JBC Papers in Press. Published on February 5, 2014 as Manuscript M113.544874 The latest version is at http://www.jbc.org/cgi/doi/10.1074/jbc.M113.544874 Evidence that the DNA Endonuclease ARTEMIS also has Intrinsic 5' Exonuclease Activity Sicong Li1, Howard H. Chang1, Doris Niewolik2, Michael P. Hedrick3, Anthony B. Pinkerton3, Chris Hassig3, Klaus Schwarz2, and Michael R. Lieber1* 1Departments of Pathology, Biochemistry & Molecular Biology, Biological Sciences, and Molecular Microbiology & Immunology, University of Southern California Keck School of Medicine Norris Comprehensive Cancer Center, Los Angeles, CA, USA 2Institute for Transfusion Medicine, University Hospital, Ulm, Germany 3Sanford Burnham Medical Research Institute La Jolla, CA Running Title: Intrinsic 5' Exonuclease of ARTEMIS. *Corresponding author. [email protected] (ph. 323 865 0568). D ow n lo a d Background: DNA-PKcs stimulates the endonucleolytic activities of ARTEMIS, but not the 5’ exonuclease ed activity. fro m Results: Point mutagenesis, small molecule inhibitors, and modulation of divalent cation concentrations all h ttp affect both the endonuclease and 5’ exonuclease activities in parallel. ://w Conclusion: ARTEMIS has intrinsic 5’ exonuclease activity, in addition to its known endonuclease activity. w w Significance: ARTEMIS may use both its endo- and its 5’ exonuclease activity in NHEJ. .jb c .org ARTEMIS is a member of the small molecule inhibitors, which do not b/ y metallo-β-lactamase protein family. block unrelated nucleases. We gu e s ARTEMIS has endonuclease activity at conclude that the 5' exonuclease is t o n DNA hairpins and at 5’ and 3’ DNA intrinsic to ARTEMIS, making it relevant A p overhangs of duplex DNA, and this to the role of ARTEMIS in ril 7 endonucleolytic activity is dependent nonhomologous DNA end joining. , 20 1 9 upon DNA-PKcs. There has been ARTEMIS is a member of the metallo-β- uncertainty about whether ARTEMIS lactamase superfamily of nucleases. The also has 5’ exonuclease activity on metallo-β-lactamase family is characterized by single-stranded DNA and 5' overhangs, a conserved metallo-β-lactamase and β-CASP because this 5’ exonuclease is not domain, and their ability to hydrolyze DNA or dependent upon DNA-PKcs. Here, we RNA. Two other mammalian members of the show that the 5’ exonuclease and the same family, SNM1A and SNM1B (Apollo), endonuclease activities co-purify. function in DNA interstrand cross-link repair (1) Second, we show that a point mutant of and protection of telomeres (2). ARTEMIS, ARTEMIS at a putative active site also known as SNM1C, functions in the residue (H115A) markedly reduces both nonhomologous end joining (NHEJ) repair the endonuclease activity and the 5’ pathway. On duplex DNA, ARTEMIS has exonuclease activity. Third, divalent endonuclease activity on 5' and 3' DNA cation effects on the 5’ exonuclease overhangs and at DNA hairpins, and these are and the endonuclease parallel one all dependent on DNA-PKcs (3). In the context another. Fourth, both the endonuclease of V(D)J recombination, the ARTEMIS:DNA- activity and 5’ exonuclease activity of PKcs complex is required for the opening of ARTEMIS can be blocked in parallel by hairpins generated by the RAG complex (3). 1 Copyright 2014 by The American Society for Biochemistry and Molecular Biology, Inc. Patients lacking ARTEMIS suffer from severe Purified ARTEMIS protein also appears combined immunodeficiency (SCID)(4). NHEJ to have 5’ exonuclease activity, which is DNA- requires nucleases for resection of broken PKcs independent (3). The 5' exonuclease DNA ends, polymerase to fill in the gaps, and activity co-fractionates with the known the XLF:XRCC4:DNA ligase IV complex for endonucleolytic activities of ARTEMIS (12). ligation (5). ARTEMIS is at least one of the Exonucleases are abundant in cells and are nucleases participating in this process and is common contaminants. As other 5' responsible for a subset of end processing exonucleases are purified away, the apparent during NHEJ in response to ionizing radiation level of 5' exonuclease activity falls, but not to (6). Patients or cells lacking ARTEMIS are zero, and the question arises whether the deficient for repair caused by ionizing radiation residual level of 5' exonuclease activity is (IR) (4,7) and for repair of breaks caused by intrinsic to ARTEMIS or a contaminating topoisomerase II inhibitors (8). enzyme that co-fractionates with ARTEMIS. In a biochemical system consisting of Like other members of the metallo-β- purified proteins of the NHEJ pathway, lactamase family, it is thought that ARTEMIS ARTEMIS, in complex with the serine/threonine requires the binding of divalent cations to two kinase DNA-PKcs, removes 5' and 3' co-catalytic portions of the active site for overhangs in order to join duplex DNA ends catalysis (13). Each catalytic site may hold (9). ARTEMIS:DNA-PKcs is also able to open one Mg2+, Mn2+, or Zn2+ ion. In the metallo-β- D hairpins both in vivo and in vitro. ARTEMIS lactamase family, the ligands for the divalent ow n opens hairpins preferentially two nucleotides cations are generally conserved, and consist of loa d past the tip of a perfect hairpin (3). At a 5’ a combination of histidine and aspartic acid ed overhang, ARTEMIS:DNA-PKcs preferentially residues (14). Our previous studies have from cuts directly at the double-strand/single-strand demonstrated that upon mutation of these http junction (3,10). At a 3’ overhang, ARTEMIS conserved residues, the endonuclease activity ://w w preferentially cuts 4 nucleotides into the single- of ARTEMIS:DNA-PKcs can be markedly w .jb stranded overhang from the double reduced. However, the 5’ exonuclease activity c.o strand/single-stranded DNA junction (11). All appeared to remain (13). This raised the brg/ y of the endonuclease activities of ARTEMIS on possibility that the 5’ exonuclease activity and g u duplex DNA ends can be unified by a model the endonuclease activity of ARTEMIS use two est o invoking binding to the single- to double-strand independent catalytic sites. However, an n A p junction and occupying 4 nt along the single- alternative explanation is that previous ril 7 stranded portion, followed by nicking attempts at testing point mutants were carried , 2 0 preferentially 3' of those 4 nt (3). out using 293T cells, a system known for high 19 In addition to its endonucleolytic action levels of background nucleases, and complete on the single-strand to double-strand removal of contaminating 5' exonucleases may transitions (overhangs and hairpins of duplex not have been successful. DNA), ARTEMIS alone has a weaker single- Here, we present a point mutant of stranded DNA endonuclease activity, such as ARTEMIS created using a baculovirus poly dT (12). Like the endonuclease activity of expression system -- a system that more ARTEMIS on duplex DNA substrates, the readily permits purification away from host cell action of ARTEMIS on entirely single-stranded 5’ exonucleases. This mutation markedly DNA is stimulated by the addition of DNA-PKcs reduces both the endonuclease as well as the (12), which forms a complex with ARTEMIS 5' exonuclease activities of ARTEMIS. We that is stable to well over 0.5 M monovalent have also identified small molecules that inhibit salt (3). Our model for ARTEMIS both the ARTEMIS endonuclease and 5' endonucleolytic action on single-stranded DNA exonuclease activities in parallel. These and (ssDNA) proposes binding by ARTEMIS at other findings shed light on ARTEMIS nuclease random locations along the ssDNA, followed by activity and expand the range of substrates nicking, and these activities of ARTEMIS are that ARTEMIS can act upon. We discuss increased by DNA-PKcs. these findings in the context of NHEJ. 2 EXPERIMENTAL PROCEDURES protein was eluted off using 250mM imidazole. Oligonucleotides & DNA Substrates - Eluted protein was directly applied to DEAE- Oligonucleotides used in this study were sepharose resin (GE Healthcare). Unbound synthesized by Operon Biotechnologies,Inc. soluble fraction was collected and dialyzed (Huntsville, AL) and Integrated DNA against Superose 12 buffer (500mM KCl, Technologies, Inc. (San Diego, CA). We 25mM Tris-HCl, 1mM DTT). ARM14 fractions purified the oligonucleotides using 12% or 15% eluted from Superose 12 were frozen down denaturing PAGE and determined the and stored at -80°C. Expression and concentration spectrophotometrically. DNA purification of DNA-PKcs from HeLa cells has substrate 5’ end labeling was done with [γ- been previously described (15). 32P]ATP (3000 Ci/mmol) (PerkinElmer Life Sciences, Boston, MA) and T4 polynucleotide Circular Dichroism - All CD studies were kinase (New England Biolabs) according to the conducted on Jasco-810 Spectro Polarimeter. manufacturer’s instructions. Substrates were A wavelength of 260nm was used at the start incubated with [γ-32P]-ATP and T4 PNK for 30 and 190nm for the end reading. A data pitch of min at 37°C. T4 PNK was subsequently 0.5nm was used with continuous reading and inactivated by incubating samples at 72 °C for 20 accumulations per scan. 20 min. Unincorporated radioisotope was removed by using G-25 Sephadex (Amersham Biochemical Nuclease Assay - The in D Biosciences, Inc.) spin-column vitro DNA nuclease assay was performed in a ow n chromatography. For the hairpin substrate, volume of 10ul with a buffer composition of loa d YM164-labeled oligonucleotide was diluted in a 25mM Tris-HCl (pH 7.5), 10mM KCl, 10mM ed buffer containing 10 mM Tris-hydrochloride (pH MgCl2, and 1mM DTT. In the reaction, 50nM of from 8.0), 1 mM EDTA (pH 8.0), and 100 mM NaCl, single-stranded DNA or 20nM of hairpin DNA http heated at 100°C for 5 min, allowed to cool to YM164 or SL23 was mixed with 50nM of ://w w room temperature for 3 h, and then incubated ARM14 or 50nM of wild-type ARTEMIS. In w .jb at 4 °C overnight. The sequences of the specified cases, 50nM of DNA-PKcs was c.o oligonucleotides used in this study are as added with 0.5mM ATP. DNase I and brg/ y follows: JG 169 5’-TTT TTT TTT TTT TTT TTT micrococcal nuclease were obtained from New g u TTT TTT TTT TTT-3’. JG 167 5’– AAA AAA England Biolabs and Sigma Chemical Co., est o AAA AAA AAA AAA AAA AAA AAA AAA -3’. respectively. Reactions were then incubated n A p YM164, 5’ -TTT TTG ATT ACT ACG GTA GTA at 37°C for 1h. After incubation, reactions ril 7 GCT ACG TAG CTA CTA CCG TAG TAA T-3’. were stopped, and analyzed on a 12% , 2 0 SL 23 5’- T*T*T*T*TT TTG CCA GCT GAC denaturing PAGE. 19 GCG CGT CAG CTG GC-3’. Cells and Transfection - HEK293T cells Protein Expression and Purification - were transfected by nucleofection using ARM14 recombinant baculovirus with a C- AMAXA nucleofector kit V (Lonza). Equal terminal 8x histidine tag was cloned into transfection efficiencies were confirmed by co- baculodirect vector and transfected into SF21 transfection of an enhanced green fluorescent according to protocol (Invitrogen). Virus was protein expression plasmid followed by FACS harvested after 48hrs and used to infect Sf9 analysis. cells in log phase growth. Wild type ARTEMIS was purified as previously described (12). Immunoprecipitation and Immuno- Harvested ARM14 infected Sf9 cells were blotting - The immunoprecipitation was resuspended in Ni-NTA binding buffer (50mM performed as described (3). Myc-tagged NaH PO (pH 7.8), 0.5M KCl, 2mM beta- ARTEMIS was immunoprecipitated and 2 4 mercaptoethanol, 10% glycerol, 0.1% Triton X- detected with anti-myc antibody (Invitrogen). 100, and 20mM imidazole (pH 7.8)) SV40 large T antigen was used as loading supplemented with protease inhibitors. Cells control and was detected with anti-SV40 T-Ag were lysed by sonication and applied to Ni- antibody (Santa Cruz Biotechnology, Santa NTA agarose resin (Qiagen). Bound ARM14 Cruz, CA). Endogenous DNA-PKcs was 3 detected using anti-DNA-PKcs antibody Additionally, we performed a pull down (Abcam). experiment to assess the ability of ARM14 protein to bind to endogenous DNA-PKcs. ARTEMIS or ARM14 expression plasmids RESULTS were transfected into HEK293T cells. Myc- ARTEMIS Endonuclease and ARTEMIS and Myc-ARM14 were 5’Exonuclease Activities Co-Purify with immunoprecipitated using anti-myc antibody. WT ARTEMIS Protein - ARTEMIS has a We assayed to determine if DNA-PKcs was well-documented set of endonuclease also co-immunoprecipitated in both cases. The activities, all of which are stimulated by DNA- presence of DNA-PKcs was assayed using PKcs (3). The 5' exonuclease activity of anti-DNA-PKcs antibody. We find that the ARTEMIS co-fractionates with the ARTEMIS ARM14 point mutant is able to bind to DNA- endonucleolytic activities (Fig. 1). The 5' PKcs with indistinguishable affinity as WT exonuclease activity is seen as the 1 nt product ARTEMIS (Fig. 2F). at the bottom of Figure 1A. The 5' exonuclease activity is not stimulated by DNA- The ARM14 Mutant of ARTEMIS does PKcs (3), and for this reason, we wondered not Possess Hairpin Opening, whether the 5' exonuclease is intrinsic to Endonuclease, or 5’ Exonuclease ARTEMIS. Activity - We tested the activity of ARM14 D fractions collected from the Superose 12 ow Expression and Purification of ARM14 n column for nuclease action on a DNA hairpin lo Point Mutant. - We created a point mutant ad and ssDNA substrates, specifically a single- ed hoifs AtidRiTnEe MatI Sa mbyin mo uatcaidtin 1g1 t5h eto c aonn saelarvneinde (Fig. stranded poly dT substrate. The WT ARTEMIS from 2A). The histidine at amino acid 115 is a endonuclease activity elutes as a symmetrical http putative binding site for divalent cations and defined peak. Unlike WT ARTEMIS ://w w protein, an equal amount of ARM14 mutant w involved in the catalysis of DNA (13). This was .jb protein had no endonuclease activity, on c subsequently cloned into a baculodirect vector .o (Invitrogen) to generate what we call the ssDNA (Fig. 2E, lane 2-5). Importantly, no 5’ brg/ y ARM14 baculovirus and mutant form of exonuclease activity was detected, as gu demonstrated by the lack of a band at the 1 nt es ARTEMIS protein. The ARM14 mutant and WT t o human ARTEMIS-His proteins were both position at the bottom of the gel (Fig. 3A, lane n A p expressed in Sf9 insect cells using a 2 versus lane 3). Therefore, the ARM14 ril 7 baculovirus expression system. Both the WT mutant does not exhibit any endonuclease , 2 0 and ARM14 mutant proteins were purified activity or 5’ exonuclease activity. 19 In some DNA DSBs, the 5’ termini that using Ni-NTA, an anion exchange step, are generated can be 5’ hydroxyl rather than 5’ followed by a Superose 12 gel filtration column. phosphate. We tested for the ability of the Purified ARM14 mutant elutes as a relatively intrinsic 5’ exonuclease activity of ARTEMIS to sharp peak off of the Superose 12 column and process a 3’ labeled substrate with a 5’ runs at 100 kDa on a 8% SDS-PAGE (Fig. 2B hydroxyl group. We labeled the substrate at and 2D). The ARM14 mutant runs at the same the 3’ end by using TdT and dATP that has a position (~100kDa) as the WT ARTEMIS on radioactive phosphate at the alpha position. 8% SDS-PAGE (Fig. 2C). To minimize addition of more than one To ensure that the ARM14 mutant was radiolabeled mononucleotide, we include a 5- not denatured, we used circular dichroism (CD) fold excess of ddTTP, which caps the 3’ on the WT ARTEMIS, on ARM14, and on terminus. However, the final substrate is a completely heat-denatured ARTEMIS. The WT mixture of lengths due to incorporation of 1, 2, ARTEMIS and ARM14 had similar CD profiles or 3 (and smaller amounts of more) (data not shown). Hence, purified ARM14 radiolabels, followed by one dideoxy mutant has a conformation that is mononucleotide. We tested the processitivity indistinguishable from WT ARTEMIS by CD. of the 5’ exonuclease activity of ARTEMIS using this 3’ labeled substrate at three time 4 points. Unlike processive nucleases, activity. To distinguish this from its 5' ARTEMIS is able to generate intermediate exonuclease activity, we utilized a ssDNA poly products even at the 30 min time point (Fig. dA substrate. Single-stranded polypurine 3B, lane 4). Therefore, ARTEMIS acts substrates are not endonucleolytically cut as nonprocessively and can process 5’ hydroxyl efficiently by ARTEMIS, which prefers termini as shown here (Fig. 3B) and substrates polypyrimidine single-stranded DNA substrates with 5’ phosphate termini, as we have shown (9). From our previous work, we knew that the previously [Fig. 3B in ref. (12)]. 5' exonuclease activity of ARTEMIS is DNA- We compared the activity of ARM14 PKcs independent (3). We tested the with WT ARTEMIS on a hairpin substrate. To exonuclease activity of ARTEMIS in the do this, we designed an oligonucleotide DNA presence and absence of DNA-PKcs (Fig. 4B, substrate that is able to fold back on to itself lane 2 vs lane 3). Consistent with our previous upon annealing to form a hairpin substrate findings, we saw no difference in the (YM164. WT ARTEMIS protein shows weak exonuclease activity of purified WT ARTEMIS hairpin opening, weak endonuclease activity, with or without DNA-PKcs. The addition of 20 as well as weak 5’ exonuclease activity on this uM of the inhibitor SBI-713 was able to inhibit substrate in Mg2+-containing reactions that lack the 5' exonuclease activity of ARTEMIS, as DNA-PKcs (Fig. 3C, lane 2). In contrast, the demonstrated by the lack of the 1 nt ARM14 protein generated no hairpin nicking, exonuclease product (Fig. 4B, lane 4). This D endonuclease products or 5’ exonuclease further demonstrates the 5’ exonuclease ow n products even when we added DNA-PKcs to activity of ARTEMIS. Furthermore, we tested lo a d the reaction (Fig. 3C, lane 5). the ability of the ARM14 point mutant to ed process the poly dA ssDNA. The ARM14 from The Endonuclease and 5’ Exonuclease mutant does not have any apparent http activities of ARTEMIS can be Inhibited endonuclease or 5’ exonuclease activity on this ://w w in Parallel - We further tested the ability of substrate when compared to WT ARTEMIS w .jb selected compounds to inhibit the hairpin (Fig. 4C, lane 3 vs. lane 2), indicating that the c .o opening, endonuclease, and 5’ exonuclease point mutation markedly reduces both brg/ y activities of ARTEMIS. Prior to the HTS, a endonuclease and 5’ exonuclease activities. g u small screen showed that ampicillin, a β– We also tested the inhibitors in titrations est o lactamase antibiotic that inhibits to determine if the 5’ exonuclease and the n A p transpeptidase, is also inhibitory, probably endonuclease of ARTEMIS are inhibited over ril 7 because ARTEMIS is a member of the β- the same concentration ranges (Fig. 5). We , 2 0 lactamase family. SBI-713 and SBI-116 are found that as the concentration of ampicillin or 19 two structurally distinct <500 mw small SBI-713 increases, there is a concomitant molecules (not shown) that were identified inhibition of both the endonuclease and the 5’ using a high through-put screen (HTS) of a exonuclease activity of ARTEMIS. The level of 31,000 compound subset of the NIH’s MLSMR inhibition of the 5’ exo- and endonuclease collection (16). activities is similar for both inhibitors. If the 5' In the absence of any inhibitor, exonuclease were a contaminant, we would ARTEMIS has endonuclease activity as well as not expect both the ARTEMIS endonuclease 5’ exonuclease activity on a hairpin substrate, and the contaminating 5’ exonuclease to be YM164 (Fig. 4A, lane 2). However, at 20 uM, inhibited over the very same concentration the SBI-713 dramatically inhibited both the ranges for each drug (Fig. 5). endonuclease and the 5’ exonuclease activities We have also tested SBI-713 and of ARTEMIS (Fig. 4A, lane 4). On the other ampicillin against other nucleases to assess hand, 20 uM of SBI-116 did not inhibit either specificity for inhibition of Artemis. These activity (Fig. 4A, lane 3). included DNase I and micrococcal nuclease Given that ARTEMIS is an (MCN). SBI-713 does not inhibit either of endonuclease, we wondered if ARTEMIS these counter-screen nucleases (Fig. 4D). cutting at the 5’ most nt could merely reflect a In related studies, in 10 mM MgCl2 preferential cutting by ARTEMIS endonuclease reaction buffers, addition of ZnCl2 above 10 uM 5 inhibits ARTEMIS endonuclease activity (Fig. alter the activity of enzymes in the metallo-β- 6). It has been shown previously in certain lactamase family of nucleases (18). We found cases that excess zinc can inhibit the that the addition of excess zinc (> 10 uM) to enzymatic activity of zinc dependent enzymes. the reaction buffer decreases the This is thought to be caused by excess zinc- endonuclease as well as the 5’ exonuclease monohydroxide binding to the active form of activity of ARTEMIS. It seems unlikely that a the zinc dependent enzyme, thereby contaminating 5’ exonuclease would also show preventing the catalytic site from binding active this behavior, particularly at the same low zinc (17). In the case of ARTEMIS, the (micromolar) concentrations of Zn2+. inhibition by excess zinc is observed for the Fifth, we explored the possibility that the endonuclease activity as well as the 5’ 5’ exonuclease activity seen in our gels is exonuclease activity over the same Zn2+ actually the result of the ARTEMIS concentration range (10 to 100 uM) (Fig. 6). endonuclease activity acting on the 5’ terminal This is consistent with the 5’ exonuclease nucleotide. We created a poly dA ssDNA activity of ARTEMIS sharing the same catalytic substrate that is not as readily site as the endonuclease activity of ARTEMIS. endonucleolytically processed by the ARTEMIS endonuclease activity, but DISCUSSION conceivably might be processed by the 5’ Though a 5' DNA exonuclease activity exonuclease activity of ARTEMIS. Indeed, we D o co-purifies with ARTEMIS (Fig. 1), the 5' do observe 5’ exonuclease activity when using w n exonuclease is not dependent on DNA-PKcs, this poly dA ssDNA substrate, indicating that loa d e wachteivrietiaess t(h5e' aAnRdT 3E' MovIeSr hDaNnAg eclnedaovnaugcel eaansde tphoely apcytrivimityid sineee nsu abts tthraet e5 ’ ise nadtt roibnu ata ble to 5’ d from h hairpin opening) do depend on DNA-PKcs. For exonuclease activity and not merely due to the ttp this reason, we did the further studies endonuclease activity of ARTEMIS. ://w w described here to evaluate whether the 5’ Importantly, this 5’ exonuclease activity is w .jb exonuclease is intrinsic to ARTEMIS. completely inhibited by the SBI-713 compound c.o Here, we have provided several (Fig. 4). brg/ y independent lines of evidence that the 5’ From the evidence presented here we g u e exonuclease is intrinsic to ARTEMIS. First, the conclude that the 5’ exonuclease activity seen st o 5’ exonuclease and the endonuclease activities in purified ARTEMIS is intrinsic to the n A p co-fractionate. ARTEMIS protein and not due to a co-purifying ril 7 Second, a point mutant of ARTEMIS is contaminating enzyme. , 2 0 1 markedly reduced in 5’ exonuclease activity in The metallo-β-lactamase family of 9 the presence and absence of DNA-PKcs. A nucleases uses several conserved catalytic point mutant of ARTEMIS is very unlikely to residues to coordinate divalent cations. The affect any contaminating 5’ exonuclease. This two most conserved residues within the alone is strong evidence of an intrinsic 5’ catalytic site are histidine and aspartic acid exonuclease. (14). We generated the ARM14 point mutant, Third, both the endonuclease and the 5’ which changed a conserved histidine residue exonuclease activity of ARTEMIS can be to an alanine residue. If both the inhibited by small molecules that are not endonuclease activity of ARTEMIS and 5’ chelators. It is unlikely that any one inhibitor is exonuclease activity of ARTEMIS shared a able to inhibit both ARTEMIS endonuclease common catalytic site, by disrupting the activity and an unidentified contaminating 5’ divalent ligand in the site there should be a exonuclease. It is even more unlikely that corresponding decrease in both endonuclease these specific inhibitors would affect a as well as the 5’ exonuclease activity of contaminating 5’ exonuclease at the same ARTEMIS, and this is what we observe here. concentrations as they affect ARTEMIS We have previously purified ARM14 endonuclease activity. mutant from 293T cells using a two step Fourth, disruption of the catalytic sites purification. Though this point mutant is by divalent cation replacement is known to inactive for the hairpin opening and 6 endonuclease activity on a hairpin substrate, nuclease activities (both endo- and 5' this version of ARM14 appeared to still retain exonuclease) are abolished by addition of most of the 5’ exonuclease activity (13). We these same zinc specific chelators (data not attribute the discrepancies in our findings with shown), suggesting that members of this family insect cell-purified ARM14 in the current study use the same divalent cations in the catalytic to our earlier study to the high levels of site (21). Biochemical studies have also contaminating nucleases in mammalian cells demonstrated that human SMN1A and SMN1B that are difficult to remove during purification. proteins, also members of the metallo-β- The baculovirus system for ARTEMIS lactamase family and involved in inter-strand production here contains much lower levels of crosslink DNA repair, are capable of acting as background contaminating nucleases, which a 5’ exonuclease on single-stranded and can be subsequently removed in our three step double-stranded DNA substrates (22), and this purification. is its only biochemically defined enzymatic Using a purification process consisting activity. Mutations of key catalytic residues of phosphocellulose, nickel-NTA agarose, and within the metallo-β-lactamase domain also hydroxyapatite, one group separated the usual abolish the 5’-3’ DNA processing ability of bulk of other cellular 5’ exonuclease activity hSMN1a and hSMN1b. Therefore, 5’ away from the endonuclease activity of exonuclease activity is intrinsic to many ARTEMIS (19). They concluded that the 5' members of the metallo-β-lactamase nuclease D exonuclease activity was not intrinsic to family, and the catalytic site involved during ow n ARTEMIS based on an assay that was much DNA processing resides in the metallo-β- loa d less sensitive than our assay. Even at their lactamase domain. Our current study clearly ed reduced sensitivity, we note that some 5' shows that ARTEMIS also has a similar 5' from exonuclease activity is still apparent in their exonuclease activity. http assay gels, which they did not point out (19). The 5’ exonuclease activity described ://w w For example, they tested their purified here would permit Artemis to act on 1 and 2 nt w .jb ARTEMIS on a single-stranded 5' radiolabeled 5’ overhangs more efficiently because such c.o DNA substrate to assay for 5' exonuclease short 5’ overhangs are not optimal substrates brg/ y activity (Fig. 5B of ref. (19)). Though they for action by Artemis in an endonucleolytic g u could detect no 5' exonuclease activity, their manner (3). This would be of particular utility es t o assay was insensitive to the extent that they for repair of ionizing radiation damage, which n A p also could not detect any of the single-stranded might have short overhangs with either a 5’ ril 7 endonuclease activity of ARTEMIS (19). This phosphate or 5’ hydroxyl. , 2 0 was due in part to exposure intensities of their 19 gels that were 10 to 20-fold lower than what we Acknowledgements: This work was use to detect both single-stranded endo- and 5' supported by an NIH grant to M.R.L and the exonuclease activities (12). Despite the fact BMBF grant 01GM1111F to K.S. We thank Dr. that their gels show apparent residual 5' Ray Mosteller (USC) and members of the exonuclease activity at their darker radiolabel Lieber lab for critically reading the manuscript. exposures (Fig. 5C, lanes 10 & 11 in ref. (19)), We thank Dr. Ralf Langen (USC) for help they nevertheless concluded that ARTEMIS obtaining the CD spectra. We thank Dr. David has no intrinsic 5' exonuclease activity (19). Auld (Harvard) and Dr. Andrew Mason (Cal Other members of the metallo-β- State Long Beach) for advice about zinc lactamase family possess nuclease activities metalloenzymes. We thank Fusayo Yamamoto similar to those of ARTEMIS. CPSF-73, an and Drs. T.C. Chung (Sanford-Burnham enzyme involved in the 3' end processing of Research Institute) for conducting the high histone pre-mRNA, has been shown to throughput screen of ARTEMIS inhibitors. possess endonuclease activity as well as 5’ Abbreviations: NHEJ (nonhomologous DNA exonuclease activity on single stranded RNA end joining); DNA-PKcs (DNA-dependent substrates (18,20). The nuclease activities of protein kinase, catalytic subunit). CPSF-73 can be abolished by the addition of zinc specific chelators. We find that ARTEMIS 7 REFERENCES 1. Dronkert, M. L. G., Wit, J. d., Boeve, M., Vasconcelos, M. L., Steeg, H. v., Tan, T. L., Hoeijmakers, J. H. J., and Kanaar, R. (2000) Disruption of mouse SNM1 causes increased sensitivity to the DNA interstrand cross-linking agent mitomycin C. Mol. Cell. Biol. 20, 4553- 4561 2. van Overbeek, M., and de Lange, T. (2006) Apollo, an Artemis-related nuclease, interacts with TRF2 and protects human telomeres in S phase. Curr Biol 16, 1295-1302 3. Ma, Y., Pannicke, U., Schwarz, K., and Lieber, M. R. (2002) Hairpin opening and overhang processing by an Artemis:DNA-PKcs complex in V(D)J recombination and in nonhomologous end joining. Cell 108, 781-794 4. Moshous, D., Callebaut, I., Chasseval, R. d., Corneo, B., Cavazzana-Calvo, M., Diest, F. L., Tezcan, I., Sanal, O., Bertrand, Y., Philippe, N., Fischer, A., and Villartay, J.-P. d. (2001) Artemis, a novel DNA double-strand break repair/V(D)J recombination protein, is mutated in human severe combined immune deficiency. Cell 105, 177-186 5. Lieber, M. R. (2010) The Mechanism of Double-Strand DNA Break Repair by the Nonhomologous DNA End-Joining Pathway. Annu Rev Biochem 79, 181-211 6. Riballo, E., Kuhne, M., Rief, N., Doherty, A., Smith, G. C. M., Recio, M.-J., Reis, C., Dahm, K., D o Fricke, A., Kempler, A., Parker, A. R., Jackson, S. P., Gennery, A., Jeggo, P. A., and Lobrich, w n M. (2004) A pathway of double-strand break rejoining dependent upon ATM, Artemis, and loa d e proteins locating to gamma-H2AX foci. Mol. Cell 16, 715-724 d fro 7. Rooney, S., Alt, F. W., Lombard, D., Whitlow, S., Eckersdorff, M., Fleming, J., Fugmann, S., m h Ferguson, D. O., Schatz, D. G., and Sekiguchi, J. (2003) Defective DNA repair and increased ttp genomic instability in Artemis-deficient murine cells. J. Exp. Med. 197, 553-565 ://w w 8. Kurosawa, A., Koyama, H., Takayama, S., Miki, K., Ayusawa, D., Fujii, M., Iiizumi, S., and w .jb Adachi, N. (2008) The requirement of Artemis in double-strand break repair depends on the c.o rg type of DNA damage. DNA Cell Biol 27, 55-61 b/ y 9. Lu, H., Shimazaki, N., Raval, P., Gu, J., Watanabe, G., Schwarz, K., Swanson, P., and Lieber, g u e M. R. (2008) A biochemically defined system for coding joint formation in human V(D)J st o n recombination. Molecular Cell 31, 485-497 A p 10. Yannone, S. M., Khan, I. S., Zhou, R. Z., Zhou, T., Valerie, K., and Povirk, L. F. (2008) ril 7 Coordinate 5' and 3' endonucleolytic trimming of terminally blocked blunt DNA double-strand , 2 0 1 break ends by Artemis nuclease and DNA-dependent protein kinase. Nucleic Acids Res 36, 9 3354-3365 11. Niewolik, D., Pannicke, U., Lu, H., Ma, Y., Wang, L. C., Kulesza, P., Zandi, E., Lieber, M. R., and Schwarz, K. (2006) DNA-PKcs dependence of artemis endonucleolytic activity: differences between hairpins and 5' or 3' overhangs. J. Biol. Chem. 281, 33900-33909 12. Gu, J., Li, S., Zhang, X., Wang, L. C., Niewolik, D., Schwarz, K., Legerski, R. J., Zandi, E., and Lieber, M. R. (2010) DNA-PKcs regulates a single-stranded DNA endonuclease activity of Artemis. DNA Repair (Amst) 9, 429-437 13. Pannicke, U., Ma, Y., Lieber, M. R., and Schwarz, K. (2004) Functional and biochemical dissection of the structure-specific nuclease Artemis. EMBO J. 23, 1987-1997 14. Aravind, L. (1999) An evolutionary classification of the metallo-beta-lactamase fold proteins. In Silico Biol 1, 69-91 15. Chan, D. W., Mody, C. H., Ting, N. S., and Lees-Miller, S. P. (1996) Purification and characterization of the double-stranded DNA-activated protein kinase, DNA-PK, from human placenta. Biochem. Cell Biol. 74, 67-73 16. PubChem AID:686596, http://pubchem.ncbi.nlm.nih.gov/assay/assay.cgi?aid=686956&loc=ea_ras. 17. Larsen, K. S., and Auld, D. S. (1989) Carboxypeptidase A: mechanism of zinc inhibition. Biochemistry 28, 9620-9625 8 18. Dominski, Z. (2007) Nucleases of the metallo-beta-lactamase family and their role in DNA and RNA metabolism. Crit Rev Biochem Mol Biol 42, 67-93 19. Pawelczak, K. S., and Turchi, J. J. (2010) Purification and characterization of exonuclease-free Artemis: Implications for DNA-PK-dependent processing of DNA termini in NHEJ-catalyzed DSB repair. DNA Repair (Amst) 9, 670-677 20. Mandel, C. R., Kaneko, S., Zhang, H., Gebauer, D., Vethantham, V., Manley, J. L., and Tong, L. (2006) Polyadenylation factor CPSF-73 is the pre-mRNA 3'-end-processing endonuclease. Nature 444, 953-956 21. Ryan, K., Calvo, O., and Manley, J. L. (2004) Evidence that polyadenylation factor CPSF-73 is the mRNA 3' processing endonuclease. Rna 10, 565-573 22. Sengerova, B., Allerston, C. K., Abu, M., Lee, S. Y., Hartley, J., Kiakos, K., Schofield, C. J., Hartley, J. A., Gileadi, O., and McHugh, P. J. (2012) Characterization of the human SNM1A and SNM1B/Apollo DNA repair exonucleases. J Biol Chem 287, 26254-26267 FIGURE LEGENDS Figure 1. Mono Q purified ARTEMIS exonuclease activity correlates with its endonuclease activity. A, 20 nM of 5' radiolabeled ssDNA substrate (JG169) was incubated with 1ul of Mono Q fractions in a D o 10ul reaction. The reaction was performed in 25 mM Tris-Cl (pH 8.0), 10 mM MnCl , 50 ug/mL BSA, w 2 n and 1 mM DTT for 60 min at 37°C. Lane 2 contains pre-Mono Q ARTEMIS, and lanes 3-18 contain loa d e Mono Q fractions 8-23. The 5' exonuclease activity is seen as the 1 nt product at the bottom of the d fro gel. m h B, The signals of remaining substrate (30nt position), endonucleolytic products (2-29nt position), and ttp exonucleoytic product (1nt position) were quantified by boxing the corresponding bands and ://w w measuring the summed band intensities using Bio-Rad Quantify One software. The signal % was w .jb plotted as a function of fraction number. The top graph plots the substrate and endonucleolytic c.o rg product signals. The bottom graph plots the 5' exonucleolytic product signal. b/ y C, An SDS PAGE gel is stained with Coomassie Blue. The gel shows that purified Artemis from the g u e Mono Q column runs at the 100 kDa position. Fractions 10 – 18 correlate with peak activity as well st o n as ARTEMIS protein abundance. A p ril 7 Figure 2. Size exclusion chromatography and characterization of a single amino acid , 2 0 1 point mutant of ARTEMIS, the ARM14 mutant. 9 A, The ARM14 mutant was constructed by creating a point mutation in the putative active site of ARTEMIS, specifically histidine 115 to alanine, followed by cloning into a baculodirect vector. B, An SDS-PAGE gel was stained with Coomassie Blue. The gel shows that the ARM14 mutant protein has a gel mobility of ~100kDa. C, ARM14 and wild type ARTEMIS were both run on an 8% SDS-PAGE gel and stained with Coomassie Blue. Purified ARM14 mutant (lane 2) and wild type ARTEMIS (lane 3) have gel mobility positions at ~100 kDa, 2 ug of each was loaded onto the SDS-PAGE. The molecular weight marker is in lane 1. D, The UV profile of ARM14 during Superose 12 size exclusion chromatography is shown. Purified ARM 14 mutants elutes as a ~450 kDa multimer at fraction 15 on the Superose 12 size exclusion column. E, Fractions across the Superose 12 UV peak were collected and assayed for nuclease activity. In each reaction, 50 nM of ARM14 mutant was incubated with 50 nM of ssDNA poly dT substrate (JG169) in a 10 ul reaction. Reactions were incubated for 60 min at 37°C, stopped, and run on 12% 9 denaturing PAGE. The ARM14 mutant fractions 15-18 do not show any endonuclease or 5’ exonuclease activity. F, ARM14 binds endogenous DNA-PKcs. Empty vector, wildtype (ART-WT) or mutant ARTEMIS (ARM14) expressing plasmids were transfected into HEK293T cells, and the myc-tagged ART- WT/ARM14 was immunoprecipitated (IP) in the presence (+) or absence (-) of anti-myc antibody from cell lysates. N.a. not applicable. ART-WT/ARM14, co-precipitated endogenous DNA-PKcs, and the loading control SV40 large T-antigen were detected with the corresponding antibodies by Western blot analysis. The position of protein standards is given at the left in kDa. Transfection efficiencies were between 83 and 86%. Figure 3. Endonuclease and 5’ exonuclease activity of WT ARTEMIS and the ARTEMIS point mutant, ARM14. A, The activity of ARM14 mutant was assayed for activity and compared to that of WT ARTEMIS. 50 nM of ARM14 mutant or WT ARTEMIS was incubated with 50 nM of ssDNA poly dT substrate in a 10 ul reaction. Reactions were incubated for 60 min at 37°C, stopped, and run on 12% denaturing PAGE. The 5' exonuclease activity is seen as the 1 nt product at the bottom of the gel. B, The activity of WT ARTEMIS was assayed for activity on a 3’ labeled substrate to determine the D processitivity of the 5’ exonuclease activity. 50 nM of WT ARTEMIS was incubated with 50 nM of ow n ssDNA poly dT substrate (3’ labeled by Tdt) in a 10 ul reaction. Reactions were incubated for either lo a d 2, 5, or 30 min as indicated. Where compound SBI-713 is used, 20 uM was added to the reaction. ed from C, 50 nM of hairpin DNA substrate (YM164) was incubated with 50 nM WT ARTEMIS or 50 nM of http ARM14 in a 10 ul reaction. Reactions were incubated at 37°C for 60 min. In lanes where DNA-PKcs ://w w was added, 50 nM of DNA-PKcs, 0.5 mM of ATP, and 0.25 uM of cold DNA YM8/9 were also added. w .jb Lane 1 contains only the hairpin substrate. Lane 2 contains ARTEMIS, which does have low dsDNA c .o endonucleolytic activity without DNA-PKcs being present, but it has high 5’ exonucleolytic activity (1 brg/ y nt product at bottom of gel). Lane 4 and 5 are the ARM14 point mutant of ARTEMIS. Lane 3 WT g u ARTEMIS plus DNA-PKcs, which generates a hairpin opened product, and 5’ overhang cleavage es t o products. Lane 6 is the control lane containing DNA-PKcs, but no ARTEMIS. n A p ril 7 Figure 4. Small molecule inhibitors block both the 5' exonuclease and the , 2 0 endonucleolytic activities of ARTEMIS on hairpin DNA. 19 A, Addition of compound SBI-713 inhibits the endonuclease and 5’ exonuclease activities of ARTEMIS. 50 nM of hairpin DNA substrate (YM164) was incubated with 60 nM ARTEMIS in a 10ul reaction. Reactions were incubated at 37°C for 60 min. Compound SBI-116 or SBI-713 was added at 20 uM in the indicated reactions. The 5' exonuclease activity is seen as the 1 nt product at the bottom of the gel. B, ARTEMIS has weak endonuclease activity on a single-stranded poly dA substrate, but has 5' exonuclease activity. 50 nM of ssDNA substrate (JG167) was incubated with 60 nM of ARTEMIS in addition to 75 nM of DNA-PKcs and 0.5 mM ATP where applicable. 20 uM of the inhibitor SBI-713 was added to the indicated reactions. Reactions were incubated for 1hr at 37°C. The 5' exonuclease activity is seen as the 1 nt product at the bottom of the gel. The Klenow ladder was generated by adding 2 units of Klenow (NEB) to 50 nM of substrate JG 167. Reactions were incubated for the indicated time periods at 37°C. C, Point mutant ARM14 does not have endonuclease or 5’ exonuclease activity on poly dA substrate. 50 nM of ssDNA substrate (JG167) was incubated with 60 nM of ARTEMIS or 60 nM of ARM14 point mutant. Reactions were incubated for 1hr at 37°C. The 5' exonuclease activity is seen as the 1 nt product at the bottom of the gel. D, The specificity of compound SBI-713 was tested against DNase I and micrococcal nuclease. 100 nM of DNase I or 2.5 uM of micrococcal nuclease (Sigma) was incubated with 20 nM of the hairpin 10
Description: