Trimming of damaged 3′ overhangs of DNA double-strand breaks by the Metnase and Artemis endonucleases

Both Metnase and Artemis possess endonuclease activities that trim 3′ overhangs of duplex DNA. To assess the potential of these enzymes for facilitating resolution of damaged ends during double-strand break rejoining, substrates bearing a variety of normal and structurally modified 3′ overhangs were constructed, and treated either with Metnase or with Artemis plus DNA-dependent protein kinase (DNA-PK). Unlike Artemis, which trims long overhangs to 4–5 bases, cleavage by Metnase was more evenly distributed over the length of the overhang, but with significant sequence dependence. In many substrates, Metnase also induced marked cleavage in the double-stranded region within a few bases of the overhang. Like Artemis, Metnase efficiently trimmed overhangs terminated in 3′-phosphoglycolates (PGs), and in some cases the presence of 3′-PG stimulated cleavage and altered its specificity. The nonplanar base thymine glycol in a 3′ overhang severely inhibited cleavage by Metnase in the vicinity of the modified base, while Artemis was less affected. Nevertheless, thymine glycol moieties could be removed by Metnase- or Artemis-mediated cleavage at sites farther from the terminus than the lesion itself. In in vitro end-joining systems based on human cell extracts, addition of Artemis, but not Metnase, effected robust trimming of an unligatable 3′-PG overhang, resulting in a dramatic stimulation of ligase IV- and XLF-dependent end joining. Thus, while both Metnase and Artemis are biochemically capable of resolving a variety of damaged DNA ends for the repair of complex double-strand breaks, Artemis appears to act more efficiently in the context of other nonhomologous end joining proteins.     

Conversion of phosphoglycolate to phosphate termini on 3' overhangs of DNA double strand breaks by the human tyrosyl-DNA phosphodiesterase hTdp1

Mammalian cells contain potent activity for removal of 3'-phosphoglycolates from single-stranded oligomers and from 3' overhangs of DNA double strand breaks, but no specific enzyme has been implicated in such removal. Fractionated human whole-cell extracts contained an activity, which in the presence of EDTA, catalyzed removal of glycolate from phosphoglycolate at a single-stranded 3' terminus to leave a 3'-phosphate, reminiscent of the human tyrosyl-DNA phosphodiesterase hTdp1. Recombinant hTdp1, as well as Saccharomyces cerevisiae Tdp1, catalyzed similar removal of glycolate, although less efficiently than removal of tyrosine. Moreover, glycolate-removing activity could be immunodepleted from the fractionated extracts by antiserum to hTdp1. When a plasmid containing a double strand break with a 3'-phosphoglycolate on a 3-base 3' overhang was incubated in human cell extracts, phosphoglycolate processing proceeded rapidly for the first few minutes but then slowed dramatically, suggesting that the single-stranded overhangs gradually became sequestered and inaccessible to hTdp1. The results suggest a role for hTdp1 in repair of free radical-mediated DNA double strand breaks bearing terminally blocked 3' overhangs.     

Coordinate 5' and 3' endonucleolytic trimming of terminally blocked blunt DNA double-strand break ends by Artemis nuclease and DNA-dependent protein kinase

Previous work showed that, in the presence of DNA-dependent protein kinase (DNA-PK), Artemis slowly trims 3′-phosphoglycolate-terminated blunt ends. To examine the trimming reaction in more detail, long internally labeled DNA substrates were treated with Artemis. In the absence of DNA-PK, Artemis catalyzed extensive 5′→3′ exonucleolytic resection of double-stranded DNA. This resection required a 5′-phosphate, but did not require ATP, and was accompanied by endonucleolytic cleavage of the resulting 3′ overhang. In the presence of DNA-PK, Artemis-mediated trimming was more limited, was ATP-dependent and did not require a 5′-phosphate. For a blunt end with either a 3′-phosphoglycolate or 3′-hydroxyl terminus, endonucleolytic trimming of 2–4 nucleotides from the 3′-terminal strand was accompanied by trimming of 6 nt from the 5′-terminal strand. The results suggest that autophosphorylated DNA-PK suppresses the exonuclease activity of Artemis toward blunt-ended DNA, and promotes slow and limited endonucleolytic trimming of the 5′-terminal strand, resulting in short 3′ overhangs that are trimmed endonucleolytically. Thus, Artemis and DNA-PK can convert terminally blocked DNA ends of diverse geometry and chemical structure to a form suitable for polymerase-mediated patching and ligation, with minimal loss of terminal sequence. Such processing could account for the very small deletions often found at DNA double-strand break repair sites.     

Resistance of 3'-phosphoglycolate DNA ends to digestion by mammalian DNase III

An essential step in the repair of free radical-mediated DNA strand breaks is the removal of sugar fragments such as phosphoglycolate from the 3' termini. While the abasic endonuclease Ape1 can remove phosphoglycolate from single-strand breaks in double-stranded DNA, an enzyme capable of removing it from 3' overhangs of double-strand breaks has yet to be identified. We therefore tested DNase III, the predominant 3' --> 5' exonuclease in mammalian cell extracts, for possible 3'-phosphoglycolate-removing activity. However, all 3'-phosphoglycolate substrates, as well as a 3'-phosphate substrate, were resistant to DNase III under conditions in which the analogous 3'-hydroxyl substrates were extensively degraded. The DNA end-binding protein Ku (an equimolar mixture of Ku70, now known as G22P1, and Ku86, now known as XRCC5) did not alter the resistance of the 3'-phosphoglycolate substrates, but the protein modulated the susceptibility of 3'-hydroxyl substrates, allowing DNase III to remove a 3' overhang but inhibiting digestion of the double-stranded portion of the substrate.     

DNA-PK autophosphorylation facilitates Artemis endonuclease activity

The Artemis nuclease is defective in radiosensitive severe combined immunodeficiency patients and is required for the repair of a subset of ionising radiation induced DNA double-strand breaks (DSBs) in an ATM and DNA-PK dependent process. Here, we show that Artemis phosphorylation by ATM and DNA-PK in vitro is primarily attributable to S503, S516 and S645 and demonstrate ATM dependent phosphorylation at serine 645 in vivo. However, analysis of multisite phosphorylation mutants of Artemis demonstrates that Artemis phosphorylation is dispensable for endonuclease activity in vitro and for DSB repair and V(D)J recombination in vivo. Importantly, DNA-dependent protein kinase catalytic subunit (DNA-PKcs) autophosphorylation at the T2609-T2647 cluster, in the presence of Ku and target DNA, is required for Artemis-mediated endonuclease activity. Moreover, autophosphorylated DNA-PKcs stably associates with Ku-bound DNA with large single-stranded overhangs until overhang cleavage by Artemis. We propose that autophosphorylation triggers conformational changes in DNA-PK that enhance Artemis cleavage at single-strand to double-strand DNA junctions. These findings demonstrate that DNA-PK autophosphorylation regulates Artemis access to DNA ends, providing insight into the mechanism of Artemis mediated DNA end processing.     

DNA-PKcs dependence of Artemis endonucleolytic activity, differences between hairpins and 5' or 3' overhangs

During V(D)J recombination, the RAG proteins create DNA hairpins at the V, D, or J coding ends, and the structure-specific nuclease Artemis is essential to open these hairpins prior to joining. Artemis also is an endonuclease for 5' and 3' overhangs at many DNA double strand breaks caused by ionizing radiation, and Artemis functions as part of the nonhomologous DNA end joining pathway in repairing these. All of these activities require activation of the Artemis protein by interaction with and phosphorylation by the DNA-dependent protein kinase catalytic subunit (DNA-PKcs). In this study, we have identified a region of the Artemis protein involved in the interaction with DNA-PKcs. Furthermore, the biochemical and functional analyses of C-terminally truncated Artemis variants indicate that the hair-pin opening and DNA overhang endonucleolytic features of Artemis are triggered by DNA-PKcs in two modes. First, autoinhibition mediated by the C-terminal tail of Artemis is relieved by phosphorylation of this tail by DNA-PKcs. Thus, C-terminally truncated Artemis derivatives imitate DNA-PKcs-activated wild type Artemis protein and exhibit intrinsic hairpin opening activity. Second, DNA-PKcs may optimally configure 5' and 3' overhang substrates for the endonucleolytic function of Artemis.     

Enzyme action at 3' termini of ionizing radiation-induced DNA strand breaks

gamma-Irradiation of DNA in vitro produces two types of single strand breaks. Both types of strand breaks contain 5'-phosphate DNA termini. Some strand breaks contain 3'-phosphate termini, some contain 3'-phosphoglycolate termini (Henner, W.D., Rodriguez, L.O., Hecht, S. M., and Haseltine, W. A. (1983) J. Biol. Chem. 258, 711-713). We have studied the ability of prokaryotic enzymes of DNA metabolism to act at each of these types of gamma-ray-induced 3' termini in DNA. Neither strand breaks that terminate with 3'-phosphate nor 3'-phosphoglycolate are substrates for direct ligation by T4 DNA ligase. Neither type of gamma-ray-induced 3' terminus can be used as a primer for DNA synthesis by either Escherichia coli DNA polymerase or T4 DNA polymerase. The 3'-phosphatase activity of T4 polynucleotide kinase can convert gamma-ray-induced 3'-phosphate but not 3'-phosphoglycolate termini to 3'-hydroxyl termini that can then serve as primers for DNA polymerase. E. coli alkaline phosphatase is also unable to hydrolyze 3'-phosphoglycolate groups. The 3'-5' exonuclease actions of E. coli DNA polymerase I and T4 DNA polymerase do not degrade DNA strands that have either type of gamma-ray-induced 3' terminus. E. coli exonuclease III can hydrolyze DNA with gamma-ray-induced 3'-phosphate or 3'-phosphoglycolate termini or with DNase I-induced 3'-hydroxyl termini. The initial action of exonuclease III at 3' termini of ionizing radiation-induced DNA fragments is to remove the 3' terminal phosphate or phosphoglycolate to yield a fragment of the same nucleotide length that has a 3'-hydroxyl terminus. These results suggest that repair of ionizing radiation-induced strand breaks may proceed via the sequential action of exonuclease, DNA polymerase, and DNA ligase. The possible role of exonuclease III in repair of gamma-radiation-induced strand breaks is discussed.     

In vitro repair of synthetic ionizing radiation-induced multiply damaged DNA sites.

When ionizing radiation traverses a DNA molecule, a combination of two or more base damages, sites of base loss or single strand breaks can be produced within 1-4 nm on opposite DNA strands, forming a multiply damaged site (MDS). In this study, we reconstituted the base excision repair system to examine the processing of a simple MDS containing the base damage, 8-oxoguanine (8-oxoG), or an abasic (AP) site, situated in close opposition to a single strand break, and asked if a double strand break could be formed. The single strand break, a nucleotide gap containing 3' and 5' phosphate groups, was positioned one, three or six nucleotides 5' or 3' to the damage in the complementary DNA strand. Escherichia coli formamidopyrimidine DNA glycosylase (Fpg), which recognizes both 8-oxoG and AP sites, was able to cleave the 8-oxoG or AP site-containing strand when the strand break was positioned three or six nucleotides away 5' or 3' on the opposing strand. When the strand break was positioned one nucleotide away, the target lesion was a poor substrate for Fpg. Binding studies using a reduced AP (rAP) site in the strand opposite the gap, indicated that Fpg binding was greatly inhibited when the gap was one nucleotide 5' or 3' to the rAP site.To complete the repair of the MDS containing 8-oxoG opposite a single strand break, endonuclease IV DNA polymerase I and Escherichia coli DNA ligase are required to remove 3' phosphate termini, insert the "missing" nucleotide, and ligate the nicks, respectively. In the absence of Fpg, repair of the single strand break by endonuclease IV, DNA polymerase I and DNA ligase occurred and was not greatly affected by the 8-oxoG on the opposite strand. However, the DNA strand containing the single strand break was not ligated if Fpg was present and removed the opposing 8-oxoG. Examination of the complete repair reaction products from this reaction following electrophoresis through a non-denaturing gel, indicated that a double strand break was produced. Repair of the single strand break did occur in the presence of Fpg if the gap was one nucleotide away. Hence, in the in vitro reconstituted system, repair of the MDS did not occur prior to cleavage of the 8-oxoG by Fpg if the opposing single strand break was situated three or six nucleotides away, converting these otherwise repairable lesions into a potentially lethal double strand break.     

Analysis of non-template-directed nucleotide addition and template switching by DNA polymerase

DNA polymerases use an uninterrupted template strand to direct synthesis of DNA. However, some DNA polymerases can synthesize DNA across two discontinuous templates by binding and juxtaposing them, resulting in synthesis across the junction. Primer/template duplexes with 3' overhangs are especially efficient substrates, suggesting that DNA polymerases use the overhangs as regions of microhomology for template synapsis. The formation of these overhangs may be the result of non-template-directed nucleotide addition by DNA polymerases. To examine the relative magnitude and mechanism of template switching, we studied the in vitro enzyme kinetics of template switching and non-template-directed nucleotide addition by the 3'-5' exonuclease-deficient large fragment of Escherichia coli DNA polymerase I. Non-template-directed nucleotide addition and template switching were compared to that of standard primer extension. We found that non-template-directed nucleotide addition and template switching showed similar rates and were approximately 100-fold slower than normal template-directed DNA synthesis. Furthermore, non-template-directed nucleotide addition showed a 10-fold preference for adding dAMP to the ends of DNA over that of the other three nucleotides. For template switching, kinetic analysis revealed that the two template substrates acted as a random bireactant system with mixed-type inhibition of substrate binding by one substrate over the other. These data are the first to establish the binding kinetics of two discontinuous DNA substrates to a single DNA polymerase. Our results suggest that although the activities are relatively weak, non-template-directed nucleotide addition and template switching allow DNA polymerases to overcome breaks in the template strand in an error-prone manner.     

T4 endonuclease VII resolves cruciform DNA with nick and counter-nick and its activity is directed by local nucleotide sequence.

Endonuclease VII of phage T4 resolves Holliday structures in vitro by nicking pairs of strands across the junction. We report here analyses of this reaction between endonuclease VII and a Holliday structure analogue, made in vitro from synthetic oligonucleotides. The enzyme cleaves the structure in a non-concerted way and nicks each strand independently. Combinations of nicks with counter-nicks in strands across the junction resolve the construct. The specificity of the enzyme for DNA secondary structures was tested with a series of branched molecules made from oligonucleotides with the same nucleotide sequence in one strand. Results show that the number, location and relative cleavage efficiencies depend largely on the local nucleotide sequence, rather than on the branch type. In particular, endonuclease VII cleaves a complete four-armed cruciform as efficiently as a three-armed Y-junction or its derivatives, a semi-Y, a fork with two single-strand overhangs, a single-strand overhang, and a nicked DNA. However, exchange or addition of one or more nucleotides within the cleavage area flanking the structural signal for endonuclease VII strongly affects the cleavage pattern as well as their relative efficiency of usage. Examples with a single-stranded overhang are presented and show in summary that the enzyme has a fivefold preference for pyrimidines rather than purines.     

The generation of long telomere overhangs in human cells: a model and its implication

MOTIVATION: Linear chromosomes carry on both ends repetitive DNA sequences called telomere. In the conventional model of semi-conservative DNA replication, the 3'-end of a linear DNA strand cannot be fully replicated, resulting in a single-stranded 3' overhang at one end of the double-stranded DNA product. In this model, the length of the overhang is expected to be about the size of an RNA primer (about nine nucleotides for human cells). However, it has been found that the telomere overhangs in human cells can be as long as several hundred nucleotides. At present, the opinion regarding how such long overhangs are produced is controversial. RESULTS: In order to gain insight into the mechanism by which long telomere overhangs are produced, we derived a mathematical model that can perfectly describe the length distribution of telomere overhangs in several human cell strains. The model suggests that the production of telomere overhangs can be explained by three contributions corresponding to three regions on the G-rich telomere template strand, namely, the region occupied by the last primer, that missed out by this primer at its 5'-side and the 3'-terminus of the template strand that is inaccessible to primase. The model can also be used to simulate incomplete telomere replication.     

Generation of telomeric G strand overhangs involves both G and C strand cleavage

Processing of telomeric DNA is required to generate the 3' G strand overhangs necessary for capping chromosome ends. We have investigated the steps involved in telomere processing by examining G overhang structure in Tetrahymena cells that lack telomerase or have altered telomeric sequences. We show that overhangs are generated by two precise cleavage steps involving nucleases that are robust but lack sequence specificity. Our data suggest that a G overhang binding protein delineates the boundaries for G and C strand cleavage. We also show that telomerase is not the nuclease responsible for G strand cleavage, although telomerase depletion alters the precision of processing. This change in processing indicates that telomerase affects multiple transactions at the telomere and provides a physical footprint for the continued association of telomerase with the telomere after repeat addition is complete.     

The Artemis:DNA-PKcs endonuclease cleaves DNA loops, flaps, and gaps

In eukaryotic cells, nonhomologous DNA end joining (NHEJ) is a major pathway for repair of double-strand DNA breaks (DSBs). Artemis and the 469kDa DNA-dependent protein kinase (DNA-PKcs) together form a key nuclease for NHEJ in vertebrate organisms. The structure-specific endonucleolytic activity of Artemis is activated by binding to and phosphorylation by DNA-PKcs. We tested various DNA structures in order to understand the range of structural features that are recognized by the Artemis:DNA-PKcs complex. We find that all tested substrates that contain single-to-double-strand transitions can be cleaved by the Artemis:DNA-PKcs complex near the transition region. The cleaved substrates include heterologous loops, stem-loops, flaps, and gapped substrates. Such versatile activity on single-/double-strand transition regions is important in understanding how reconstituted NHEJ systems that lack DNA polymerases can join incompatible DNA ends and yet preserve 3' overhangs. Additionally, the flexibility of the Artemis:DNA-PKcs nuclease may be important in removing secondary structures that hinder processing of DNA ends during NHEJ.     

Transient generation of displaced single-stranded DNA during nick translation

We show that displaced single-stranded overhangs are transiently generated and destroyed during nick translation by E. coli DNA polymerase I. Evidence that hyper-rec mutants have an increased frequency of such overhang structures is discussed. The transient generation of overhangs may be significant for general recombination. The 5' leads to 3' exonuclease activity of polymerase I specifically hydrolyzes such overhangs to yield a nick. Overhangs are generated by polymerization, but after every polymerization step, either polymerase or exonuclease can act--55% of the time, polymerization occurred first. At this frequency overhangs of greater than or equal to 12 nucleotides are generated every 1300 nucleotides polymerized. We suggest that many DNA strand discontinuities are displaced single-stranded overhangs, rather than gaps or simple nicks.     

Hairpin opening and overhang processing by an Artemis/DNA-dependent protein kinase complex in nonhomologous end joining and V(D)J recombination

Mutations in the Artemis protein in humans result in hypersensitivity to DNA double-strand break-inducing agents and absence of B and T lymphocytes (radiosensitive severe combined immune deficiency [RS-SCID]). Here, we report that Artemis forms a complex with the 469 kDa DNA-dependent protein kinase (DNA-PKcs) in the absence of DNA. The purified Artemis protein alone possesses single-strand-specific 5' to 3' exonuclease activity. Upon complex formation, DNA-PKcs phosphorylates Artemis, and Artemis acquires endonucleolytic activity on 5' and 3' overhangs, as well as hairpins. Finally, the Artemis:DNA-PKcs complex can open hairpins generated by the RAG complex. Thus, DNA-PKcs regulates Artemis by both phosphorylation and complex formation to permit enzymatic activities that are critical for the hairpin-opening step of V(D)J recombination and for the 5' and 3' overhang processing in nonhomologous DNA end joining.     

G-overhang dynamics at Tetrahymena telomeres

To learn more about the structure of the DNA terminus at Tetrahymena thermophila telomeres, we have devised a ligation-mediated primer extension protocol to accurately measure the length of the G-strand overhang. We show that overhang length and the identity of the 3'-terminal nucleotide are tightly regulated. The majority of overhangs terminate in the sequence 5'-TTGGGGT and >80% are either 14-15 or 20-21 nucleotides in length. No significant changes in overhang length were detected as cells traversed the cell cycle. However, changes in length distribution were observed when cells exited the cell cycle, indicating an altered balance between DNA synthesis and degradation or end protection. We also provide evidence that rDNA molecules have overhangs on both telomeres. Full-length rDNA could be cloned by a strategy that depends on overhangs being present at both ends. Moreover, analysis of leading strand telomeres revealed that a significant fraction have overhangs > or =5 nucleotides. Our results indicate that generation of the terminal telomeric DNA structure is highly regulated and requires several distinct DNA-processing events.     

Human HeLa cell enzymes that remove phosphoglycolate 3''-end groups from DNA.

We have purified three chromatographically distinct human enzyme activities from HeLa cells, that are capable of converting bleomycin-treated DNA into a substrate for E. coli DNA polymerase I. The bleomycin-treated DNA substrate used in this study has been characterized via a 32P-postlabeling assay and shown to contain strand breaks with 3'-phosphoglycolate termini as greater than 95% of the detectable dose-dependent lesions. The purified HeLa cell enzymes were shown to be capable of removing 3'-phosphoglycolates from this substrate. Also 3'-phosphoglycolate removal and nucleotide incorporation were enzyme dependent. In addition, all three Hela cell enzymes have been determined to possess Class II AP endonuclease activity. The enzymes lack 3'----5' exonuclease activity and are, therefore, dissimilar to exonuclease III--an E. coli enzyme that can remove 3'-phosphoglycolate.     

The actions of Neurospora endo-exonuclease on double strand DNAs

Neurospora crassa endo-exonuclease, an enzyme implicated in recombinational DNA repair, was found previously to have a distributive endonuclease activity with a high specificity for single strand DNA and a highly processive exonuclease activity. The activities of endo-exonuclease on double strand DNA substrates have been further explored. Endo-exonuclease was shown to have a low bona fide endonuclease activity with completely relaxed covalently closed circular DNA and made site-specific breaks in linear double strand DNA at a low frequency while simultaneously generating a relatively high level of single strand breaks (nicks) in the DNA. Sequencing at nicks induced by endo-exonuclease in pBR322 restriction fragments showed that the highest frequency of nicking occurred at the mid-points of two sites with the common sequence, p-AGCACT-OH. In addition, sequencing revealed less frequent nicking at identical or homologous hexanucleotide sequences in all other 54 cases examined where these sequences either straddled the break site itself or were within a few nucleotides on either side of the break site. The exonucleolytic action of endo-exonuclease on linear DNA showed about 100-fold preference for acting in the 5' to 3' direction. Removal of the 5'-terminal phosphates substantially reduced this activity, internal nicking, and the ability of endo-exonuclease to generate site-specific double strand breaks. On the other hand, nicking of the dephosphorylated double strand DNA with pancreatic DNase I stimulated the exonuclease activity by almost 5-fold, but no stimulation was observed when the DNA was nicked by Micrococcal nuclease. Thus, 5'-p termini either at double strand ends or at nicks in double strand DNA are entry points to the duplex from which endo-exonuclease diffuses linearly or "tracks" in the 5' to 3' direction to initiate its major endo- and exonucleolytic actions. The results are interpreted to show how it is possible for endo-exonuclease to generate single strand DNA for switching into a homologous duplex either at a nick or while remaining bound at a double strand break in the DNA. Such mechanisms are consistent with current models for recombinational double strand break repair in eukaryotes.     

Stability of 3' double nucleotide overhangs that model the 3' ends of siRNA

Thermodynamic parameters are reported for duplex formation in 1 M NaCl for 16 RNA sequences, each containing a core tetramer duplex, GGCC, and a 3' overhang consisting of two bases. The results indicate additional double-helical stability is conferred by the double 3' terminal overhang relative to the single 3' terminal overhang. A nearest-neighbor analysis of the data indicates that the free energy contribution at 37 degrees C of the second base in the double 3' terminal overhang varies from 0 to 0.7 kcal/mol. The second base in the 3' double overhang can contribute nearly the same stability to a duplex as a base pair or a 3' dangling overhang. Stability contribution of a dangling base, two nucleotides removed from the 3' end of a duplex, is dependent upon both the identity of the base as well as that of the dangling base that it neighbors. A second dangling base only increases the stability of the duplex when it is neighboring a 3' purine dangling nucleotide. Furthermore, a second dangling pyrimidine provides a greater contribution to duplex stability than a purine. A nearest-neighbor model was developed to predict the influence of 3' double overhang on the stability of duplex formation. The model improves the prediction of free energy and melting temperature when tested against six sequences with different core duplexes.