Footprinting of Yeast DNA Topoisomerase II Lysyl Side Chains Involved in Substrate Binding and Interdomainal Interactions
1997; Elsevier BV; Volume: 272; Issue: 49 Linguagem: Inglês
10.1074/jbc.272.49.31190
ISSN1083-351X
Autores Tópico(s)Synthesis and bioactivity of alkaloids
ResumoFootprinting of yeast DNA topoisomerase II and its NH2- and COOH-terminal truncation derivatives was carried out to map the locations of lysyl side chains that are involved in enzyme-DNA interaction, in the binding of ATP, or in interaction between domains of the same enzyme molecule. Several conclusions were drawn based on these measurements and the crystal structures of a 92-kDa fragment of the yeast enzyme and a 43-kDa fragment ofEscherichia coli gyrase B-subunit. First, the footprinting results support the model previously inferred from the 92-kDa fragment crystal structure that the main site of DNA binding is comprised of a pair of semicircular grooves. Second, the binding of a nonhydrolyzable ATP analog to the yeast enzyme appears to affect citraconylation at a minimum of six lysines in the ATPase domain of each polypeptide. Two of these lysines are probably involved in contacting the nucleotide directly, and one probably becomes buried when the two ATPase domains of a dimeric enzyme come into contact upon ATP binding; for the others, changes in lysine reactivity appear to reflect allosteric changes following ATP binding. Third, from a comparison of the footprint of the intact enzyme and those of the truncated polypeptides comprised of either the NH2- or the COOH-terminal half of the intact polypeptide, it appears that there are few contacts between the NH2- and COOH-terminal half of yeast DNA topoisomerase II. Footprinting of yeast DNA topoisomerase II and its NH2- and COOH-terminal truncation derivatives was carried out to map the locations of lysyl side chains that are involved in enzyme-DNA interaction, in the binding of ATP, or in interaction between domains of the same enzyme molecule. Several conclusions were drawn based on these measurements and the crystal structures of a 92-kDa fragment of the yeast enzyme and a 43-kDa fragment ofEscherichia coli gyrase B-subunit. First, the footprinting results support the model previously inferred from the 92-kDa fragment crystal structure that the main site of DNA binding is comprised of a pair of semicircular grooves. Second, the binding of a nonhydrolyzable ATP analog to the yeast enzyme appears to affect citraconylation at a minimum of six lysines in the ATPase domain of each polypeptide. Two of these lysines are probably involved in contacting the nucleotide directly, and one probably becomes buried when the two ATPase domains of a dimeric enzyme come into contact upon ATP binding; for the others, changes in lysine reactivity appear to reflect allosteric changes following ATP binding. Third, from a comparison of the footprint of the intact enzyme and those of the truncated polypeptides comprised of either the NH2- or the COOH-terminal half of the intact polypeptide, it appears that there are few contacts between the NH2- and COOH-terminal half of yeast DNA topoisomerase II. DNA topoisomerase II of the budding yeastSaccharomyces cerevisiae is one of the most extensively studied type II DNA topoisomerases (E.C. 5.99.1.3). The enzyme is a dimer of two identical polypeptides encoded by a single-copy gene (1Goto T. Wang J.C. J. Biol. Chem. 1982; 257: 5866-5872Abstract Full Text PDF PubMed Google Scholar, 2Goto T. Wang J.C. Cell. 1984; 36: 1073-1080Abstract Full Text PDF PubMed Scopus (119) Google Scholar; for a recent review, see Ref. 3Wang J.C. Annu. Rev. Biochem. 1996; 65: 635-692Crossref PubMed Scopus (2076) Google Scholar). Recent biochemical and crystallographic studies have led to a detailed model on how this enzyme couples ATP binding and hydrolysis to the transport of one DNA double helix through another (4Roca J. Wang J.C. Cell. 1992; 71: 833-840Abstract Full Text PDF PubMed Scopus (297) Google Scholar, 5Roca J. Wang J.C. Cell. 1994; 77: 609-616Abstract Full Text PDF PubMed Scopus (251) Google Scholar, 6Berger J.M. Gamblin S.J. Harrison S.C. Wang J.C. Nature. 1996; 379: 225-232Crossref PubMed Scopus (748) Google Scholar, 7Roca J. Berger J.M. Harrison S.C. Wang J.C. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 4057-4062Crossref PubMed Scopus (164) Google Scholar). The enzyme is believed to act as an ATP-modulated clamp with two molecular gates at opposite ends. One of the two gates, termed the N-gate, is formed by the pair of amino-terminal ATPase domains of the protein. This gate is mainly in the open state in the absence of bound ATP. Upon ATP-binding, the NH2-terminal domains come into contact to close the N-gate (4Roca J. Wang J.C. Cell. 1992; 71: 833-840Abstract Full Text PDF PubMed Scopus (297) Google Scholar, 5Roca J. Wang J.C. Cell. 1994; 77: 609-616Abstract Full Text PDF PubMed Scopus (251) Google Scholar, 8Wigley D.B. Davies G.J. Dodson E.J. Maxwell A. Dodson G. Nature. 1991; 351: 624-629Crossref PubMed Scopus (487) Google Scholar, 9Ali J.A. Orphanides G. Maxwell A. Biochemistry. 1995; 34: 9801-9808Crossref PubMed Scopus (90) Google Scholar). The second gate, termed the C-gate, is formed mainly by amino acid residues 1030–1045 and 1113–1129 of the yeast enzyme near its COOH terminus (6Berger J.M. Gamblin S.J. Harrison S.C. Wang J.C. Nature. 1996; 379: 225-232Crossref PubMed Scopus (748) Google Scholar, 7Roca J. Berger J.M. Harrison S.C. Wang J.C. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 4057-4062Crossref PubMed Scopus (164) Google Scholar). 1Numbering of amino acid residues was corrected for the presence of an extra asparagine at position 72 of the reported sequence. According to the two-gate protein-clamp model, the C-gate is usually in the closed state and opens only transiently for the exit of a duplex DNA segment, termed the T-segment, after its entrance through the N-gate and passage through an enzyme-bound DNA segment termed the G-segment. The passage of the T-segment through the G-segment involves transesterification between a pair of active-site tyrosines, Tyr782 (10Worland S.T. Wang J.C. J. Biol. Chem. 1989; 264: 4412-4416Abstract Full Text PDF PubMed Google Scholar), and a pair of DNA phosphoryl groups that are 5′-staggered and separated by 4 base pairs. This reaction breaks the pair of phosphodiester bonds in the G-segment and at the same time forms a pair of phosphotyrosine links between the enzyme and the DNA; an opening or gate in the G-segment is thus created. Following the passage of the T-segment, a second transesterification disrupts the phosphotyrosine links and rejoins the DNA strands to close the G-segment gate (3Wang J.C. Annu. Rev. Biochem. 1996; 65: 635-692Crossref PubMed Scopus (2076) Google Scholar). The protein-clamp model postulates that upon the entrance of a DNA T-segment through an open N-gate of a DNA-bound enzyme, closure of the N-gate triggered by ATP-binding entraps the T-segment. Steric repulsion between the NH2-terminal domains of the protein and the entrapped T-segment in turn widens the gate in the G-segment for the passage of the T-segment into a large hole between the G-segment and the C-gate. Following this DNA passage event, the widely separated protein-linked DNA ends retract toward each other to reform the DNA phosphodiester bonds, reducing the size of the hole containing the T-segment and forcing the expulsion of the T-segment through the C-gate (6Berger J.M. Gamblin S.J. Harrison S.C. Wang J.C. Nature. 1996; 379: 225-232Crossref PubMed Scopus (748) Google Scholar, 7Roca J. Berger J.M. Harrison S.C. Wang J.C. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 4057-4062Crossref PubMed Scopus (164) Google Scholar). Because of the high degree of homology between the amino acid sequences of yeast DNA topoisomerase II and its homologues (see, for example, Ref. 11Caron P.R. Wang J.C. Liu L. DNA Topoisomerase and Their Applications in Pharmacology: A Volume of Advances in Pharmacology. 29B. Academic Press, San Diego1994: 271-297Google Scholar), many features of the model deduced mainly from studies of the yeast enzyme are most likely to hold for its homologues. The three-dimensional structural details of the above model are mainly based on the crystal structures of a 92-kDa yeast DNA topoisomerase II fragment (6Berger J.M. Gamblin S.J. Harrison S.C. Wang J.C. Nature. 1996; 379: 225-232Crossref PubMed Scopus (748) Google Scholar), which lacks the ATPase domain of a functional enzyme, and the ATPase domain of a bacterial gyrase B-subunit fragment (8Wigley D.B. Davies G.J. Dodson E.J. Maxwell A. Dodson G. Nature. 1991; 351: 624-629Crossref PubMed Scopus (487) Google Scholar). Electron microscopic images of several intact and truncated type II DNA topoisomerases also provided low resolution structural features (12Kirchhausen T. Wang J.C. Harrison S.C. Cell. 1985; 41: 933-943Abstract Full Text PDF PubMed Scopus (94) Google Scholar, 13Schultz P. Olland S. Oudet P. Hancock R. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 5936-5940Crossref PubMed Scopus (43) Google Scholar, 14Benedetti P. Silvestri A. Fiorani P. Wang J.C. J. Biol. Chem. 1997; 272: 12132-12137Abstract Full Text Full Text PDF PubMed Scopus (23) Google Scholar). In the crystal structure of the yeast protein, interaction between the DNA G-segment and the yeast enzyme was modeled by fitting 10–12 base pairs of B-form duplex DNA into a positively charged semicircular groove in each of the 92-kDa yeast DNA topoisomerase II fragment. It was also found that if a duplex DNA possesses a 4-nucleotide 5′-overhang, the single-stranded overhang could be joined to the active site tyrosine 782 through a narrow channel at one end of the semicircular groove (6Berger J.M. Gamblin S.J. Harrison S.C. Wang J.C. Nature. 1996; 379: 225-232Crossref PubMed Scopus (748) Google Scholar). These observations led to the hypothesis that the conformation of the enzyme observed in the particular crystal closely resembles that of the enzyme after it has cleaved the DNA G-segment and widened the DNA gate (6Berger J.M. Gamblin S.J. Harrison S.C. Wang J.C. Nature. 1996; 379: 225-232Crossref PubMed Scopus (748) Google Scholar). Direct evidence in support of this model was lacking, however. In the present work, we applied the two-step lysine footprinting method (15Hanai R. Wang J.C. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 11904-11908Crossref PubMed Scopus (41) Google Scholar) to map the lysyl residues of yeast DNA topoisomerase II that became protected against citraconylation upon the binding of DNA. In addition, modulation of the lysine reactivities by the binding of AMP-PNP, 2The abbreviations used are: AMP-PNP, adenosine 5′-(β,γ-imido)triphosphate; HMK, heart muscle kinase; HA, a flu virus hemagglutinin epitope with amino acid sequence YPYDVP; HEPPS, 4-(2-hydroxyethyl)-1-piperazinepropanesulfonic acid. the nonhydrolyzable β,γ-imido analog of ATP, and by interdomainal interactions within a dimeric enzyme molecule, were also studied. The results of these measurements in solution support several structural and mechanistic features of the enzyme that were implicated by previous x-ray crystallographic and electron microscopic results (6Berger J.M. Gamblin S.J. Harrison S.C. Wang J.C. Nature. 1996; 379: 225-232Crossref PubMed Scopus (748) Google Scholar, 8Wigley D.B. Davies G.J. Dodson E.J. Maxwell A. Dodson G. Nature. 1991; 351: 624-629Crossref PubMed Scopus (487) Google Scholar, 13Schultz P. Olland S. Oudet P. Hancock R. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 5936-5940Crossref PubMed Scopus (43) Google Scholar, 14Benedetti P. Silvestri A. Fiorani P. Wang J.C. J. Biol. Chem. 1997; 272: 12132-12137Abstract Full Text Full Text PDF PubMed Scopus (23) Google Scholar). All expression clones were derived from YEpTOP2-PGAL1, which was originally constructed for overexpression of full-length yeast DNA topoisomerase II from an inducible promoter of the yeast GAL1 gene (16Giaever G.N. Snyder L. Wang J.C. Biophys. Chem. 1988; 29: 7-15Crossref PubMed Scopus (66) Google Scholar). NH2-terminal tagging of the polypeptides was done through the use of a cloned segmentGTA(TACCCATATGACGTCCCG)GACTAC(GCGCGCCGTGCATCT) GTACA-3′ containing codons for a flu virus hemagglutinin epitope YPYDVP and a heart muscle kinase (HMK) phosphorylation site RRASV (codons in parentheses; see Refs. 17Kolodziej P.A.S. Young R.A. Methods Enzymol. 1991; 194: 508-519Crossref PubMed Scopus (423) Google Scholar and 18Kennelly P.J. Krebs E.G. J. Biol. Chem. 1991; 266: 15555-15558Abstract Full Text PDF PubMed Google Scholar). After digestion withBst1107I and BsrGI at the underlined hexanucleotide sequences shown above, the BsrGI end was repaired and SmaI/XmaI linker molecules 5′-ACCCGGGT-3′ were added to both ends. Cutting within the added linkers by XmaI yielded a fragment with 5′-CCGG overhangs, which was inserted into the AgeI site near the beginning of the yeast DNA topoisomerase II gene. A cloned sequence 5′-GCGCGCCGTGCATCTGTACACCACCACCACCACCACTAGT-3′, which contains codons for an alanine, the HMK phosphorylation site, six histidines and termination, was used for adding an HMK phosphorylation site and/or hexahistidine tag to the COOH termini of polypeptides (the three underlined hexanucleotide sequences GCGCGC,TGTACA, and ACTAGT are the restriction sites for BssHII, BsrGI, and SpeI, respectively). To fuse the HMK-(His)6 tag to codon 1196 of yeast DNA topoisomerase II, YEpTOP2-PGAL1 was cut withHindIII and repaired, and the linker 5′-TGCGCGCA-3′ was then added to introduce a BssHII site downstream of theHindIII site. The segment from this BssHII site to an NheI site further downstream was then replaced by theBssHII to SpeI fragment from the tag sequence. Fusion of the hexahistidine tag to codon 1196 was similarly done, using a linker 5′-TGCTGTACAGCA-3′ to provide a BsrGI site downstream of the HindIII site for replacement by theBsrGI to SpeI segment containing the hexahistidine tag. For the construction of expression clones for HA-HMK-Top2(1–659)-(His)6 and Top2(660–1196)-HMK-(His)6, site-directed mutagenesis was first used to convert codons 659 and 660 (CCT-GGT) in a full-length yeast TOP2 clone to CCCGGG, thus introducing a unique SmaI/XmaI site. In the construction of Top2(660–1196)-HMK-(His)6, this Top2(SmaI/XmaI) plasmid was cut withSmaI, and a pair of complementary oligonucleotides 5′-GATCCGTAACCATGTCA-3′ and 5′-TGACATGGTTACG-3′ were used to link codon 660 to the BamHI site downstream of the GAL1 promoter in YEpTOP2-PGAL1. The insertion of this linker also added an initiation methionine codon and a serine codon before codon 660 of the yeast protein. Replacing the COOH-terminal codons of this clone with those encoding an HMK-(His)6 tag after codon 1196 yielded the desired clone. To express HA-HMK-Top2(1–659)-(His)6, the clone for the expression of Top2(1–1196)-HMK-(His)6 was first cut with BsrGI, and a linker 5′-ACCCCCGGGGGT-3′ was added to the repaired ends. After cutting withBamHI and XmaI, the BamHI toXmaI fragment from Top2(SmaI/XmaI) was inserted to yield a clone expressing Top2(1–659)-(His)6. To add the HA-HMK tag to the NH2-terminal region of this clone, the BamHI to KpnI segment in it was replaced by the same segment from the plasmid for the expression of HA-HMK-Top2(1–1428). Overexpression and purification of HA-HMK-Top2(1–1428) followed the procedures previously described for the full-length yeast enzyme, in which cell lysate cleared by centrifugation was fractionated by chromatography on a phosphocellulose column (10Worland S.T. Wang J.C. J. Biol. Chem. 1989; 264: 4412-4416Abstract Full Text PDF PubMed Google Scholar, 19.Berger, J. M., Structural Determination of a 92-kDa Fragment of Yeast Topoisomerase II by X-ray Crystallography at 2.7 Å Resolution. Ph.D. thesis, 1995, Harvard University.Google Scholar). For the hexahistidine-tagged proteins, chromatography on the Ni-NTA column (Qiagen) was incorporated as one of the purification steps: HA-HMK-Top2(1–659)-(His)6 was purified successively on Ni-NTA, heparin, and phosphocellulose, Top2(660–1196)-HMK-(His)6 on Ni-NTA and MonoQ (Sigma), and Top2(1–1196)-HMK-(His)6 on phosphocellulose, Ni-NTA and MonoQ. Purified proteins were dialyzed into 50 mm HEPPS, pH 8.2, 50 mm KCl, 1 mm each of EDTA and EGTA, 2 mm 2-mercaptoethanol, 0.4 μg/ml each of leupeptin and pepstatin, and 10% (v/v) glycerol, flash frozen in liquid nitrogen, and stored at −70 °C until use. The previously published procedures (15Hanai R. Wang J.C. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 11904-11908Crossref PubMed Scopus (41) Google Scholar, 20Hanai R. Liu Z. Benedetti P. Caron P.R. Lynch A.S. Wang J.C. J. Biol. Chem. 1996; 271: 17469-17475Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar) were followed with minor modifications. The addition of Tris buffer after treatment with citraconic anhydride was omitted, and the citraconylated protein was used directly in lysine acetylation byN-hydroxysuccinimide. After the acetylation step, Tris·HCl, pH 8.5, was added to 100 mm final concentration to stop the reaction, and the samples were dialyzed against 50 mm Tris·HCl, pH 8.5, for 3 h. Citraconylated lysines were then deblocked by dialysis at 4 °C against two changes of 30% acetic acid over a period of 20 h. The polypeptide was then dialyzed against two changes of 50 mm Tris·HCl, pH 8.5, 5 mm EDTA, and 0.01% SDS over a period of 6 h. Lys-C endoprotease (100 ng) was added to each peptide solution recovered from the dialysis bag, and the mixture was incubated at 30 °C for 12 h. Proteolyzed samples were stored at −20 °C until use. Radiolabeling of the stored samples by treatment with heart muscle kinase and sizing of the labeled peptides by SDS-polyacrylamide gel electrophoresis were done as described before (20Hanai R. Liu Z. Benedetti P. Caron P.R. Lynch A.S. Wang J.C. J. Biol. Chem. 1996; 271: 17469-17475Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar). Length measurements of the footprinting products were done as before, using partial cleavage at cysteines or methionines of the same tagged polypeptides as length standards (15Hanai R. Wang J.C. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 11904-11908Crossref PubMed Scopus (41) Google Scholar). The mobilities of these length standards were plotted directly against their sizes (expressed as the positions of cleavage), and a smooth calibration curve was drawn through these points. Because the cysteine reaction at a residuei breaks the polypeptide on the NH2-terminal side, the labeled cleavage product from an NH2-terminal-labeled protein was assigned a nominal length of i-1 and that from a COOH-terminal tagged protein a lengthi; for the methionine cleavage reaction, which occurs on the COOH-terminal side of a methionine, the respective products were assigned lengths i and i-1. The mobilities of the footprinting products, which were analyzed on the same gel as the calibration standards tagged in the same way, were then used to determine their sizes by interpolation between adjacent points on the calibration curve. In case extrapolation rather than interpolation was involved, the calibration curve was plotted in the standard semilogarithmic form to permit linear extrapolation. In all size assignments, no correction was made for the presence of modified amino acid residues at the ends of the size markers. Treatment with 2-nitro-5-cyanobenzoic acid gives COOH-terminal fragments that begin with 2-imidothiazolidine-4-carboxyl groups, and treatment with cyanogen bromide gives NH2-terminal fragments that end with a lactone ring (see for example, Ref. 21Lundblad R.L. Techniques in Protein Modification. CRC Press, Roca Raton, FL1994Google Scholar). The effects of these modifications on the electrophoretic mobilities of the fragments in an SDS-polyacrylamide gel were negligible within experimental error. Because of its large size (1428 amino acids), footprinting of the single-polypeptide yeast DNA topoisomerase II was carried out with both NH2-terminal and COOH-terminal-tagged derivatives of it to improve the resolution of the footprinting products by SDS-polyacrylamide gel electrophoresis. In Fig. 1, the schematics a-dillustrate the various tagged derivatives used in these experiments and schematics e and f depict the positions of the methionines, cysteines, and lysines in yeast DNA topoisomerase II. The fragments a and b were constructed to carry two motifs HA and HMK at their NH2 termini: the former constitutes a flu virus hemagglutinin-derived epitope that is recognized by a commercially available monoclonal antibody (17Kolodziej P.A.S. Young R.A. Methods Enzymol. 1991; 194: 508-519Crossref PubMed Scopus (423) Google Scholar), and the latter constitutes a heart muscle kinase phosphorylation site (18Kennelly P.J. Krebs E.G. J. Biol. Chem. 1991; 266: 15555-15558Abstract Full Text PDF PubMed Google Scholar). In the polypeptides depicted in c and d, the HMK-tag was placed near their COOH termini. For convenience of purification, three of the polypeptides were also tagged with a stretch of six histidines at their COOH termini (22Hochuli E. Bannwarth W. Döbeli H. Gentz R. Stüber D. BioTechnology. 1988; 6: 1321-1325Crossref Scopus (964) Google Scholar). The HA tag allows the detection of fragments containing it by immunoblotting with a monoclonal antibody, and the HMK tag by direct 32P labeling of the tag by treatment with the kinase (18Kennelly P.J. Krebs E.G. J. Biol. Chem. 1991; 266: 15555-15558Abstract Full Text PDF PubMed Google Scholar). The use of the HA and HMK tags in protein footprinting corresponds to, respectively, the indirect and direct end-labeling approach in nucleic acid sequencing or footprinting. Control experiments with untagged yeast DNA topoisomerase II and its truncation derivatives showed that they were neither detectable by immunoblotting with the HA-specific antibody nor phosphorylated by heart muscle kinase (data not shown). In all experiments described below, only the 32P labeling method was employed. For the NH2-terminal tagged protein HA-HMK-Top2(1–1428) and the COOH-terminal tagged protein Top2(1–1196)-HMK-(His)6, measurements of their relaxation of supercoiled DNA were carried out, using the untagged proteins as references. Tagging at these positions was observed to have no significant effect on the relaxation activities of the purified polypeptides (results not shown). Fig.2 depicts the result of a representative experiment with purified Top2(1–1196)-HMK-(His)6, which is comprised of the first 1196 amino acid residues of yeast DNA topoisomerase II and an HMK site and a hexahistidine tag at the COOH terminus. For the sample run in lane 1, the protein was first unfolded in 6 m guanidium chloride, and acetylation of the lysyl residues was carried out by treatment withN-hydroxysuccinimide. The acetylated protein was then radiolabeling by heart muscle kinase and subject to SDS-polyacrylamide gel electrophoresis after the same dialysis steps that were employed for samples subjected to lysine footprinting. A single band was observed at the position expected for the intact protein, indicating that the various steps in sample processing caused no appreciable breakage of the polypeptide chain. Lanes 2 and 3 of Fig. 2 contained the same polypeptide after partial cleavage at the cysteines by treatment with 2-nitro-5-cyanobenzoic acid (lane 2), or at methionines with cyanogen bromide (lane 3). Because these partial cleavage products were to be used as size markers in the protein footprinting experiments, they were subjected to the same unfolding and acetylation treatments to make their electrophoretic mobilities directly comparable to those of the two-step lysine footprinting products (15Hanai R. Wang J.C. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 11904-11908Crossref PubMed Scopus (41) Google Scholar). There are a total of 9 cysteines in yeast Top2(1–1196): Cys48, -149, -180, -355, -381, -443, -471, -710, and -731 (Fig. 1 e). Seven end-labeled products from partial cleavage at the cysteines were observed, including two doublets that were better resolved in autoradiograms that were not as overexposed as the one shown in Fig. 2. Six of these bands were unequivocally assigned to cleavages at Cys355, Cys381, and the cysteine pairs Cys443 and -471, and Cys710 and -731. A standard semilogarithmic plot of peptide sizes versusdistances migrated in the gel gave a linear line (not shown). The end-labeled product due to cleavage at Cys48 was probably unresolved from the uncleaved polypeptide in this gel, and the faint band below the uncleaved polypeptide in lane 2 was probably a mixture of products of cleavage at Cys149 and Cys180. The mobilities of the six assignable cysteine-cleavage products were in turn used to guide the assignment of the end-labeled bands derived from cleavage at the methionines (see "Materials and Methods" for details). For the bands shown inlane 3 of Fig. 2, about half of the 28 methionine-cleavage products (see Fig. 1 e) were well resolved, and these were readily assigned. To check for potential errors introduced by extrapolation beyond Cys731, the cysteine closest to the COOH terminus, the fastest migrating band in the quartet in lane 3 was isolated from a similar gel and analyzed by NH2-terminal peptide sequencing. The residues in the first five cycles of Edman reactions were found to be YGRIE, the expected sequence from cleavage after Met886, in agreement with the assignment based on its electrophoretic mobility. In three separate experiments, uncleaved but acetylated HA-HMK-Top2(1–659)-(His)6 was also included in the calibration measurements. Relative to the markers from the cysteine and methionine cleavage reactions, this fragment was found to correspond to cleavage at position 666 of HA-HMK-Top2(1–1428). This assignment is in good agreement with the calculated value 668 (659 from the NH2-terminal region of yeast DNA topoisomerase II plus six histidines and three additional amino acids between the yeast polypeptide and the histidine tag that were introduced during cloning). Samples of Top2(1–1196)-HMK-(His)6 footprinted with citraconic anhydride in the absence and presence of DNA were analyzed in lanes 4 and 6 of Fig. 2, respectively. As described previously, citraconylation of a particular lysine protects it from irreversible acetylation, and thus subsequent decitraconylation would make this site sensitive to cleavage by Lys-C endoprotease (15Hanai R. Wang J.C. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 11904-11908Crossref PubMed Scopus (41) Google Scholar). A comparison of the sample run in lane 1 with those run inlanes 4 and 6 shows clearly that citraconylation occurs at many sites along the polypeptide, leading to the formation of many end-labeled fragments following Lys-C treatment of the decitraconylated protein. The relative reactivities of the lysines toward citraconic anhydride in the absence and presence of DNA can be assessed from a comparison of the intensity patterns of these end-labeled cleavage products. Protection of a particular lysine against citraconylation by DNA binding would reduce the intensity of the end-labeled band generated by Lys-C cleavage at this lysine. By comparing the patterns of samples analyzed in lanes 4 and6, it is clear that the presence of DNA reduced the intensities of at least four bands marked a-d in theright margin of Fig. 2. Assignments of these bands were carried out by comparing their mobilities with those of the calibration markers. Data from a total of six sets of experiments with NH2-terminal and COOH-terminal-tagged proteins were analyzed; in each experiment, size calibration standards derived from the same tagged proteins were used in band assignments. In all measurements, the four bands shown in Fig.2 consistently showed DNA-dependent intensity changes. These bands were assigned to be the products of cleavages around residues 331, 512, 657, and 780. The probable errors in sizing the bands were estimated to be about ±5 amino acid residues from the six sets of measurements. In addition to the four bands indicated in Fig.2, in several other gels the presence of DNA also appear to reduce the intensity of the band corresponding to cleavage around residue 449 (data not shown). DNA-mediated protection of the same lysines against citraconylation was also observed when both DNA and AMP-PNP were added sequentially to the enzyme (compare the patterns shown in lanes 4 and5 for the enzyme itself and the enzyme-DNA complex after the addition of AMP-PNP, respectively). There appears to be, however, a reduction of the degree of DNA-mediated protection of cleavage at positions 332 and 512 upon AMP-PNP addition (compare the intensities of bands a and c in lane 5 with those inlane 6). There is strong evidence that the pair of ATPase domains in the dimeric enzyme come into contact upon binding of ATP or AMP-PNP to either the free enzyme or the enzyme-DNA complex (4Roca J. Wang J.C. Cell. 1992; 71: 833-840Abstract Full Text PDF PubMed Scopus (297) Google Scholar, 5Roca J. Wang J.C. Cell. 1994; 77: 609-616Abstract Full Text PDF PubMed Scopus (251) Google Scholar, 8Wigley D.B. Davies G.J. Dodson E.J. Maxwell A. Dodson G. Nature. 1991; 351: 624-629Crossref PubMed Scopus (487) Google Scholar,9Ali J.A. Orphanides G. Maxwell A. Biochemistry. 1995; 34: 9801-9808Crossref PubMed Scopus (90) Google Scholar). According to the protein-clamp model, the closing of the N-gate following ATP binding is responsible for triggering a conformational cascade in the enzyme, leading to the transport of one duplex DNA segment through another. Because of these ATP-mediated conformational and contact changes, both in and outside the vicinity of the ATP-binding pocket, the footprint of the enzyme is expected to be modulated by AMP-PNP. For the sample Top2(1–1196)-HMK-(His)6 used in the experiment shown in Fig. 2, the ATPase domain of yeast DNA topoisomerase II is located far away from the COOH-terminal HMK tag, and any change in the reactivity of a lysine in the vicinity of the ATPase domain would result in intensity changes of poorly resolved bands. Experiments were therefore carried out with two NH2-te
Referência(s)