| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
From the Departments of Medicine and Biochemistry, Stanford University School of Medicine, Stanford, California 94305
INTRODUCTION
Mutant Cell Lines
Ku and DNA-dependent Protein Kinase
Genes for Double Strand Break Repair
DNA End Joining in Intact Cells
DNA End Joining versus Homologous Recombination
DNA End Joining in Cell-free Extracts
Model for the End-joining Reaction
Other Roles for DNA End-joining Proteins
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
INTRODUCTION
DNA double strand breaks
(DSBs)1 may be the most
disruptive form of DNA damage. If left unrepaired, they lead to broken
chromosomes and cell death. If repaired improperly, they can lead to
chromosome translocations and cancer. Humans are at risk for DSBs from
exogenous agents. The paradigm agent, ionizing radiation, is present in the environment mainly from the decay of radon gas, which accumulates in homes to different levels depending on the uranium content of the
underlying soil. Ionizing radiation is also utilized in medicine for
diagnostic x-rays and for treating cancer patients. Anticancer drugs
will generate DSBs as well: bleomycin produces oxidative free radicals,
which induce strand breaks; etoposide and adriamycin inhibit
topoisomerase II to create protein-bridged DSBs. Humans are also at
risk for DSBs from endogenous agents. Oxidative metabolism generates
free radicals and subsequent strand breaks. V(D)J recombination
generates DSBs during rearrangement of genes encoding B cell
immunoglobulins and T cell receptors (2).
In response to the threat of DSBs, cells have evolved at least two independent pathways for repairing DSBs, by homologous recombination or by nonhomologous DNA end joining. This review will focus on recent progress in understanding the dominant mechanism in mammalian cells, nonhomologous end joining.
Mutant Cell Lines
Immunological diversity is generated by V(D)J recombination (Fig.
1), a site-specific cleavage of the
chromosome followed by an end-joining reaction to bring the free DNA
ends together (2). The possibility that DSB repair and V(D)J
recombination might share the same biochemical pathway was first
recognized in the severe combined immunodeficiency (scid)
mouse. The scid mouse lacks mature B and T cells due to a
defect in V(D)J recombination (3) and is also hypersensitive to
ionizing radiation due to a defect in DSB repair (4-6). Screening of
other x-ray-sensitive cell lines (Table
I) led to the identification of three
genetic complementation groups with defects in both DSB repair and
V(D)J recombination (7, 8). Since efforts were under way to complement these rodent cell lines with human genes, the genes for the
complementation groups were designated XRCC, for x-ray
cross-complementing. XRCC7 mutant cells (scid
mouse cells and V3 hamster cells) were severely defective in coding
joint formation but only mildly defective in signal joint formation. By
contrast, XRCC4 and XRCC5 mutant cells were
severely defective in both coding and signal joint formation.
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Ku and DNA-dependent Protein Kinase
The path toward isolating the XRCC5 and XRCC7 genes began with identification of the Ku protein as an autoantigen in scleroderma-polymyositis overlap syndrome (9, 10). In the hope of gaining insight into the pathogenesis of autoimmune disease, Ku was extensively characterized.
Ku is a heterodimer of 70 and 86 kDa (Ku70 and Ku86) that binds with
strong affinity to DNA ends, stem-loop or bubble structures, or
transitions between double-stranded DNA and two single strands (11,
12). Once Ku binds to DNA, it can translocate along the DNA, so that
three or more Ku molecules can bind to a single linear DNA fragment
(13, 14). If the linear DNA is then ligated into a circle, Ku can no
longer dissociate, consistent with a model in which Ku molecules bind
to DNA like beads on a string. Ku is a DNA-dependent ATPase
that is activated by both double- and single-stranded DNA (15). Both
Ku70 and Ku86 contain motifs for potential ATP binding sites. Ku has
also been reported to have an ATP-dependent 3
to 5
helicase activity (16).
Ku is the regulatory subunit for DNA-dependent protein kinase (DNA-PK), which has the unusual property of remaining quiescent until activated by DNA ends (17). DNA-PK contains an enormous catalytic subunit of 465 kDa (DNA-PKcs) that is activated when Ku binds to DNA. DNA-PK will phosphorylate serine or threonine residues that immediately precede glutamine in a large number of protein substrates in vitro (18). Interestingly, DNA-PK will phosphorylate Ku in vitro, activating its ATPase activity (15). However, attempts to identify in vivo substrates have been inconclusive. For example, DNA-PK will phosphorylate p53 in vitro but is not required for the accumulation or activation of p53 (19, 20).
Genes for Double Strand Break Repair
A DNA end binding (DEB) factor was detected by an electrophoretic mobility shift assay (21). This assay was used to screen a large number of x-ray-sensitive cell lines, and DEB factor was absent in three different cell lines, all belonging to the complementation group for XRCC5 (21). DEB factor proved to be both biochemically and antigenically similar to Ku (22, 23). Furthermore, the XRCC5 and Ku86 genes were independently mapped to the same human chromosome locus, 2q33-35 (24-26).
To demonstrate that Ku86 and XRCC5 are identical, an expression vector for Ku86 was transfected into the mutant XRCC5 hamster cells. Transfection of human Ku86 rescued the mutant hamster cells for DEB activity, DNA-PK enzymatic activity, x-ray resistance, and V(D)J recombination (27, 28). Transfection of Ku86 also restored resistance to the topoisomerase II inhibitor etoposide (29). Thus Ku is involved in the repair of DNA DSBs produced by ionizing radiation, V(D)J recombination, or etoposide.
XRCC5 cells were found to contain mutations in the Ku86 gene (30, 31), leading to functionally significant alterations in Ku86 (Table I). As expected, the generation of Ku86 knockout mice produced severe immunodeficiency due to an absence of both T and B cells (32, 33). Both coding and signal joint formation were impaired, and the coding joint defect was accompanied by accumulation of hairpin coding ends in Ku86 knockout thymocytes. Unexpectedly, the Ku86 knockout mice also displayed marked growth retardation (32), suggesting that Ku might have an additional unforseen role.
The discovery that Ku86 was defective in XRCC5 mutant cells raised the possibility that DNA-PKcs might be defective in XRCC7 cells (34-36). Indeed, DNA-PK enzymatic activity and DNA-PKcs protein levels were severely reduced in both scid and V3 cells. When genomic yACs containing the DNA-PKcs gene were transfected into scid and V3 cells, DNA-PK enzymatic activity, ionizing radiation resistance, and coding joint formation were restored. Furthermore, the DNA-PKcs gene was mutated in scid cells, producing a premature termination codon in the putative kinase domain that truncates the C-terminal 83 amino acids (37, 38).
As in Ku86 knockout thymocytes, scid thymocytes accumulate hairpin coding ends (33). Therefore, Ku and DNA-PK are required for the proper processing of hairpin ends. Ku or DNA-PK do not have an endonuclease activity that opens the hairpin ends directly. Instead, DNA-PK enzymatic activity may make the hairpin accessible to a still unidentified hairpin endonuclease.
The only known XRCC4 cell line, XR-1, is rescued by transfection of a cDNA encoding a 37-kDa protein, restoring both ionizing radiation resistance and V(D)J recombination to wild-type levels (39). XRCC4 is a novel gene, lacking homology to other known genes. It is deleted in XR-1 cells and therefore not essential for growth. The sensitivity of XR-1 cells increases dramatically in the G1 phase of the cell cycle (40), and any proposed role for XRCC4 must account for this phenomenon.
DNA End Joining in Intact Cells
End joining of naked DNA has been studied by introducing linear
DNA into intact cells. When linearized plasmid DNA was injected into
Xenopus oocytes, a low level of plasmid recircularization occurred (41). The end-joining reaction resulted in junctions containing deletions back to regions of microhomology of 1-10 bases
within 20 bp of the ends. When linearized SV40 DNA was transfected into
mammalian cells, end joining either occurred by direct joining of the
ends or by the deletions back to regions of microhomology of 1-6 bases
(42, 43). Direct joining occurred even for ends with protruding ends of
opposite polarity (5 protruding and 3 protruding). The joining events
utilizing microhomology were proposed to be directed by base pairing of
the microhomology regions (Fig. 2), thus
aligning the DNA ends for subsequent steps in the end-joining reaction.
Interestingly, end joining also occurred occasionally with the
insertion of nucleotides from a number of sources, including free
oligonucleotides in the nucleus, copying errors near the DSB by slipped
mispairing and repair synthesis, and misincorporation of A residues
during repair of a 5 single-stranded extension (44).
To study end joining of chromosomal DNA, restriction enzymes have been used to introduce DSBs at specific chromosomal sites. Restriction enzymes were electroporated into Chinese hamster ovary cells, which are hemizygous at the APRT (adenine phosphoribosyltransferase) locus (45). Colonies of viable cells containing mutations in APRT were then analyzed at the DNA sequence level. Mutations consisted of insertions, small deletions up to 36 bp, and combinations of insertions and deletions at the cleavage sites. Most of the deletions involved the utilization of microhomology of 1-4 bases at the recombination junctions. Because this method scored for induced mutations in APRT, the full spectrum of end-joining events could not be observed. Nevertheless, the end-joining reactions observed with transfected naked DNA and chromosomal DNA shared key characteristics: nucleotide insertions and nucleotide deletions directed by microhomology.
Rare cutting restriction enzymes have been used to introduce a single or very few DSBs into the mammalian genome (46). Chinese hamster ovary cells were engineered to contain a single I-SceI site disrupting an integrated neo gene. Repair of the induced DSB reconstructed an intact neo gene if the end joining occurred with deletion of the ends back to a region of 4-base microhomology. In this system, XRCC5 mutant cells were severely defective in rejoining the DSB by at least 3 orders of magnitude compared with wild-type hamster cells.
V(D)J recombination junctions provide additional insight into end
joining of broken chromosomes (2). Coding joint formation occurs with
both insertion and deletion of nucleotides. Many insertions occur by
mechanisms specific for lymphoid cells. Palindromic (P-nucleotide) insertions are generated by the asymmetric cleavage of the hairpin coding ends. Nontemplated (N-nucleotide) insertions are catalyzed by
lymphoid-specific terminal deoxynucleotidyltransferase (TdT). However,
nucleotide insertion will still occur in TdT knockout mice (47).
Nucleotide deletion is also best studied in TdT knockout mice, since
TdT-catalyzed insertions can obscure the nature of the deletions. In
the absence of TdT, 75% of the joining events contain deletions
directed by microhomology regions of 1-5 nucleotides (47, 48). Thus,
the joining reaction for coding ends during V(D)J recombination
confirms the characteristics observed for random DSBs in nonlymphoid
cells. The proposed DNA intermediates for the end-joining reaction are
shown in Fig. 3.
In contrast to coding joint formation, signal joint formation is precise and depends on Ku but not on an intact kinase domain in DNA-PKcs. Thus, the mechanisms that lead to the insertion and deletion of nucleotides are somehow suppressed during the formation of signal ends.
DNA End Joining versus Homologous Recombination
Cells are capable of repairing DSBs either by homologous recombination or by nonhomologous end joining. Genetic screens suggest that the relative importance of these two mechanisms is quite different in yeast and mammalian cells. In Saccharomyces cerevisiae, screens for mutants sensitive to ionizing radiation have produced many mutant alleles of genes involved in homologous recombination (49), but not genes involved in nonhomologous end joining. Furthermore, yeast knocked out for the genes homologous to mammalian Ku have normal sensitivity to ionizing radiation (50-53). Nevertheless, yeast Ku is involved in DSB repair. Yeast Ku70 mutants fail to recircularize linearized plasmids (50, 52). They also fail to survive when induced to express EcoRI, which causes DSBs at the same position in both sister chromatids, precluding repair by homologous recombination (51). Furthermore, when homologous recombination is eliminated in a rad52 mutant, disruption of yeast Ku70 confers additional radiation hypersensitivity (53). Thus, S. cerevisiae repairs DSBs preferentially by homologous recombination, while using Ku-dependent end joining as a secondary pathway when homologous recombination is not possible.
By contrast, mutant screens in mammalian cells for sensitivity to ionizing radiation have yielded mutant alleles for Ku and DNA-PKcs but not alleles for genes involved in homologous recombination. When cells are engineered with DNA chromosomal substrates that permit the repair of DSBs by either homologous recombination or by nonhomologous end joining, the great majority of the joining events occurs by nonhomologous end joining (54-57). Thus, in contrast to yeast, mammalian cells repair DSBs preferentially by nonhomologous end joining.
DNA End Joining in Cell-free Extracts
Progress toward defining nonhomologous end joining in a cell-free
system has been relatively slow. End joining was studied in
Xenopus extracts by recircularization of linear DNA
molecules with different types of DNA ends (58, 59). In contrast to results in intact mammalian cells, nucleotide deletion was not observed
beyond losses from protruding single strands. Short regions of
microhomology were utilized but only if the homologous base pairs were
available within the protruding single-stranded regions of the same
polarity. Interestingly, substrates with opposite polarities (5
protruding and 3 protruding ends) were joined by the alignment of
single strands, gap filling, and ligation. Such an unusual activity was
found for Klenow fragment of DNA polymerase I from Escherichia
coli (60), suggesting that a vertebrate polymerase might
accomplish the same thing.
End joining in mammalian extracts has been studied by the recircularization of pUC18 plasmid DNA cleaved with different restriction enzymes. Almost all events involved direct joining of the cohesive ends, but there were a small number of misrejoined molecules with insertions or deletions utilizing regions of microhomology. By making DSBs with noncohesive ends, it was possible to fractionate extracts into four different end-joining activities (61). The activity that utilized microhomology near the ends could be purified extensively (62). However, none of the activities was dependent on the presence of Ku protein, suggesting that this cell-free system had detected a secondary end-joining mechanism.
There is reason for optimism that Ku-dependent end joining can be studied in cell-free systems for the special case of signal joint formation (63). When a substrate containing V(D)J recombination signal sequences was incubated with purified RAG1 and RAG2 and HeLa extracts, the substrate was cleaved and the signal ends were joined with appropriate precision. Furthermore, signal joining could be blocked by anti-Ku antibodies. This result, while promising, must be verified with extracts from Ku-deficient cells.
Model for the End-joining Reaction
The properties of Ku and DNA-PK can be used to formulate a model
for how these proteins might facilitate DNA end joining (Fig. 4). When the chromosome is broken, Ku
protects the DNA ends from degradation until end joining is completed.
In support of this, when Ku86 mutant cells were tested for
V(D)J recombination, the few signal joining events that were recovered
contained large deletions (8). Similarly, when Ku mutant cells were
transfected with linearized plasmid DNA, they exhibited excess
degradation of the DNA ends (64). Finally, purified Ku protein
protected free DNA ends from exonuclease degradation in
vitro (23).
On binding to the DNA ends, Ku recruits DNA-PKcs. Synapsis of the two DNA ends might be mediated by two independent DNA binding sites on DNA-PK or by the association of two DNA-PK molecules.
DNA-PK phosphorylates both DNA-PKcs and Ku subunits in vitro (65). We hypothesize that this phosphorylation occurs in trans, so that the kinase assembled on one DNA end phosphorylates the DNA-PK subunits assembled on the other DNA end. Activity in trans would regulate the kinase activity so that processing of the DNA ends would occur only after synapsis of the two ends is accomplished.
Upon phosphorylation by DNA-PK in vitro, DNA-PKcs dissociates from Ku (65), and Ku acquires helicase activity (15, 16). We postulate that this helicase activity unwinds DNA ends in vivo so that exposed regions of microhomology can anneal by base pairing. The unpaired DNA flaps could then be removed either by an exonuclease or a flap endonuclease. A flap endonuclease, FEN-1, has the appropriate enzymatic activity (66), although its role in DSB repair and V(D)J recombination has not been established. Gaps are then filled in by a DNA polymerase and the nicks sealed by ligase to complete the end-joining process.
The hypothesis that DNA-PK phosphorylates in trans has implications for V(D)J recombination. Recall that the joining of signal ends occurs precisely and does not require intact kinase activity. When a pair of signal and coding ends is created by RAG1/RAG2 cleavage, DNA-PK would assemble on the signal end and phosphorylate Ku bound to the opposing coding end, activating the Ku helicase. The hairpin end would then be processed to allow cleavage by a presumed hairpin endonuclease. The reverse reaction is quite different: DNA-PK assembled on the hairpin end remains inactive as a kinase in vitro. Thus, Ku bound to the signal end would remain unphosphorylated, preventing processing of the signal ends. Therefore, the transphosphorylation hypothesis explains why the joining of signal ends is precise and why it is unaffected by the scid mutation.
To test this model, experiments are needed to determine the effect of abrogating Ku helicase activity. Of note, mutation of the conserved ATP binding site in Ku86 has no effect (31), but a similar experiment has not been performed for the ATP binding site in Ku70. Experiments must also determine whether the helicase activity of Ku will unwind DNA ends and whether Ku is an in vivo substrate for DNA-PK. A knockout of the entire DNA-PKcs gene is needed to determine whether the enormous catalytic subunit has additional biochemical activities. To complete the model, many questions must be answered. What is the role of the XRCC4 protein in the end-joining reaction? What protein has hairpin endonuclease activity? What proteins are required for alignment of the ends by microhomology base pairing? What proteins process the unpaired DNA flaps after alignment?
Other Roles for DNA End-joining Proteins
The pathway for nonhomologous DNA end joining has evolved for other roles beyond its primary role in joining broken chromosomes. It has been hypothesized that RAG1 and RAG2 were originally introduced into the mammalian genome by an ancient retroviral infection (67). V(D)J recombination could then have evolved by combining the cleavage activity of RAG1/RAG2 with the pre-existing Ku-dependent end-joining activity for broken chromosomes. Such a dual use of end joining is also seen in Drosophila melanogaster, which utilizes Ku for both P-element transposition and DNA repair (68).
Mammalian cells may have additional roles for Ku. Ku86 knockout mice display marked growth retardation (32), suggesting a role for Ku other than in DNA repair or V(D)J recombination. Such roles include involvement in transcription by RNA polymerase I (69, 70) or RNA polymerase II (12, 71). Data from yeast suggest roles in regulating replication and in maintaining telomere length (51).
In conclusion, the study of how cells repair double strand breaks has revealed a primordial end-joining mechanism. Further study promises to illuminate not just the repair of broken chromosomes but other critical pathways in DNA metabolism.
FOOTNOTES
To whom correspondence should be addressed: M211, Dept. of
Medicine (Oncology), Stanford University Medical Center, Stanford, CA
94305-5306. Tel.: 415-725-6442; Fax: 415-725-1420; E-mail: chu{at}cmgm.stanford.edu.
ACKNOWLEDGEMENTS
I thank Kim Rathmell, Vaughn Smider, and Ola Hammarsten for helpful discussions and critical reading of the manuscript.
REFERENCES
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  M. R. Lieber The Mechanism of Human Nonhomologous DNA End Joining J. Biol. Chem., January 4, 2008; 283(1): 1 - 5. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. MARTIEN and C. ABBADIE Acquisition of Oxidative DNA Damage during Senescence: The First Step toward Carcinogenesis? Ann. N.Y. Acad. Sci., November 1, 2007; 1119(1): 51 - 63. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Windbichler, P. A. Papathanos, F. Catteruccia, H. Ranson, A. Burt, and A. Crisanti Homing endonuclease mediated gene targeting in Anopheles gambiae cells and embryos Nucleic Acids Res., September 27, 2007; 35(17): 5922 - 5933. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Shi, D. Qiu, G. Zhao, B. Corthesy, S. Lees-Miller, W. H. Reeves, and P. N. Kao Dynamic binding of Ku80, Ku70 and NF90 to the IL-2 promoter in vivo in activated T-cells Nucleic Acids Res., April 1, 2007; 35(7): 2302 - 2310. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I.-Y. Chang, C.-K. Youn, H.-B. Kim, M.-H. Kim, H.-J. Cho, Y. Yoon, Y.-S. Lee, M.-H. Chung, and H. J. You Oncogenic H-Ras Up-regulates Expression of Ku80 to Protect Cells from {gamma}-Ray Irradiation in NIH3T3 Cells Cancer Res., August 1, 2005; 65(15): 6811 - 6819. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. A. Smith, B. C. Waldman, and A. S. Waldman A Role for DNA Mismatch Repair Protein Msh2 in Error-Prone Double-Strand-Break Repair in Mammalian Chromosomes Genetics, May 1, 2005; 170(1): 355 - 363. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. D. Kennedy, J. E. Quinn, P. B. Mullan, P. G. Johnston, and D. P. Harkin The Role of BRCA1 in the Cellular Response to Chemotherapy J Natl Cancer Inst, November 17, 2004; 96(22): 1659 - 1668. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Adachi, S. So, and H. Koyama Loss of Nonhomologous End Joining Confers Camptothecin Resistance in DT40 Cells: IMPLICATIONS FOR THE REPAIR OF TOPOISOMERASE I-MEDIATED DNA DAMAGE J. Biol. Chem., September 3, 2004; 279(36): 37343 - 37348. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. E. Liberti and L. J. Rasmussen Is hEXO1 a Cancer Predisposing Gene? Mol. Cancer Res., August 1, 2004; 2(8): 427 - 432. [Full Text] [PDF]
|
|
|
|
|
|
M. Kampmann and D. Stock Reverse gyrase has heat-protective DNA chaperone activity independent of supercoiling Nucleic Acids Res., July 6, 2004; 32(12): 3537 - 3545. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Yang, Y. Yu, H. E. Hamrick, and P. J. Duerksen-Hughes ATM, ATR and DNA-PK: initiators of the cellular genotoxic stress responses Carcinogenesis, October 1, 2003; 24(10): 1571 - 1580. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Adachi, H. Suzuki, S. Iiizumi, and H. Koyama Hypersensitivity of Nonhomologous DNA End-joining Mutants to VP-16 and ICRF-193: IMPLICATIONS FOR THE REPAIR OF TOPOISOMERASE II-MEDIATED DNA DAMAGE J. Biol. Chem., September 19, 2003; 278(38): 35897 - 35902. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N.G. Oliveira, M. Castro, A.S. Rodrigues, I.C. Goncalves, O.M. Gil, A.P. Fernandes, J.M. Toscano-Rico, and J. Rueff Wortmannin enhances the induction of micronuclei by low and high LET radiation Mutagenesis, January 1, 2003; 18(1): 37 - 44. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Sak, M. Stuschke, R. Wurm, G. Schroeder, B. Sinn, G. Wolf, and V. Budach Selective Inactivation of DNA-dependent Protein Kinase with Antisense Oligodeoxynucleotides: Consequences for the Rejoining of Radiation-induced DNA Double-Strand Breaks and Radiosensitivity of Human Cancer Cell Lines Cancer Res., November 15, 2002; 62(22): 6621 - 6624. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. C. Stern, D. M. Umbach, R. M. Lunn, and J. A. Taylor DNA Repair Gene XRCC3 Codon 241 Polymorphism, Its Interaction with Smoking and XRCC1 Polymorphisms, and Bladder Cancer Risk Cancer Epidemiol. Biomarkers Prev., September 1, 2002; 11(9): 939 - 943. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Q. Zhong, C.-F. Chen, P.-L. Chen, and W.-H. Lee BRCA1 Facilitates Microhomology-mediated End Joining of DNA Double Strand Breaks J. Biol. Chem., August 2, 2002; 277(32): 28641 - 28647. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Mori, S. M. Colman, Z. Xiao, A. M. Ford, L. E. Healy, C. Donaldson, J. M. Hows, C. Navarrete, and M. Greaves Chromosome translocations and covert leukemic clones are generated during normal fetal development PNAS, June 11, 2002; 99(12): 8242 - 8247. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
|
