Portal de Periódicos da CAPES

Interaction of DNA Fragmentation Factor (DFF) with DNA Reveals an Unprecedented Mechanism for Nuclease Inhibition and Suggests That DFF Can Be Activated in a DNA-bound State

2004; Elsevier BV; Volume: 280; Issue: 7 Linguagem: Inglês

10.1074/jbc.m413035200

1083-351X

Christian Korn, Sebastian Scholz, Oleg Gimadutdinow, Rudi Lurz, Alfred Pingoud, Gregor Meiß,

RNA Interference and Gene Delivery

DNA fragmentation factor (DFF) is a complex of the DNase DFF40 (CAD) and its chaperone/inhibitor DFF45 (ICAD-L) that can be activated during apoptosis to induce DNA fragmentation. Here, we demonstrate that DFF directly binds to DNA in vitro without promoting DNA cleavage. DNA binding by DFF is mediated by the nuclease subunit, which can also form stable DNA complexes after release from DFF. Recombinant and reconstituted DFF is catalytically inactive yet proficient in DNA binding, demonstrating that the nuclease subunit in DFF is inhibited in DNA cleavage but not in DNA binding, revealing an unprecedented mode of nuclease inhibition. Activation of DFF in the presence of naked DNA or isolated nuclei stimulates DNA degradation by released DFF40 (CAD). In transfected HeLa cells transiently expressed DFF associates with chromatin, suggesting that DFF could be activated during apoptosis in a DNA-bound state. DNA fragmentation factor (DFF) is a complex of the DNase DFF40 (CAD) and its chaperone/inhibitor DFF45 (ICAD-L) that can be activated during apoptosis to induce DNA fragmentation. Here, we demonstrate that DFF directly binds to DNA in vitro without promoting DNA cleavage. DNA binding by DFF is mediated by the nuclease subunit, which can also form stable DNA complexes after release from DFF. Recombinant and reconstituted DFF is catalytically inactive yet proficient in DNA binding, demonstrating that the nuclease subunit in DFF is inhibited in DNA cleavage but not in DNA binding, revealing an unprecedented mode of nuclease inhibition. Activation of DFF in the presence of naked DNA or isolated nuclei stimulates DNA degradation by released DFF40 (CAD). In transfected HeLa cells transiently expressed DFF associates with chromatin, suggesting that DFF could be activated during apoptosis in a DNA-bound state. DNA fragmentation is a biochemical hallmark of apoptotic cell death that can be achieved by the action of several nucleases involved in various apoptotic signal transduction pathways (1Nagata S. Exp. Cell Res. 2000; 256: 12-18Crossref PubMed Scopus (740) Google Scholar, 2Nagata S. Nagase H. Kawane K. Mukae N. Fukuyama H. Cell Death Differ. 2003; 10: 108-116Crossref PubMed Scopus (360) Google Scholar, 3Counis M.F. Torriglia A. Biochem. Cell Biol. 2000; 78: 405-414Crossref PubMed Scopus (88) Google Scholar, 4Wyllie A.H. Nature. 1980; 284: 555-556Crossref PubMed Scopus (4152) Google Scholar). Of paramount importance for cell autonomous apoptotic DNA degradation is the DNA fragmentation factor (DFF) 1The abbreviations used are: DFF, DNA fragmentation factor; CAD, caspase-activated DNase; ICAD-L, inhibitor of CAD large form; ICAD-S, inhibitor of CAD small form; PIPES, 1,4-piperazinediethane-sulfonic acid; CHAPS, 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid; PBS, phosphate-buffered saline; NTA, nitrilotriacetic acid; EMSA, electrophoretic mobility shift assay; GST, glutathione S-transferase; Z, benzyloxycarbonyl; fmk, fluoromethylketone; GFP, green fluorescent protein. (5Liu X. Zou H. Slaughter C. Wang X. Cell. 1997; 89: 175-184Abstract Full Text Full Text PDF PubMed Scopus (1649) Google Scholar). DFF is a heterodimeric complex of the nuclease DFF40 (CAD) and its specific chaperone/inhibitor DFF45 (ICAD-L) (6Liu X. Li P. Widlak P. Zou H. Luo X. Garrard W.T. Wang X. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 8461-8466Crossref PubMed Scopus (501) Google Scholar, 7Enari M. Sakahira H. Yokoyama H. Okawa K. Iwamatsu A. Nagata S. Nature. 1998; 391: 43-50Crossref PubMed Scopus (2807) Google Scholar). During DFF40 (CAD) biosynthesis, DFF45 (ICAD-L) serves as a specific chaperone for the formation of a catalytically competent nuclease (8Sakahira H. Iwamatsu A. Nagata S. J. Biol. Chem. 2000; 275: 8091-8096Abstract Full Text Full Text PDF PubMed Scopus (57) Google Scholar, 9Sakahira H. Nagata S. J. Biol. Chem. 2002; 277: 3364-3370Abstract Full Text Full Text PDF PubMed Scopus (55) Google Scholar). Once the complex between the nuclease and its chaperone/inhibitor is formed, the inhibitory subunit can be cleaved by caspase-3 or granzyme B and in turn dissociates from the nuclease (5Liu X. Zou H. Slaughter C. Wang X. Cell. 1997; 89: 175-184Abstract Full Text Full Text PDF PubMed Scopus (1649) Google Scholar, 10Sakahira H. Enari M. Nagata S. Nature. 1998; 391: 96-99Crossref PubMed Scopus (1424) Google Scholar, 11Thomas D.A. Du C. Xu M. Wang X. Ley T.J. Immunity. 2000; 12: 621-632Abstract Full Text Full Text PDF PubMed Scopus (183) Google Scholar, 12Sharif-Askari E. Alam A. Rheaume E. Beresford P.J. Scotto C. Sharma K. Lee D. DeWolf W.E. Nuttall M.E. Lieberman J. Sekaly R.P. EMBO J. 2001; 20: 3101-3113Crossref PubMed Scopus (115) Google Scholar). The released nuclease is now able to participate in high molecular weight and nucleosomal apoptotic DNA fragmentation (13Widlak P. Acta Biochim. Pol. 2000; 47: 1037-1044Crossref PubMed Scopus (87) Google Scholar). Although initially purified from cytosolic fractions of human and murine cell lines, several lines of evidence indicate that DFF can be a nuclear factor as well (1Nagata S. Exp. Cell Res. 2000; 256: 12-18Crossref PubMed Scopus (740) Google Scholar, 2Nagata S. Nagase H. Kawane K. Mukae N. Fukuyama H. Cell Death Differ. 2003; 10: 108-116Crossref PubMed Scopus (360) Google Scholar, 6Liu X. Li P. Widlak P. Zou H. Luo X. Garrard W.T. Wang X. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 8461-8466Crossref PubMed Scopus (501) Google Scholar, 7Enari M. Sakahira H. Yokoyama H. Okawa K. Iwamatsu A. Nagata S. Nature. 1998; 391: 43-50Crossref PubMed Scopus (2807) Google Scholar, 14Lechardeur D. Drzymala L. Sharma M. Zylka D. Kinach R. Pacia J. Hicks C. Usmani N. Rommens J.M. Lukacs G.L. J. Cell Biol. 2000; 150: 321-334Crossref PubMed Scopus (78) Google Scholar, 15Samejima K. Earnshaw W.C. Exp. Cell Res. 1998; 243: 453-459Crossref PubMed Scopus (45) Google Scholar, 16Ramuz O. Isnardon D. Devilard E. Charafe-Jauffret E. Hassoun J. Birg F. Xerri L. Int. J. Exp. Pathol. 2003; 84: 75-81Crossref PubMed Scopus (33) Google Scholar). DFF40 (CAD) and DFF45 (ICAD-L), but not the small isoform of the inhibitor DFF35 (ICAD-S), display nuclear localization signals (NLS) at their C termini that contribute in an additive manner to the nuclear accumulation of this nuclease/inhibitor complex (14Lechardeur D. Drzymala L. Sharma M. Zylka D. Kinach R. Pacia J. Hicks C. Usmani N. Rommens J.M. Lukacs G.L. J. Cell Biol. 2000; 150: 321-334Crossref PubMed Scopus (78) Google Scholar, 17Samejima K. Earnshaw W.C. Exp. Cell Res. 2000; 255: 314-320Crossref PubMed Scopus (32) Google Scholar). Several nuclear proteins interact with and possibly regulate the activity of DFF40 (CAD), one such factor being topoisomerase IIα, implying a putative cooperation with DFF40 (CAD) in large scale chromosomal DNA fragmentation (18Durrieu F. Samejima K. Fortune J.M. Kandels-Lewis S. Osheroff N. Earnshaw W.C. Curr. Biol. 2000; 10: 923-926Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar, 19Widlak P. Li P. Wang X. Garrard W.T. J. Biol. Chem. 2000; 275: 8226-8232Abstract Full Text Full Text PDF PubMed Scopus (137) Google Scholar). DFF40 (CAD) activity is also enhanced by architectural chromatin proteins such as histone H1 and HMGB1 and 2. For example, DFF40 (CAD) directly binds to histone H1, suggesting that this protein perhaps targets the nuclease to the DNA linker region where nucleosomal DNA fragmentation occurs (6Liu X. Li P. Widlak P. Zou H. Luo X. Garrard W.T. Wang X. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 8461-8466Crossref PubMed Scopus (501) Google Scholar, 19Widlak P. Li P. Wang X. Garrard W.T. J. Biol. Chem. 2000; 275: 8226-8232Abstract Full Text Full Text PDF PubMed Scopus (137) Google Scholar, 20Liu X. Zou H. Widlak P. Garrard W. Wang X. J. Biol. Chem. 1999; 274: 13836-13840Abstract Full Text Full Text PDF PubMed Scopus (161) Google Scholar, 21Toh S.Y. Wang X. Li P. Biochem. Biophys. Res. Commun. 1998; 250: 598-601Crossref PubMed Scopus (40) Google Scholar). In the present study, we provide evidence that DFF in addition to the above mentioned interactions can directly interact with DNA in vitro without inducing DNA cleavage and is able to associate with chromatin in transfected HeLa cells. This is the first known case of a DNase-inhibitor complex in which DNA binding occurs but DNA cleavage is blocked. Intriguingly, DNA binding by DFF prior to or concomitant with its activation by caspase-3 stimulates the activity of the released nuclease DFF40 (CAD) on naked DNA and isolated nuclei, suggesting that during apoptosis nuclear DFF can be activated in a DNA-bound state. Bacterial and Mammalian Expression Vectors—Vectors for the expression of differentially tagged subunits of DFF in mammalian cells, pCI-EGFP-DFF40, pCI-GST-DFF45, and pCS2-MT-DFF40, were described previously (22Scholz S.R. Korn C. Gimadutdinow O. Knoblauch M. Pingoud A. Meiss G. Nucleic Acids Res. 2002; 30: 3045-3051Crossref PubMed Scopus (12) Google Scholar). pcDNA-3.1-EGFP-DFF45 was constructed by inserting the cDNA of DFF45 into a modified pcDNA-3.1 vector, allowing expression of GFP fusion proteins. pEGFP-C2-CTCF was a kind gift from Ru Zhang (Institute for Genetics, JLU-Giessen). For bacterial expression of DFF subunits, the following vectors were used: pGEX-2T-CAD and pACET-DFF45, allowing coexpression of GST-CAD and DFF45 in Escherichia coli as described elsewhere (22Scholz S.R. Korn C. Gimadutdinow O. Knoblauch M. Pingoud A. Meiss G. Nucleic Acids Res. 2002; 30: 3045-3051Crossref PubMed Scopus (12) Google Scholar, 23Meiss G. Scholz S.R. Korn C. Gimadutdinow O. Pingoud A. Nucleic Acids Res. 2001; 29: 3901-3909Crossref PubMed Scopus (22) Google Scholar). pET-Duet1-HisCAD, for the expression of N-terminally His-tagged CAD, was constructed by inserting CAD cDNA into the expression vector pET-Duet-1 (Merck Biosciences). pRSETB-ICAD-L (a kind gift from W. C. Earnshaw) was used to express His-tagged ICAD-L (24Samejima K. Tone S. Kottke T.J. Enari M. Sakahira H. Cooke C.A. Durrieu F. Martins L.M. Nagata S. Kaufmann S.H. Earnshaw W.C. J. Cell Biol. 1998; 143: 225-239Crossref PubMed Scopus (110) Google Scholar). His-tagged ICAD-S was expressed using vector pLK-His-ICAD-S as described previously (22Scholz S.R. Korn C. Gimadutdinow O. Knoblauch M. Pingoud A. Meiss G. Nucleic Acids Res. 2002; 30: 3045-3051Crossref PubMed Scopus (12) Google Scholar). Non-detergent Cell Lysis and Immunodetection of Endogenous DFF— Non-detergent cell lysis was performed using the CNM compartmental protein extraction kit (BioChain Institute, Inc.) according to the supplier's recommendations. Western blotting was performed using polyclonal anti-DFF40 and anti-DFF45/35 antibodies (ProSci, Inc.) in combination with horseradish peroxidase-conjugated secondary antibodies (Merck Biosciences-Calbiochem) and enhanced chemiluminescence detection reagents (ECL) (Amersham Biosciences). Cell Culture and Transfection—Mammalian cells were cultured in a humidified atmosphere of 5% CO2 in Dulbecco's modified Eagle's medium with 10% fetal calf serum, 100 units of penicillin, and 100 μg/ml streptomycin. For in situ nuclear matrix preparations, HeLa cells were grown as described above on glass slides in 9.4-cm dishes and transfected by Ca2+-phosphate co-precipitation using a total of 30 μg of DNA. For preparative nuclear matrix preparations, HeLa cells grown in maxidishes were transfected by calcium phosphate co-precipitation with 60 μg of DNA. Nuclear Matrix Preparations—In situ nuclear matrix preparations were performed as described previously with a slight modification (25Mattern K.A. van der Kraan I. Schul W. de Jong L. van Driel R. Exp. Cell Res. 1999; 246: 461-470Crossref PubMed Scopus (38) Google Scholar). HeLa cells transfected with appropriate GFP fusion constructs were extracted with 0.25% Triton X-100, instead of 1%, in CSK100 buffer (10 mm PIPES, pH 6.8, 0.3 m sucrose, 100 mm NaCl, 3 mm MgCl2, 1 mm EGTA, 1.2 mm phenylmethylsulfonyl fluoride, protease inhibitor mixture) to account for the Triton X-100 sensitivity of DFF. Cells were examined with a Leica TCS4D confocal laser-scanning microscope as described previously (22Scholz S.R. Korn C. Gimadutdinow O. Knoblauch M. Pingoud A. Meiss G. Nucleic Acids Res. 2002; 30: 3045-3051Crossref PubMed Scopus (12) Google Scholar). For in-batch nuclear matrix preparations, transfected HeLa cells were incubated in 0.5% Triton X-100 in CSK100 buffer for 5 min on ice and then passed five times through a 22-gauge injection needle. Cellular debris was removed by centrifugation at 6000 rpm for 5 min, and the pellet washed with 1× CSK100 buffer. Matrix stabilization, chromosomal DNA cleavage, and chromatin depletion were achieved as described above. Chromatin-depleted nuclei were pelleted by centrifugation at 10,000 × g for 20 min and washed twice with CSK50 buffer. The Triton X-100 soluble fraction, matrix-stabilized nuclei, and chromatin-depleted nuclei were analyzed by immunoblotting using rabbit GFP antiserum (α-GFP; Invitrogen) and an anti-rabbit horseradish peroxidase-conjugated secondary antibody (Merck Biosciences-Calbiochem) in combination with enhanced chemiluminescence detection reagents (ECL) (Amersham Biosciences). Production of Recombinant Proteins and in Vitro Mutagenesis—Recombinant His-tagged DFF was produced as described previously with minor modifications (23Meiss G. Scholz S.R. Korn C. Gimadutdinow O. Pingoud A. Nucleic Acids Res. 2001; 29: 3901-3909Crossref PubMed Scopus (22) Google Scholar). Briefly, His-tagged DFF was expressed in E. coli BL21Gold (DE3) cells transformed with the two compatible plasmids pACET-DFF45 and pET-Duet1-HisCAD. The soluble protein fraction containing His-tagged DFF was purified by standard Ni2+-NTA affinity chromatography. Eluted DFF was concentrated using centricon microconcentrators and subjected to anion-exchange chromatography using a Mono-Q HR 5/5 column. Bound protein was eluted in buffer A (20 mm HEPES-KOH, pH 7.4, 100 mm NaCl, 1 mm EDTA, 10% glycerol, 0.01% CHAPS, 10 mm dithiothreitol) using a 15-ml gradient of NaCl (100–300 mm) at a flow rate of 1 ml/min. Peak fractions of recombinant DFF were collected and concentrated as before in buffer A. GST-tagged DFF containing wild-type CAD or the K155Q and H263N CAD mutants were produced as described before (23Meiss G. Scholz S.R. Korn C. Gimadutdinow O. Pingoud A. Nucleic Acids Res. 2001; 29: 3901-3909Crossref PubMed Scopus (22) Google Scholar, 26Korn C. Scholz S.R. Gimadutdinow O. Pingoud A. Meiss G. Nucleic Acids Res. 2002; 30: 1325-1332Crossref PubMed Scopus (25) Google Scholar). Where desired, GST-tagged DFF bound to glutathione-Sepharose 4-B beads in PBS was incubated with thrombin (50 units) overnight to remove the GST tag. The processed complex from the supernatant was purified by anion exchange chromatography as described above. Caspase-3 was expressed and purified as described before (23Meiss G. Scholz S.R. Korn C. Gimadutdinow O. Pingoud A. Nucleic Acids Res. 2001; 29: 3901-3909Crossref PubMed Scopus (22) Google Scholar). His-tagged ICAD-L was expressed using plasmid pRSET-ICAD-L (see above) and His-tagged ICAD-S as described elsewhere (22Scholz S.R. Korn C. Gimadutdinow O. Knoblauch M. Pingoud A. Meiss G. Nucleic Acids Res. 2002; 30: 3045-3051Crossref PubMed Scopus (12) Google Scholar). Briefly, the proteins were first purified by standard Ni2+-NTA affinity chromatography and further purified by anion-exchange chromatography similarly as described above. Electrophoretic Mobility Shift Assays (EMSA)—Standard DNA binding reactions were performed for 30 min in shift buffer (20 mm HEPES-KOH, pH 7.4, 100 mm NaCl, 2 mm EDTA, 10% glycerol, 0.01% CHAPS) containing 12.5 ng/μl plasmid DNA (pBS-VDEX) and indicated amounts of proteins (see figures) in a final volume of 20 μl. Aliquots of the binding reactions were transferred into shift loading buffer (75% sucrose, 0.1% bromphenol blue), loaded onto 0.8% agarose gels containing 0.05 μg/ml ethidium bromide and separated by electrophoresis in TBE buffer (Tris borate/EDTA, pH 8.3). Aliquots of the binding reactions with PCR products were analyzed by electrophoresis on 6 or 5% polyacrylamide gels in TPE buffer (Tris phosphate, pH 8.2). The gels were stained with ethidium bromide. UV Cross-linking—A-39 bp double-stranded oligodeoxynucleotide at a concentration of 5 μm with photoactivatable 5-iododeoxyuridine incorporated at three adjacent sites in the center of one DNA strand radioactively labeled with 32P at its 5′-end was incubated with 5 μm DFF in buffer A. UV irradiation of the DFF-DNA complex with a helium/cadmium laser at 325 nm for 30 min was applied to induce cross-linking. Aliquots of the reaction mix were separated by PAGE and analyzed by autoradiography. Transmission Electron Microscopy—Plasmid DNA and protein were incubated at 37 °C for 30 min in a 10-μl reaction volume containing 30 ng of DNA and different concentrations of DFF ranging from 0 to 300 ng (molar ratios of enzyme to DNA between 30:1 and 300:1) in shift buffer. Complexes were fixed with 0.2% (v/v) glutaraldehyde for 10 min at 37 °C and, after 3-fold dilution in 10 mm triethanolamine chloride, pH 7.5, and 10 mm MgCl2, adsorbed to freshly cleaved mica. Micrographs were taken using a Philips CM100 electron microscope at 100 kV and a Fastscan CCD camera (Tietz Video and Image Processing Systems GmbH, Gauting, Germany). Reconstitution of DFF—For reconstitution of DFF we activated GST-tagged DFF (5 μm final concentration) in shift buffer (see above) or shift buffer supplemented with Mg2+ (5 mm final concentration) containing plasmid DNA (pBS-VDEX, 12.5 ng/μl) with caspase-3, then added pancaspase inhibitor z-VAD-fmk (88 μm final concentration) and substituted the reaction mix with recombinant ICAD-L or ICAD-S, respectively, at 15 and 30 μm final concentration, corresponding to 3- and 6-fold molar excess of inhibitory subunits over free nuclease. Aliquots of the reaction were mixed with shift loading buffer and analyzed by agarose gel electrophoresis in TBE buffer (Tris borate/EDTA, pH 8.3). Time Order of Addition Experiments—To examine the cleavage activity of DFF premixed with DNA prior to caspase-3 activation, we incubated His-DFF (20 nm) with plasmid DNA (pBS-VDEX, 25 ng/μl) for 5 min, added recombinant caspase-3 and then stopped caspase-3 processing of DFF after 15 min by addition of the pan-caspase inhibitor z-VAD-fmk (Sigma-Aldrich). DNA cleavage was initiated by the addition of MgCl2 to a final concentration of 5 mm. To analyze the cleavage activity of DFF activated by caspase-3 in the absence of DNA, we incubated recombinant DFF (20 nm) with caspase-3 for 15 min and stopped caspase-3 activation of DFF by adding pan-caspase inhibitor. Activated DFF was then incubated with DNA (pBS-VDEX, 25 ng/μl) for 20 min, and the cleavage reaction started by addition of MgCl2 as above. To analyze DNA cleavage activity of DFF incubated with DNA concomitant with caspase-3 activation, we mixed DFF (20 nm), DNA (pBS-VDEX, 25 ng/μl), and caspase-3, stopped caspase-3 activation of DFF after 15 min by addition of pan-caspase inhibitor and then started the DNA cleavage reaction after a further 5 min incubation by addition of MgCl2. Aliquots of the DNA cleavage reactions were taken at 0 (immediately after addition of MgCl2), 1, 3, 9, and 15 min time points, loaded onto 0.8% agarose gels and analyzed by electrophoresis as described above. To quantify DNA cleavage, the decrease of the supercoiled plasmid DNA band in the gels was measured and transformed into relative cleavage rates. To use chromatin as substrate, nuclei were isolated from HeLa cells by standard centrifugation procedures using Tris-buffered saline (TBS, 50 mm Tris-HCl, 200 mm NaCl, 3 mm KCl, 0.02% sodium azide, pH 7.5) and a cell lysis buffer (0.325 m sucrose, 10 mm Tris-HCl, pH 7.8, 5 mm MgCl2, 1% Triton X-100) and cleaved using 50 nm DFF. Inhibited and Released DFF40 (CAD) but Not DFF45 (ICAD-L) Interact with DNA—In an attempt to identify the DNA binding domain of DFF40 (CAD) we to our surprise found that not only the released nuclease subunit itself but also the nuclease/inhibitor complex DFF forms stable DNA complexes in vitro. In the EMSA shown in Fig. 1B, we used GST-tagged wild-type DFF and the active site variant H263N of murine DFF40 (CAD) as well as recombinant DFF45 and GST with plasmid DNA as substrate. His263 is an important catalytic residue of murine DFF40 (CAD) situated in the active center close to the C terminus of the protein (Fig. 1A) (23Meiss G. Scholz S.R. Korn C. Gimadutdinow O. Pingoud A. Nucleic Acids Res. 2001; 29: 3901-3909Crossref PubMed Scopus (22) Google Scholar, 27Sakahira H. Takemura Y. Nagata S. Arch Biochem. Biophys. 2001; 388: 91-99Crossref PubMed Scopus (33) Google Scholar, 28Woo E.J. Kim Y.G. Kim M.S. Han W.D. Shin S. Robinson H. Park S.Y. Oh B.H. Mol. Cell. 2004; 14: 531-539Abstract Full Text Full Text PDF PubMed Scopus (89) Google Scholar). The H263N variant of murine DFF40 (CAD) was used to see, whether this catalytically inactive enzyme retains the ability to bind DNA despite of its inability to cleave it. Recombinant DFF45 and GST were analyzed for control. As shown in Fig. 1B, DFF and the nuclease subunits released from DFF, but neither the inhibitory subunit DFF45 nor GST, form stable DNA complexes. Formation of these complexes is not dependent on the presence of Mg2+, which is an essential cofactor for DNA cleavage by the nuclease DFF40 (CAD) as seen in Fig. 1B (27Sakahira H. Takemura Y. Nagata S. Arch Biochem. Biophys. 2001; 388: 91-99Crossref PubMed Scopus (33) Google Scholar). In the absence of Mg2+, the wild-type nuclease and the H263N variant induce a similar mobility shift in the DNA. In the presence of Mg2+, the wild-type nuclease readily cleaves the DNA, loosing its ability to induce a mobility shift (Fig. 1B, lane 13), whereas the active site variant H263N cannot cleave but can still bind the DNA (Fig. 1B, lane 15). These results clearly show that not only the nuclease subunit released from DFF but also the unprocessed nuclease/inhibitor complex DFF can form stable DNA complexes, which is highly unusual both for nonspecific nucleases and nuclease/inhibitor complexes. The data also suggest that the nuclease subunit in DFF mediates DNA binding of the nuclease/inhibitor complex, since the inhibitory subunit alone does not form DNA complexes (Fig. 1B, lanes 7 and 16). EMSA with a second inactive variant of murine DFF40 (CAD) with substitution of Lys155 by Gln supports this suggestion. Lys155 in murine DFF40 (CAD) is not an active site residue but closer to the CAD/CIDE-N-domain of the enzyme and appears to be involved in maintaining the protein structure and/or DNA binding (Fig. 1A) (23Meiss G. Scholz S.R. Korn C. Gimadutdinow O. Pingoud A. Nucleic Acids Res. 2001; 29: 3901-3909Crossref PubMed Scopus (22) Google Scholar, 26Korn C. Scholz S.R. Gimadutdinow O. Pingoud A. Meiss G. Nucleic Acids Res. 2002; 30: 1325-1332Crossref PubMed Scopus (25) Google Scholar). In this EMSA we used thrombin-treated GST-DFF complexes in order to rule out that the binding activity seen is an artifact caused by the GST fusion (Fig. 1C). Interestingly, DFF with the K155Q mutant of CAD, DFFN-K155Q, does not form stable DNA complexes whereas DFFN-H263N and wild-type DFF do (Fig. 1C). This corroborates that the nuclease subunit in DFF mediates DNA binding and suggests that Lys157 (Lys155) in DFF40 (CAD) directly or indirectly plays an important role in DNA binding. The recently published crystal structure of murine DFF40 (CAD) illustrates that Lys155 is a buried residue stabilizing the N-terminal end of α-helix 4, which presumably binds in the major groove of the DNA (Fig. 1A) (28Woo E.J. Kim Y.G. Kim M.S. Han W.D. Shin S. Robinson H. Park S.Y. Oh B.H. Mol. Cell. 2004; 14: 531-539Abstract Full Text Full Text PDF PubMed Scopus (89) Google Scholar). Recombinant DFF Binds to Various DNA Substrates in Vitro—So far our results revealed that DFF forms stable DNA complexes in vitro via its nuclease subunit DFF40 (CAD). In order to further characterize the DNA binding capacity of DFF we produced recombinant DFF with the nuclease subunit carrying an N-terminal extension of 12 amino acid residues [MGRSH6KL] including a hexahistidine tag to facilitate purification and replacing the GST tag (Fig. 2A). This protein complex was used to carry out EMSA as well as UV cross-linking experiments with DFF and various DNA substrates including plasmid DNA as well as synthetic PCR products and oligodeoxynucleotides. As shown in Fig. 2B, His-tagged DFF, and the nuclease released from DFF by caspase-3 processing induce a mobility shift in the DNA. As seen from lane 4 in Fig. 2B, nuclease released from DFF by caspase-3 treatment in the absence of Mg2+ led to high molecular weight complexes that stick to the wells of the agarose gel, which can be explained by the fact that released DFF40 (CAD) has the propensity to form oligomers (20Liu X. Zou H. Widlak P. Garrard W. Wang X. J. Biol. Chem. 1999; 274: 13836-13840Abstract Full Text Full Text PDF PubMed Scopus (161) Google Scholar, 29Widlak P. Lanuszewska J. Cary R.B. Garrard W.T. J. Biol. Chem. 2003; PubMed Google Scholar). Such high molecular weight DFF40 (CAD)-DNA complexes might be unable to enter the gel matrix. In some cases with free DFF40 (CAD) we also observed a smear instead of homogenous complex formation indicating a non-homogenous distribution of DNA complexes and suggesting that under the conditions applied the DFF40 (CAD)-DNA complexes are less stable than the DFF-DNA complexes. DFF also binds to synthetic DNA such as PCR products of various sizes (Fig. 2C) as well as to a 39-bp double-stranded oligodeoxynucleotide as demonstrated by UV cross-linking (Fig. 2D), apparently with a greater affinity toward larger DNA fragments. Titration of a 273-bp double-stranded PCR-derived DNA fragment with DFF in the presence (data not shown) or absence of Mg2+ revealed a dissociation constant Kd ≈ 5 × 10-6m (Fig. 2E). The sequences of the plasmid DNA, PCR products, and the oligodeoxynucleotide were unrelated, indicating that, as expected for a nonspecific nuclease, DFF binding to DNA is not sequence specific. Visualization of DFF-DNA Complex Formation and DNA Cleavage by Transmission Electron Microscopy—In order to obtain additional information about the nature of the DFF-DNA complexes formed we analyzed DNA binding and cleavage by DFF using transmission electron microscopy. When plasmid DNA was incubated with increasing amounts of DFF at physiological salt concentration the decrease in electrophoretic mobility of the DFF-DNA complex compared with the free plasmid DNA analyzed by EMSA was dependent on the concentration of DFF (Fig. 3A). When such complexes were analyzed by transmission electron microscopy, similar results were obtained (Fig. 3A). Depending on concentration, DFF binds to plasmid DNA in a nonspecific manner, independent of the absence or presence of Mg2+. As seen from Fig. 3A, DFF forms large aggregates on the DNA rather than sitting side by side on its substrate and more than one plasmid molecule is bound by such aggregates. When the nuclease subunit released from DFF by caspase-3 treatment was analyzed in the same way (Fig. 3B), we again observed nonspecific DNA binding in the absence of Mg2+ but, as expected, cleavage of the DNA into short linear fragments in the presence of Mg2+, corroborating the results obtained from the gel shift experiments. Recombinant and Reconstituted DFF Allow Formation of Stable DFF-DNA Complexes—Our results indicated that the free nuclease released from DFF and recombinant DFF obtained from co-expression of its subunits interact with DNA. It was pertinent to investigate if reconstitution of DFF, i.e. addition of inhibitory subunit to the nuclease subunit released from activated DFF results in the restoration of a DNA binding proficient but catalytically deficient nuclease/inhibitor complex. To this end, we have activated recombinant DFF using caspase-3, then added the pan-caspase inhibitor z-VAD-fmk and finally supplemented the reaction mixture with recombinant ICAD-S (DFF35) or ICAD-L (DFF45) again (Fig. 4A). As shown, reconstituted DFF is able to bind to the DNA substrate similar to recombinant DFF obtained from co-expression of its subunits, allowing formation of a stable DNA complex but inhibiting the activity of the nuclease subunit (Fig. 4B). When the reconstituted DFF complexes were activated by excess caspase-3 (indicated by asterisks in Fig. 4), DNA cleavage by DFF40 (CAD) was regained. This result clearly shows that inhibition of the nuclease subunit in DFF is due to blocking the catalytic activity but not the DNA binding capacity of DFF40 (CAD). The results also suggest that ICAD-S (DFF35) and ICAD-L (DFF45) use the same mechanism for inhibition of DFF40 (CAD). To our knowledge, this mode of nuclease inhibition has not been seen yet with any other known nuclease/inhibitor complex (see below). DNA Binding by DFF Stimulates the Activity of Released DFF40 (CAD)—To elucidate if binding of DFF to DNA prior to or concomitant with its activation by caspase-3 has an influence on the activity of DFF40 (CAD), we performed time order of addition experiments. We compared the DNA cleavage kinetics of DFF40 (CAD) released from DFF incubated with DNA prior to or concomitant with its activation by caspase-3 with DNA cleavage kinetics of DFF40 (CAD) released from DFF activated by caspase-3 in the absence of DNA. Under the conditions applied, incubation of DFF with DNA prior to or concomitant with its activation by caspase-3 clearly stimulated the activity of released DFF40 (CAD) (Fig. 5A). This suggests that activation of DNA-bound DFF leads to an apparently higher cleavage efficiency of the released nuclease subunit compared with the nuclease released in the absence of DNA. When the experiment was conduct