for cleavage, which produces blunt signal ends and hairpin coding ends of precisely the form detected in vivo [14"',15°']. These data also strongly suggest that hairpin coding ends are normal intermediates in V(D)J recombination. More detailed biochemical studies using oligonucleotide substrates revealed that cleavage can be separated into two steps: introduction of a single-stranded nick immediately 5' to the RSS; and conversion of this nick into a hairpin coding end, which liberates a blunt signal end (Fig. 1). Both steps require an RSS, RAG-l, and RAG-2 - - no additional proteins or high energy cofactors are required [15°°]. T h e possibility that RAG-1 and RAG-2 might func-
tion as a complex is supported by co-immunoprecipitation studies demonstrating an in vivo association between the two proteins [16].
Joining broken ends: connections between V(D)J recombination and DNA repair T h e realization that the mouse scid mutation affects both DSB repair and V(D)J recombination stimulated a search for additional mutations that might affect both reactions. Several mutants (comprising four complementation groups, termed XRCC4 through XRCC7) with defects in both DSB repair and V(D)J recombination have been identified (for review, see [17°,18"]). All mutants are defective for both signal and coding joint formation except murine scid (XRCC7, see below). XRCC5 mutants lack expression of a D N A end binding activity termed Ku [19,20], a heterodimer composed of 70kDa and 80kDa subunits that binds to altered DNA structures such as single-stranded gaps, broken ends, and hairpins [21-23]. In addition to its end-binding properties, Ku is a DNA helicase [24] and serves as the DNA-binding component of a DNA-dependent protein kinase, DNA-PK [21,25,26]. XRCC5 mutants do not produce a functional 80kDa Ku subunit, lack DNA-PK activity [27°], and are complemented by Ku80 eDNA [28"',29°,30°]. The sxi-1 cell line (XRCC6) exhibits defects in DSB repair, DNA end binding, and V(D)J recombination, and the DSB
Mechanism of V(D)J recombinationBogueand Roth 177
repair phenotype can be complemented by introduction of a cDNA encoding Ku70 [17",31]. T h e identification of the DNA-binding component of DNA-PK as a critical factor in both V(D)J recombination and DSB repair suggested that mutations in the catalytic subunit of the kinase, DNA-PKcs, might also affect these processes. This prediction was fulfilled by analysis of XRCC7 mutants (a group that includes the mouse scid mutation), which lack DNA-PK activity [32°°,33°°,34]. T h e murine scid defect has recently been mapped to DNA-PKcs [32°°,33"°]. It is remarkable that three of the four classes of mutation that affect V(D)J recombination and DSB repair also affect DNA-PK. T h e remaining gene, XRCC4, encodes a novel protein (see Note added on proof). When activated by binding to DNA lesions, DNA-PK is capable of phosphorylating a wide range of substrates in vitro, including several transcription factors, Ku, replication protein A, and p53 [34-36]. Possible roles of DNA-PK include signaling the presence of DNA damage, recruiting other components of DNA repair machinery, and inhibiting transcription in the vicinity of DNA lesions [18°,33",37,38]. Another possibility is that this large (465 kDa) protein [39°] might perform a scaffolding function, assembling components of the repair apparatus in the vicinity of DNA damage. Analysis of the nucleotide sequence of the human DNA-PKcs gene provided an unexpected clue about possible function. T h e carboxyl terminus contains a region that is homologous to the catalytic domains of proteins in the phosphatidylinositol-3 kinase superfamily [39°], which includes the AtaxiaTelangiectasia protein (ATM) and several other proteins that regulate cell-cycle progression (reviewed in [40]). Whether DNA-PK plays a similar role remains to be determined, although it is noteworthy that cell lines deficient for DNA-PK activity do not exhibit obvious checkpoint defects [41].
A w o r k i n g model for V(D)J r e c o m b i n a t i o n Although many questions remain, we have attempted to organize the recent wealth of information into a working model (Fig. 2). T h e first step involves recognition of RSS and cleavage by the RAG-1 and RAG-2 proteins, perhaps with the help of accessory (regulatory?) factors. After cleavage RAG-1 and RAG-2 may remain associated with the signal ends, protecting them from degradation and possibly facilitating joining [15°°]. T h e suggestion that all four ends might remain associated after cleavage is supported by the intriguing observation that the processing of coding ends may depend upon whether they are attached to an RSS with a 12 nucleotide or a 23 nucleotide spacer [42]. After cleavage, Ku binds to hairpin coding ends and signal ends and recruits DNA-PKcs, which somehow promotes the joining of hairpin coding ends.
T h e following observations must be kept in mind when considering the role of DNA-PK in coding joint formation. Hairpin coding ends accumulate to high levels in thymocytes of scid mice [3,13°]. However, the ability of scid cell lines to open exogenously introduced hairpin substrates is not detectably impaired [43,44]. One solution to this paradox is to propose that authentic hairpin coding ends are specifically protected by a DNA-protein complex that renders them unavailable for hairpin opening. Perhaps DNA-PKcs regulates accessibility of the hairpin ends in this complex to nucleases [13°,45]. This could be accomplished in several ways. Perhaps phosphorylation of substrate(s) serves to alter the complex, making the hairpins accessible to nuclease activities. Alternatively, phosphorylation events could serve to recruit a specific nuclease into the complex. DNA-PKcs could have functions in addition to kinase activity, particularly in view of its whopping size. The mechanism of hairpin opening remains uncharacterized. Analysis of coding junction sequences suggests that hairpin opening occurs within a few nucleotides of the terminus [1,43,46]. It is thought that hairpins are frequently opened to give ends with short single-stranded extensions, the presumed precursors of a particular class of junctional inserts termed P nucleotides (reviewed in [1]). T h e hairpin-opening nuclease could be a specialized enzyme that specifically recognizes hairpins or a more general enzyme that recognizes a variety of alternative DNA structures. Precedent for the latter possibility is provided by the observation that both mung bean nuclease and P1 nuclease, which recognize distortions in DNA structure, can efficiently open hairpins to leave short single-stranded extensions [47]. Opened coding ends then become accessible for modification, including loss of nucleotides and addition of extra nucleotides by terminal deoxynucleotidyl transferase, prior to joining (reviewed in [1]). Once hairpins are opened, they are joined much more rapidly than signal ends [12°]. Access of signal ends to modification appears to be much more restricted, perhaps because they remain associated with RAG-1 and RAG-2, as suggested above.
Conclusions and future prospects The past year has seen a number of exciting advances in our knowledge of the mechanism of V(D)J recombination. T h e initial step of the reaction, cleavage, is now amenable to detailed biochemical investigation, and the basic mechanism of cleavage should be elucidated in the near future. However, important aspects of the cleavage reaction as it occurs in vivo have not yet been reconstituted in cell-free systems. Several lines of evidence (discussed in [15°°]) suggest that in living cells cleavage occurs at pairs of signals, rather than individual signals. In contrast, in vitro cleavage at one RSS is not affected by the presence of a second RSS in the same substrate [14°°,15°°]. It will be interesting to see whether factors in addition to RAG-1 and RAG-2 are required for regulation of cleavage and
178 Lymphocyte development
Figure 2 A proposed role for DNA-PKcs: regulation of hairpin opening by controlling accessibility of hairpins to nucleases. We propose that RAG-1 and RAG-2 cleave at the RSS, and the resulting hairpin coding ends and blunt signal ends are bound by Ku. At this stage, it is possible that all four broken ends are in proximityto one another. The proposed DNA-protein complex renders the hairpin ends inaccessible to the hairpin-openingmachinery. DNA-PKcs (the murine scid factor) binds to Ku, is activated, and phosphorylates target molecules, resulting in increased accessibility of the coding ends to the hairpin opening machinery. Once accessible, hairpins are rapidly opened and joined.
M II
Cleavage at RSS Ku binds to broken ends
~
G
Hairpincoding end
I Hairpins inaccessible I
Ku 70•80 -1 + RAG-2
Bind and activate DNA-PKcs
~7
DNA-PKcs
Phosphorylationof target(s)
~7
IHairpins accessible I
Nuclease
Open hairpins join coding and signal ends
© 1996CurrentOpinioninImmunolog~ •
RSS
[]
for enforcement of the 12/23 rule. Other important levels of regulation include cell cycle control of recombinase activity (reviewed in [48"]) and substrate accessibility (reviewed in [1]).
Coding segments
0
Ku 70•80
O
RAG-1 + RAG-2
Although substantial progress has been made in identification of factors that participate in the joining reaction, the biochemical mechanisms of the steps following cleavage and hairpin formation remain undefined. Although the
Mechanism of V(D)J recombination Bogue and Roth
genes encoding three factors involved in the joining reaction have been identified, the XRCC4 gene remains to be characterized, and perhaps other complementation groups will be identified. More detailed analysis of some recently identified mutations in other species may be informative. For example, cells homozygous for the equine scid mutation lack detectable DNA-PK activity or immunoreactive DNA-PKcs protein [49"], and, unlike cells from scid mice, both coding and signal joint formation are severely affected [49"]. It will also be interesting to examine the effects of the recently described human cell line bearing an XRCC7 mutation [50] on coding and signal joint formation. Ultimately, biochemical experiments will be necessary to obtain detailed information about the joining mechanism. T h e next few years should prove to be exciting, as investigations on many fronts reveal more about the mechanism of V(D)J recombination and illuminate the connections between this lymphoid-specific reaction and general DNA repair mechanisms.
Note added in proof Since the submission of this review, a paper reporting that the XRCC4 gene encodes a novel protein has been published [51].
Acknowledgements The authors apologize to our colleagues whose work could not be cited due to space limitations. We are grateful to Martin Gellert and Katberyn Meek for communicating unpublished results. We thank Mark Landree, Sharon Roth and Sharri Steen for critical review of the manuscript. Mary Lowe-provided secretarial assistance. Work in the author's laboratory is supported by grants from the NIH (AI-36420) and the American Cancer Society (DB-118). DB Roth is a Charles E Culpeper Medical Scholar.
and a deficiency in DNA double-strand break repair. Proc Natl Acad Sci USA 1991, 88:1394-1397. 8.
FuIop GM, Phillips RA: The $cid mutation in mice causes a general defect in DNA repair. Nature 1990, 347:4?9-482.
9.
Hendrickson EA, Ctin X-Q, Bump EA, Schatz DG, Oettinger M, Weaver DT: A link between double-strand break-related repair and V(D)J recombination: the sc/d mutation. Proc Nat/Acad Sci USA 1991, 88:4061-4065.
10.
Pergola F, Zdzienicka MZ, Lieber MR: V(D)J recombination In mammalian mutants defective in DNA double-strand break repair. Mol Cell Bio/1993, 13:3464-3471.
11.
Taccioli GE, Rathbun G, Oltz E, Stamato T, Jeggo PA, AIt FW: Impairment of V(D)J recombination In double-strand break repair mutants. Science 1993, 260:207-210.
RamsdenDA, Gelled M: Formation and resolution of doublestrand break intermediates in V(D)J rearrangemenL Genes Dev 1995, 9:2409-2420. These elegant studies make use of a cell line with inducible V(D)J recombination activity to demonstrate that broken molecules are recombination intermediates. The first demonstration of hairpin coding ends in vivo in the absence of the scid mutation. 12.
•
13. •
Zhu C, Roth DB: Characterization of coding ends In thymocytes of scld mice: implications for the mechanism of V(D)J recombination. Immunity 1995, 2:101-112. A detailed investigation of the structure of coding and signal ends isolated from scid mouse thymocytes. The first analysis of hairpin coding ends at the nucleotide sequence level. The structures of ends produced in vivo are exactly the same as those produced by cleavage in cell-free systems. A hairpin accessibility model for scid defect is proposed. 14. ••
Van Gent DC, McBlane JF, Remsden DA, Sadofsky MJ, Hesse JE, Gelled M: Initiation of V(D)J recombination in a cell-free system. Cell 1995, 81:925-934. A landmark study, the first to demonstrate specific cleavage at RSS to give blunt signal ends and hairpin coding ends in a cell-frse system. Also showed that cleavage depends on both RAG-1 and RAG-2. 15. •,
McBlane JF, Van Gent DC, Ramsden DA, Romeo C, Cuomo CA, Gelled M, Oettinger MA: Cleavage at a V(D)J recombination signal requires only RAG-1 and RAG-2 proteins end occurs In two steps. Cell 1995, 83:387-395. Purified RAG-t and RAG-2 proteins and oligonucleotide substrates are used to dissect the details of the cleavage reaction. 16.
References and recommended reading Papers of particular interest, published within the annual period of review, have been highlighted as: * of special interest == of outstanding interest 1.
Lewis SM: The mechanism of V(D)J joining: lessons from molecular, immunological and comparative analyses. Adv Immunol 1994, 56:27-150.
2.
Roth DB, Nakajima PB, Menetski JP, Bosma MJ, Gelled M: V(D)J recombination in mouse thymocytes: double-strand breaks near T cell receptor rearrangement signals. Cell 1992, 69:41-53.
3.
Roth DB, Menetski JP, Nakajima PB, Bosma MJ, Gelled M: V(D)J recombination: broken DNA molecules with covalentiy sealed (hairpin) coding ends in scid mouse thymocytes. Cell 1992, 70:983-991.
4.
Roth DB, Zhu C, Gelled M: Characterization of broken DNA molecules associated with V(D)J recombination. Proc Nat/Acad Sci USA 1993, 90:10788-10792.
5.
Schlissel M, Constantinescu A, Morrow T, Baxter M, Pang A: Double-strand signal sequence breaks in V(D)J recombination are blunt, 5'-phosphorylated, RAG-dependent, and cell cycle regulated. Genes Dev 1993, 7:2520-2532.
6.
Sadofsky MJ, Hesse JE, Van Gent DC, Gelled M: RAG1 mutations that affect the target specificity of V(D)J recombination: a possible direct role of RAG-1 in site recognition. Genes Dev 1995, 9:2193-2199.
7.
Biedermann KA, Sun J, Giaccia AJ, Tosto LM, Brown JM: scid mutation in mice confers hypersensitivity to Ionizing radiation
179
Leu TM, Schatz DG: RAG-1 and RAG-2 are components of a high-molecular-weight complex, and association of RAG-2 with this complex is RAG-1 dependent. Mol Cell Biol 1995, 15:5657-5670.
17. An
Weaver DT: What to do at an end: DNA double-strand-break repair. Trends Genet 1995, 11:388-392. excellent recent review of of factors involved in both DSB repair and V(D)J recombination. Summarizes very recent data on XR-1 (XRCC4), sxi-1 (XRCC6), and sxi-2/sxi-3 (XRCCS) mutants. 18. •
JacksonSP, Jeggo PA: DNA double-strand break repair and V(D)J recombination: involvement of DNA-PK. Trends Biochem Sci 1995, 20:412-415. Very up-to-date review, with emphasis on role of DNA-PK. 19.
Getts RC, Stamato TD: Absence of a Ku-Ilke DNA end binding activity in the xrs double-strand DNA repair-deficient mutant, J Biol Chem 1994, 269:15981-15984.
20.
RathmellWK, Chu G: A DNA end-binding factor involved in double-strand break repair and V(D)J recombination. Mol Cell Biol 1994, 14:4741-4?48.
21.
Morozov VE, Falzon M, Anderson CW, Kuff EL: DNA-dependent protein kinase is activated by nicks and larger single-stranded gaps. J Biol Chem 1994, 269:16684-16688.
22.
Blier P, Griffith AJ, Craft J, Hardin JA: Binding of Ku protein to DNA: measurement of affinity for ends and demonstration of binding to nicks. J Biol Chem 1993, 268:7594-7601.
23.
Palllard S, Strauss F: Analysis of the mechanism of interaction of simian Ku protein with DNA. Nucleic Acids Res 1991, 19:5619-5624.
24.
Tuteja N, Tuteja R, Ochem A, Taneja P, Huang NW, Simoncsits A, Susic S, Rahman K, Marusic L, Chen J et el.: Human DNA heiicase II - a novel DNA unwinding enzyme identified as the Ku autoantigen. EMBO J 1994, 13:4991-5001.
25.
Gottlieb TM, Jackson SP: The DNA-dependent protein Idnase: requirement for DNA ends and association with Ku antigen. Cell 1993, 72:131-142.
180
Lymphocyte development
26.
Dvir A, Peterson SR, Knuth MW, Lu H, Dynan WS: Ku autoantlgen Is the regulatory component of a templateassociated protein kinase that phosphorylates RNA polymerase II. Proc Nat/Acad Sci USA 1992, 89:11920-11924.
27. •
Rnnie NJ, Gottlieb TM, Blunt T, Jeggo PA, Jackson SP: DNAdependent protein kinase activity Is absent in xrs-6 cells: implications for site-specific recombination and DNA doublestrand break repair. Proc Nat/Acad Sci USA 1995, 92:320-324. A sensitive new assay for DNA-PK activity is used to show that xrs-6 cells (XRCC5 mutants) which lack Ku end-binding, also lack DNA-PK activity. 28. ••
Taccioli GE, Gottlieb TM, Blunt T, Priestley A, Demengsot J, Mizuta R, Lehmann AR, AIt FW, Jackson SP, Jeggo PA: Ku80: product of the XRCC5 9ene and its role in DNA repair and V(D)J recombination. Science 1994, 265:1442-1445. First report identifying XRCC5 as Ku80. A cDNA encoding human Ku80 complements the V(D)J recombination, DNA end binding, and X-ray sensitivity phenotypes of XRCC5 mutant hamster cells. 29. •
Boubnov NV, Hall KT, Wills Z, Lee SE, He DM, Benjamin DM, Pulaski CR, Band H, Reeves W, Hendrickson EA, Weaver DT: Complementetion of the Ionizing radiation sensitivity, DNA end binding, and V(D)J recombination defects of double-strand break repair mutants by the p86 Ku autoanflgen. Proc Natl Acad Sci USA 1995, 92:890-894. Two new mutants with defects in DSB repair and V(D)J recombination, sxi-2 and sxi-3, are found to be in the XRCC5 group. Complemantation with a cDNA encoding Ku80 restores defects in V(D)J recombination and DNA repair.
15 and 37 in the amino-terminal transactlvation domain of human p53. Mo/Cell Bio11992, 12:5041-5049. 37.
Kuhn A, Gottlieb TM, Jackson SP, Grummt i: DNA-dependent protein kinase: a potent Inhibitor of transcription by RNA polymerase I. Genes Dev 1995, 9:193-203.
38.
Labhart P: DNA-dependent protein klnase specifically represses promoter-dlracted transcription Initiation by RNA polymerase I. Proc Nat/Acad Sci USA 1995, 92:2934-2938.
39. •
Hartley KO, Gell D, Smith GCM, Zhang H, Divecha N, Connelly MA, Admon A, Lees-Miller SP, Anderson CW, Jackson SP: DNAdependent protein kinase catalytic subunit: a relative of phosphatidyllnositol 3-klnase and the ataxla telanglectasla 9ene product. Cell 1995, 82:849-856. The sequence of the gene encoding DNA-PKcs reveals homology to a family of kinases related to phosphatidylinositol-3 kinase. No lipid kinase activity was detected, indicating that the catalytic function of DNA-PKcs may be restricted to protein kinase activity. 40.
Keith CT, Schreiber SL: PlK-ralatad kinases: DNA repelr, recombination, and cell cycle checkpoints. Science 1995, 270:50-51.
41.
Jeggo PA: X-ray sensitive mutants of Chinese hamster ovary cell line: radio-sensitivity of DNA synthesis. Mutation Res 1985, 145:171-176.
42.
Smider V, Rathmell WK, Lieber MR, Chu G: Restoration of X-ray resistance and V(D)J recombination In mutant cells by Ku cDNA. Science 1994, 266:288-291. Complementation with a cDNA encoding Ku80 restores defects in V(D)J recombination and DNA repair in XRCC5 mutant cells.
Ezekiel UR, Engler P, Stern D, Storb U: Asymmetric processing of coding ends and the effect of coding end nucleotlde composition on V(D)J recombination. Immunity 1995, 2:381-389.
43.
Lewis SM: P nucleotide insertions and the resolution of hairpin DNA structures in mammalian calls. Proc Natl Acad Sci USA 1994, 91:1332-1336.
31.
44.
Staunton JE, Weaver DT: scld cells efficiently Integrate hairpin and linear DNA substrates. Mol Cell Bio11994, 14:3876-3883.
45.
Roth DB, Lindahl T, Gellert M: How to make ends meet. Curr Biol 1995, 5:496-499.
46.
Meier JT, Lewis SM: P nucleotldes in V(D)J recombination: a fine-structure analysis. Mol Cell Biol 1993, 13:1078-1092.
47.
Kabotyanski EB, Zhu C, Roth DB: Hairpin opening by single-strand-specific nucleases. Nucleic Acids Res 1995, 23:3872-3881.
30. •
Lee SE, Pulaski CR, He DM, Benjamin DM, Voas M, Um J, Hendrickson EA: Isolation of mamma,an call mutants that are X-ray sensitive, impaired in DNA double-strand break repair and defective for V(D)J recombination. Murat Res 1995, 336:279-291.
32. ••
Kirchgassner CU, Patil CK, Evans JW, Cuomo CA, Fried LM, Carter T, Oettinger MA, Brown JM: DNA-dependent klnase (p350) as a candidate gene for the murine scld defect. Science 1995, 267:1178-1183. Positional cloning and complementation studies identified DNA-PKcs as a candidate for the murine scid gene. Analysis of DNA-PKcs protein levels by western blotting using two monoclonal antibodies revealed low levels of protein, suggesting that the mutation may cause decreased synthesis or decreased stability of the DNA-PKcs protein. 33. ••
48. Lin W-C, Desiderio S: V(D)J recombination and the cell cycle. • Immunol Today 1995, 16:279-289. Comprehensive review of regulation of V(D)J recombination activity, with emphasis on call-cycle regulation.
Blunt T, Finnio NJ, Taccioli GF_,Smith GCM, Demengeot J, Gottlieb TM, Mizuta R, Varghase AJ, AIt FW, Jeggo PA, Jackson SP: Defective DNA-dependent protein klnase activity is linked to V(D)J recombination and DNA repair defects associated with the murlne scld mutation. Cell 1995, 80:813-823. XRCC7 mutants lack DNA-PK activity due to deficiency in the catalytic subunit. Beth DSB repair and V(D)J recombination defects were complemented by yeast artiFmialchromosomes encoding DNA-PKcs.
Wiler R, Leber R, Moore BB, VanDyk LF, Perryman LE, Meek K: Equine severe combined Immunodefidency: a defect In V(D)J recombination and DNA-dependent protein Idnase activity. Proc Nat/Acad Sci USA 1995, 92:11485-11489. Defective DNA-PK activity in Arabian foals with autosomal recessive severe combined immunodeficiency is linked to defects in both coding and signal joint formation.
34.
Boubnov NV, Weaver DT: sc/d cells are deficient in Ku and replication protein A phosphorylation by the DNA-dependent protein kinase. Mol Cell Biol 1995, 15:5700-5706.
50.
Lees-Miller SP, Godbout R, Chan DW, Weinfeld M, Day RS, Barren GM, Allalunis-Tumer J: Absence of p350 subunlt of DNAactivated protein klnase from a radlosensitive human cell line. Science 1995, 267:1183-1185.
35.
Gottlieb TM, Jackson SP: Protein klnases and DNA damage. Trends Biochem Sci 1994, 19:500-503.
51.
36.
Lees-Miller SP, Sakaguchi K, UIIrich SJ, Appella E, Anderson CW: Human DNA-actlvated protein kinase phosphorylates serlnes
Li Z, Otevrel T, Gao Y, Chang H-L Seed B, Stamato TD, Taccioli GE, AIt FW: The XRCC4 9ene encodes a novel protein Involved in DNA double-stranded break repair and V(D)J recombination. Cell 1995, 83:1079-1089.
49. •