Human Protein MCM6 on HeLa Cell Chromatin
1998; Elsevier BV; Volume: 273; Issue: 13 Linguagem: Inglês
10.1074/jbc.273.13.7320
ISSN1083-351X
AutoresHans Peter Holthoff, Martina Baack, Andreas S. Richter, Marion Ritzi, Rolf Knippers,
Tópico(s)RNA modifications and cancer
ResumoMinichromosome maintenance (Mcm) proteins perform essential functions regulating the replication of chromatin. Human cells, like other eukaryotic cells, express at least six Mcm proteins conserved in the central region. We have earlier described the primary structures of five human Mcm proteins, but the primary structure of the sixth human Mcm protein, MCM6, was identified only recently. We now use antibodies, specific for the MCM6 protein, to assess its intranuclear distribution. We find that a fraction of MCM6 protein occurs in the nucleosol, forming multiprotein complexes with other Mcm proteins. More importantly, we use for the first time micrococcal nuclease as a tool to investigate the association of MCM6 protein with chromatin. After short digestion times, a considerable fraction of the MCM6 protein is released from chromatin as a multiprotein complex that includes other Mcm proteins as well. In addition, fractions of MCM3 and MCM6 proteins are released by nuclease digestion as monomeric proteins indicating that at least these two Mcm proteins may also occur as single molecules on chromatin. The data also suggest that the chromatin regions with bound Mcm proteins are more vulnerable to nuclease attack than bulk chromatin and may therefore differ in the arrangement of nucleosomes. Minichromosome maintenance (Mcm) proteins perform essential functions regulating the replication of chromatin. Human cells, like other eukaryotic cells, express at least six Mcm proteins conserved in the central region. We have earlier described the primary structures of five human Mcm proteins, but the primary structure of the sixth human Mcm protein, MCM6, was identified only recently. We now use antibodies, specific for the MCM6 protein, to assess its intranuclear distribution. We find that a fraction of MCM6 protein occurs in the nucleosol, forming multiprotein complexes with other Mcm proteins. More importantly, we use for the first time micrococcal nuclease as a tool to investigate the association of MCM6 protein with chromatin. After short digestion times, a considerable fraction of the MCM6 protein is released from chromatin as a multiprotein complex that includes other Mcm proteins as well. In addition, fractions of MCM3 and MCM6 proteins are released by nuclease digestion as monomeric proteins indicating that at least these two Mcm proteins may also occur as single molecules on chromatin. The data also suggest that the chromatin regions with bound Mcm proteins are more vulnerable to nuclease attack than bulk chromatin and may therefore differ in the arrangement of nucleosomes. Mcm proteins have recently attracted considerable attention because they are thought to perform an essential function in the regulation of chromatin replication (reviewed in Refs. 1Chong J.P.J. Thömmes P. Blow J.J. Trends Biochem. Sci. 1996; 21: 102-106Abstract Full Text PDF PubMed Scopus (157) Google Scholar and 2Kearsey S.E. Maiorano D. Holmes E.C. Todorov I.T. BioEssays. 1996; 18: 183-190Crossref PubMed Scopus (122) Google Scholar). Eukaryotic cells express six evolutionarily conserved Mcm proteins. They were so named because the first members of this family of nuclear proteins were discovered through genetic screens of yeast mutants unable to maintain the extrachromosomal replication of minichromosomes (Mcm, minichromosome maintenance) 1The abbreviations used are: Mcm, minichromosome maintenance; bp, base pair(s). 1The abbreviations used are: Mcm, minichromosome maintenance; bp, base pair(s). (3Tye B.K. Trends Cell Biol. 1994; 4: 160-166Abstract Full Text PDF PubMed Scopus (108) Google Scholar). Mcm proteins are abundant nuclear proteins with copy numbers ranging 104 to 106 per cell nucleus. They occur either free in the nucleosol or are bound to nuclear chromatin. The fraction of chromatin-bound Mcm proteins is highest at the end of the G1 phase or the beginning of the S phase of the cell cycle but gradually decreases as S phase proceeds (4Kimura H. Nozaki N. Sugimoto K. EMBO J. 1994; 13: 4311-4320Crossref PubMed Scopus (147) Google Scholar, 5Todorov I.T. Attaran A. Kearsey S.E. J. Cell Biol. 1995; 129: 1433-1445Crossref PubMed Scopus (204) Google Scholar, 6Krude T. Musahl C. Laskey R.A. Knippers R. J. Cell Sci. 1996; 109: 309-318Crossref PubMed Google Scholar). In postreplicative cells, essentially all Mcm proteins are free in the nucleosol and disperse throughout the cell upon nuclear breakdown in mitosis, but they rapidly reassemble on chromatin sites in nuclei after mitotic telophase. Binding of Mcm proteins appears to make chromatin competent for replication. This can be concluded from biochemical work withXenopus egg extracts (7Chong J.P.J. Mahbubani H.M. Khoo V.Y. Blow J.J. Nature. 1995; 375: 418-421Crossref PubMed Scopus (316) Google Scholar, 8Madine M.A. Khoo C.Y. Mills A.D. Laskey R.A. Nature. 1995; 375: 421-424Crossref PubMed Scopus (232) Google Scholar, 9Kubota Y. Mimura S. Nishimoto S. Takisawa H. Nojima H. Cell. 1995; 81: 601-609Abstract Full Text PDF PubMed Scopus (243) Google Scholar, 10Kubota Y. Mimura S. Nishimoto S. Masuda T. Nojima H. Zakisawa H. EMBO J. 1997; 16: 3320-3331Crossref PubMed Scopus (94) Google Scholar) as well as from genetic experiments with yeast mutants. Data strongly suggest that an association of Mcm proteins with chromatin depends on the previous binding of the origin recognition complex (ORC) and requires the function of additional proteins such as the Cdc6 protein (11Cocker J.H. Piatti S. Santocanale C. Nasmyth K. Diffley J.F.X. Nature. 1996; 379: 180-182Crossref PubMed Scopus (296) Google Scholar, 12Donovan S. Harwood J. Drury L.S. Diffley J.F.X. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 5611-5616Crossref PubMed Scopus (432) Google Scholar, 13Romanowski P. Madine M.A. Rowles A. Blow J.J. Laskey R.A. Curr. Biol. 1996; 6: 1416-1425Abstract Full Text Full Text PDF PubMed Scopus (172) Google Scholar) and other proteins as well (14Zou L. Mitchell J. Stillman B. Mol. Cell. Biol. 1997; 17: 553-563Crossref PubMed Scopus (137) Google Scholar). Since only Mcm-containing chromatin, but not Mcm-free chromatin, is able to replicate, the replication-dependent release of bound chromatin appears to be a powerful mechanism that prevents the re-replication of chromatin during S phase. Thömmes et al. (15Thömmes P. Fett R. Schray B. Burkhart R. Barnes M. Kennedy C. Brown N.C. Knippers R. Nucleic Acids Res. 1992; 20: 1069-1074Crossref PubMed Scopus (109) Google Scholar) detected the first mammalian Mcm protein by serendipity, whereas Hu et al. (16Hu B. Burkhart R. Schulte D. Musahl C. Knippers R. Nucleic Acids Res. 1993; 21: 5289-5293Crossref PubMed Scopus (73) Google Scholar) used an antibody directed against a central protein region, common to all Mcm proteins, to isolate additional Mcm sequences from a human HeLa cell cDNA library. Holthoff et al. (17Holthoff H.P. Hameister H. Knippers R. Genomics. 1996; 37: 131-134Crossref PubMed Scopus (15) Google Scholar) isolated and sequenced the last of the six human Mcm cDNAs, originally described as p105Mcm (18Musahl C. Schulte D. Burkhart R. Knippers R. Eur. J. Biochem. 1995; 230: 1096-1101Crossref PubMed Scopus (80) Google Scholar), but now known as MCM6. Human MCM6 protein (MCM6p) consists of 821 amino acids with an electrophoretic molecular mass of 105 kDa and is very similar in sequence to protein Mis5, first discovered as a replication protein in Schizosaccharomyces pombe (19Takahashi K. Yamada H. Yanagida M. Mol. Biol. Cell. 1994; 5: 1145-1158Crossref PubMed Scopus (186) Google Scholar). A cDNA, encoding human MCM6p, was also isolated and described by Tsuruga et al. (20Tsuruga H. Yabuta N. Hosoya S. Tamura K. Endo Y. Nojima H. Genes Cells. 1997; 2: 381-399Crossref PubMed Scopus (29) Google Scholar). Here, we describe the preparation of MCM6-specific antibodies which we have used to assess the distribution of MCM6p in the nucleus and the interaction of MCM6p with chromatin from HeLa cells. For this purpose, we degraded chromatin with micrococcal nuclease and separated the resulting chromatin fragments by glycerol gradient centrifugation. We partially recovered MCM6p in association with chromatin fragments while another part of MCM6p appeared as free protein. We conclude that MCM6p resides in chromatin regions that differ in structure or organization from bulk chromatin. HeLa cells were cultivated on plastic dishes unter previously described conditions (21Schulte D. Burkhart R. Musahl C. Hu B. Schlatterer C. Hameister H. Knippers R. J. Cell Sci. 1995; 108: 1381-1389Crossref PubMed Google Scholar). For synchronization, cells were arrested in early S phase by two subsequent thymidine blocks (22Johnson R.T. Downes C.S. Meyn R.E. Fantes P. Brooks R. The Cell Cycle. A Practical Approach. IRL Press, Oxford1993: 1-24Google Scholar) and released into the cell cycle by removing excess thymidine. Progression through S phase was monitored by pulse-labeling with [3H]-thymidine (1 μCi/60 min) (21Schulte D. Burkhart R. Musahl C. Hu B. Schlatterer C. Hameister H. Knippers R. J. Cell Sci. 1995; 108: 1381-1389Crossref PubMed Google Scholar). Incorporated radioactivity was determined in acid precipitates by scintillation counting. The number of mitotic cells was determined according to Ohyashiki et al. (23Ohyashiki J.H. Ohyashiki K. Toyama K. Cancer Genet. Cytogenet. 1988; 34: 159-163Abstract Full Text PDF PubMed Scopus (11) Google Scholar). Nuclei were prepared either in buffers with 40 μg/ml digitonin essentially as described by Adam et al. (24Adam S.A. Sterne-Marr R. Gerace L. Methods Enzymol. 1992; 219: 97-110Crossref PubMed Scopus (125) Google Scholar), or after swelling of washed HeLa cells in hypotonic buffer A (10 mmHepes, pH 7.5; 0.5 mm EDTA; 0.1% Nonidet P-40) by Dounce homogenization as described previously in detail (25Richter A. Knippers R. Eur. J. Biochem. 1997; 247: 136-141Crossref PubMed Scopus (24) Google Scholar). Protein extracts were prepared from digitonin nuclei by 0.5% Nonidet P-40 in 150 mm sodium acetate, 10 mm Tris-HCl (pH 7.5), and 1 mm EDTA. Unsoluble material was removed by centrifugation at 100,000 × g. The supernatant was analyzed by immunoblotting (see below) or through glycerol gradient centrifugation (40 to 5% glycerol in the sodium acetate buffer described above). Chromatin was prepared as originally described by Hancock (26Hancock R. J. Mol. Biol. 1974; 86: 649-663Crossref PubMed Scopus (87) Google Scholar). Briefly, cells on plates were washed three times with 50 ml of hypotonic buffer B (1 mm Hepes, pH 7.5, 0.25 mm EDTA). The cells, still attached to the plates, were incubated with 0.25 mm EDTA. Nuclei, released from the swollen cells were washed and resuspended in buffer B plus 0.5% Nonidet P-40. Nuclei were then centrifuged for 10 min at 4,000 ×g through a 20-ml sucrose cushion (0.1 m sucrose in 1 mm Tris-HCl, pH 8.0). The pellet was resuspended in hypotonic buffer B at a final concentration corresponding to 700 μg of DNA/ml. DNA concentrations were determined with Hoechst 33258 by fluorimetry (Hoefer Scientific Instruments, San Francisco). Chromatin was treated with micrococcal nuclease (Boehringer Mannheim) in 10 mmTris-HCl, pH 7.5, 2 mm CaCl2 under the conditions described below (see Fig. 5). The reaction was stopped with 8 mm EDTA. Insoluble material was removed by centrifugation. Fragmented chromatin in the supernatants were analyzed by centrifugation (15 h at 40,000 rotations/min in a Beckman SW40 rotor) through glycerol gradients (40 to 5%; 5 mmTris-HCl, pH 7.5, 5 mm EDTA). The gradients were pumped after centrifugation through a quartz cuvette to monitor the absorbance of ultraviolet light at 254 nm. DNA fragments were analyzed after deproteinization of chromatin fragments by agarose gel electrophoresis (27Krude T. Knippers R. Mol. Cell. Biol. 1991; 11: 6257-6267Crossref PubMed Scopus (44) Google Scholar). To prepare MCM6-specific antibodies, a synthetic peptide, corresponding to amino acid positions 9–26 of the MCM6 coding sequence (17Holthoff H.P. Hameister H. Knippers R. Genomics. 1996; 37: 131-134Crossref PubMed Scopus (15) Google Scholar), was linked to hemocyanin as carrier and used for the immunization of rabbits. Antibodies were prepared according to Harlow and Lane (28Harlow E. Lane D. Antibodies: A Laboratory Handbook. Cold Spring Harbor Laboratory Press, New York1988Google Scholar). Affinity-purified MCM3-specific antibodies (21Schulte D. Burkhart R. Musahl C. Hu B. Schlatterer C. Hameister H. Knippers R. J. Cell Sci. 1995; 108: 1381-1389Crossref PubMed Google Scholar) and antibodies directed against a region in the conserved Mcm domain (DEFD-antibodies) have been described (16Hu B. Burkhart R. Schulte D. Musahl C. Knippers R. Nucleic Acids Res. 1993; 21: 5289-5293Crossref PubMed Scopus (73) Google Scholar, 18Musahl C. Schulte D. Burkhart R. Knippers R. Eur. J. Biochem. 1995; 230: 1096-1101Crossref PubMed Scopus (80) Google Scholar). Antibodies were used for immunoblotting (Western) experiments essentially as originally described by Towbinet al. (29Towbin H. Staehelin T. Gordon J. Proc. Natl. Acad. Sci U. S. A. 1979; 76: 4350-4354Crossref PubMed Scopus (44911) Google Scholar) with minor modifications (30Burkhart R. Schulte D. Hu B. Musahl C. Göhring F. Knippers R. Eur. J. Biochem. 1995; 228: 431-438Crossref PubMed Scopus (119) Google Scholar). Membranes, carrying the transferred proteins, were washed extensively in a Tris-NaCl buffer containing 0.5% Tween 20 and developed using the Enhanced Chemiluminescence System (ECL) according to the manufacturer instructions (Amersham Pharmacia Biotech). A synthetic peptide, encoding amino acid residues 9 to 26 of the MCM6p sequence (Fig. 1 A), was coupled to hemocyanin as a carrier and used to raise antibodies in rabbits. Since the antigenic peptide corresponds to a section of MCM6p that is not conserved among Mcm proteins, we expected that the antibodies obtained are specifically directed against MCM6p while ignoring the other Mcm proteins. To test this prediction, we subjected a crude protein extract from HeLa cell nuclei to denaturing polyacrylamide gel electrophoresis and transferred the polypeptide bands to a nylon membrane for immunostaining with the affinity-purified MCM6-specific antibodies (Western blotting). As shown in Fig. 1 B, the antibodies recognized only one polypeptide with an apparent molecular mass of 105 kDa as expected from previous experiments (18Musahl C. Schulte D. Burkhart R. Knippers R. Eur. J. Biochem. 1995; 230: 1096-1101Crossref PubMed Scopus (80) Google Scholar). Preincubation of the antibodies with the peptide used as antigen effectively suppressed the immune reaction (not shown), and preimmune antibodies gave negative results (Fig. 1 B). For comparisons, we used affinity purified antibodies specifically directed against protein MCM3 (MCM3p), which has electrophoretic properties similar to MCM6p (18Musahl C. Schulte D. Burkhart R. Knippers R. Eur. J. Biochem. 1995; 230: 1096-1101Crossref PubMed Scopus (80) Google Scholar). The MCM3p-specific antibodies reacted with a polypeptide band (Fig. 1 B) that migrated slightly slower under the electrophoretic conditions used and could therefore be clearly distinguished from MCM6p. Finally, polyclonal rabbit antibodies, directed against a peptide from the central protein region containing the highly conserved Mcm “signature sequence” (including the amino acid sequence motif DEFD; see Fig. 1 A), recognized in the crude nuclear extracts the six known human Mcm proteins including MCM6p (16Hu B. Burkhart R. Schulte D. Musahl C. Knippers R. Nucleic Acids Res. 1993; 21: 5289-5293Crossref PubMed Scopus (73) Google Scholar). Thus, the results of Fig. 1 show that the newly prepared antibodies are specific for MCM6p and can therefore be used as a tool to investigate the intracellular distribution of MCM6p by immunoblotting. To determine the intranuclear distribution of MCM6p, we used two different procedures to prepare nuclei from proliferating HeLa cells. In one procedure, we treated cells with digitonin at concentrations which effectively disrupt the cytoplasmic membrane but not the nuclear envelope (24Adam S.A. Sterne-Marr R. Gerace L. Methods Enzymol. 1992; 219: 97-110Crossref PubMed Scopus (125) Google Scholar). Therefore all nuclear proteins, soluble and structure-bound, can be recovered in the nuclear pellet. In the second procedure, nuclei were prepared in the presence of EDTA and the nonionic detergent Nonidet P-40 to partially degrade the nuclear envelope (25Richter A. Knippers R. Eur. J. Biochem. 1997; 247: 136-141Crossref PubMed Scopus (24) Google Scholar). Consequently, free nucleosolic Mcm proteins diffuse into the surrounding medium while chromatin-bound Mcm proteins remain in the nuclear pellet. We determined the presence of MCM6p in the nuclear pellets and in the postnuclear supernatants by Western blotting using MCM6-specific antibodies as well as the DEFD-antibodies that recognize all Mcm proteins. As shown in Fig. 2 (lane 3), most MCM6p, together with other Mcm proteins, remained in nuclei prepared by the digitonin method. The minor fraction of MCM6p that appeared in the cytosolic supernatant of these preparations (Fig. 2, lane 2) originated most probably from mitotic cells where Mcm proteins disperse throughout the cell after nuclear breakdown (25Richter A. Knippers R. Eur. J. Biochem. 1997; 247: 136-141Crossref PubMed Scopus (24) Google Scholar). In contrast, when nuclei were prepared with EDTA and Nonidet P-40, approximately one-half of the Mcm proteins, including MCM6p, appeared in the postnuclear supernatant (Fig. 2, lane 4) while the second half remained associated with the nuclear pellet (Fig. 2,lane 5). This result is in agreement with earlier observations showing that only a fraction of Mcm proteins is associated with chromatin in asynchronously proliferating HeLa cells (18Musahl C. Schulte D. Burkhart R. Knippers R. Eur. J. Biochem. 1995; 230: 1096-1101Crossref PubMed Scopus (80) Google Scholar, 30Burkhart R. Schulte D. Hu B. Musahl C. Göhring F. Knippers R. Eur. J. Biochem. 1995; 228: 431-438Crossref PubMed Scopus (119) Google Scholar). However, in the present experiments, MCM2 protein (MCM2p) was not recovered in the nuclear pellet but appeared entirely in the postnuclear supernatant (Fig. 2, compare lanes 4 and 5). This was unexpected since an association of MCM2p with mammalian cell chromatin has been unambiguously shown by Todorov et al. (5Todorov I.T. Attaran A. Kearsey S.E. J. Cell Biol. 1995; 129: 1433-1445Crossref PubMed Scopus (204) Google Scholar). We could show though that the behavior of MCM2p in the present experiments is most likely due to the low ionic strength conditions used because 40 mm NaCl rather than 10 mm NaCl in hypotonic buffer A (see “Experimental Procedures”) prevented the loss of MCM2p (not shown). It has been noted before that human MCM2p only weakly interacts with other Mcm proteins (10Kubota Y. Mimura S. Nishimoto S. Masuda T. Nojima H. Zakisawa H. EMBO J. 1997; 16: 3320-3331Crossref PubMed Scopus (94) Google Scholar, 18Musahl C. Schulte D. Burkhart R. Knippers R. Eur. J. Biochem. 1995; 230: 1096-1101Crossref PubMed Scopus (80) Google Scholar) and may have different chromatin binding properties (20Tsuruga H. Yabuta N. Hosoya S. Tamura K. Endo Y. Nojima H. Genes Cells. 1997; 2: 381-399Crossref PubMed Scopus (29) Google Scholar), but the conditions under which the binding of MCM2p to chromatin becomes sensitive to low salt remains to be investigated in detail. Following the observations of Fujita et al. (32Fujita M. Kiyono T. Hayashi Y. Ishibashi M. J. Biol. Chem. 1997; 272: 10928-10935Abstract Full Text Full Text PDF PubMed Scopus (76) Google Scholar), we could already demonstrate that ATP (1–5 mm) stabilizes the interaction of MCM2p with chromatin in buffer A. According to previous experiments, nucleosolic MCM6p can be expected to participate in the formation of multiprotein complexes with other Mcm proteins (18Musahl C. Schulte D. Burkhart R. Knippers R. Eur. J. Biochem. 1995; 230: 1096-1101Crossref PubMed Scopus (80) Google Scholar). However, these earlier conclusions were based on indirect evidence because specific tools for the detection of MCM6p had not been available. To directly demonstrate the presence of nucleosolic MCM6p in high molecular weight complexes with other Mcm proteins, we centrifuged the protein supernatants, obtained after disintegration of the nuclear envelope (see “Experimental Procedures”), through glycerol gradients. MCM6p could be recovered in two peaks, one of which sedimented faster while the other sedimented slower than the catalase sedimentation marker (Fig. 3, lower panel). Using the DEFD-antibody for Western blotting, we could show that the faster moving MCM6p cosedimented with the five other Mcm proteins, while the slower moving MCM6p fraction cosedimented with MCM4p and MCM7p (Fig. 3, upper panel). We interprete the data of Fig. 3 in light of previous reports according to which Mcm proteins are able to interact forming large multiprotein complexes as well as more stable subcomplexes. One of these subcomplexes had been shown to contain MCM4p and MCM7p as well as a third 105-kDa Mcm-related protein (10Kubota Y. Mimura S. Nishimoto S. Masuda T. Nojima H. Zakisawa H. EMBO J. 1997; 16: 3320-3331Crossref PubMed Scopus (94) Google Scholar, 18Musahl C. Schulte D. Burkhart R. Knippers R. Eur. J. Biochem. 1995; 230: 1096-1101Crossref PubMed Scopus (80) Google Scholar, 25Richter A. Knippers R. Eur. J. Biochem. 1997; 247: 136-141Crossref PubMed Scopus (24) Google Scholar, 33Thömmes P. Kubota Y. Takisawa H. Blow J.J. EMBO J. 1997; 16: 3312-3319Crossref PubMed Scopus (117) Google Scholar) that we are now able to identify as MCM6p. In initial experiments, we treated nuclei, prepared in the presence of EDTA and Nonidet P-40, with pancreatic deoxyribonuclease or with micrococcal nuclease to investigate the chromatin-bound fraction of MCM6p. Using incubation conditions optimal for enzymatic activity, we observed a degradation of chromatin but no release of Mcm proteins from the nuclei. Kimura et al. (4Kimura H. Nozaki N. Sugimoto K. EMBO J. 1994; 13: 4311-4320Crossref PubMed Scopus (147) Google Scholar) have reported similar observations before. Following these authors, we assumed that Mcm proteins may aggregate or precipitate within the nuclei, and we therefore decided to use isolated chromatin as substrate for nuclease digestion. However, chromatin preparations in buffers, containing 10–20 mm Tris or Hepes plus 40–120 mm NaCl or KCl, tend to loose bound Mcm proteins when incubated at temperatures required for nuclease activity, as has also been recently observed by Fujita et al. (32Fujita M. Kiyono T. Hayashi Y. Ishibashi M. J. Biol. Chem. 1997; 272: 10928-10935Abstract Full Text Full Text PDF PubMed Scopus (76) Google Scholar). This prompted us to use the chromatin preparation procedure of Hancock (27Krude T. Knippers R. Mol. Cell. Biol. 1991; 11: 6257-6267Crossref PubMed Scopus (44) Google Scholar) who had shown that a treatment of carefully washed cells with Nonidet P-40 in very low ionic strength buffers removes the nuclear envelope and yields spherical “chromatin bodies” that can be well handled by routine laboratory procedures such as pipetting and centrifugation. As already demonstrated by Hancock (27Krude T. Knippers R. Mol. Cell. Biol. 1991; 11: 6257-6267Crossref PubMed Scopus (44) Google Scholar), the low ionic strength conditions do not lead to detectable rearrangements in chromatin structure and composition when compared with isolated nuclei. To determine whether MCM6p is bound to these chromatin preparations in a physiologically meaningful way, we isolated chromatin bodies from HeLa cells arrested at the G1/S phase boundary by a double-thymidine block as well as from cells at different times after release from the thymidine block. We expected that MCM6p behaves much like other Mcm proteins (4Kimura H. Nozaki N. Sugimoto K. EMBO J. 1994; 13: 4311-4320Crossref PubMed Scopus (147) Google Scholar, 5Todorov I.T. Attaran A. Kearsey S.E. J. Cell Biol. 1995; 129: 1433-1445Crossref PubMed Scopus (204) Google Scholar, 6Krude T. Musahl C. Laskey R.A. Knippers R. J. Cell Sci. 1996; 109: 309-318Crossref PubMed Google Scholar) and dissociates from chromatin as a consequence of chromatin replication. We found indeed that relatively high amounts of MCM6p were present on chromatin from cells in early S phase and that MCM6p dissociated from chromatin later in S phase but rapidly bound again after completion of mitosis (Fig. 4). When the experiment of Fig. 4 was carried out in the presence of aphidicolin, a potent inhibitor of replicative DNA polymerases, MCM6p remained on chromatin prepared at different cell cycle phases (not shown; see Ref. 13Romanowski P. Madine M.A. Rowles A. Blow J.J. Laskey R.A. Curr. Biol. 1996; 6: 1416-1425Abstract Full Text Full Text PDF PubMed Scopus (172) Google Scholar). Therefore, the dissociation of MCM6p from chromatin required active DNA replication. We conclude that the presence of MCM6p on isolated chromatin faithfully reflects the in vivo cycle of replication-dependent dissociation of Mcm proteins from chromatin and their post-mitotic reassociation with chromatin. Next, we treated early S phase chromatin with micrococcal nuclease and found that the amounts of released MCM6p as well as of other Mcm proteins increased in parallel with the fraction of digested DNA in a manner depending on enzyme concentration (Fig. 5 A) and on time of incubation (Fig. 5 B). Micrococcal nuclease is a useful and important tool in chromatin research (reviewed in Ref. 34Van Holde K.E. Chromatin. Springer Verlag, New York1988Google Scholar). The enzyme preferentially attacks the linker DNA between nucleosomes and therefore produces chromatin fragments carrying one, two, or more nucleosomes depending on the extent of nucleolytic degradation. In the experiment leading to Fig. 6, we have treated chromatin bodies for 5 min with micrococcal nuclease. Insoluble material was removed by low speed centrifugation while the supernatant was centrifuged through a glycerol gradient. The position of chromatin fragments in the gradient was determined by measuring the optical density at 254 nm (Fig. 6 A) and verified by an analysis of DNA fragments after deproteinization (Fig. 6 B). Western blotting experiments with the DEFD-antibodies revealed that Mcm proteins were distributed over many fractions of the glycerol gradient (Fig. 6 C), but a maximum, including all Mcm proteins (except Mcm2p; see above), appeared in gradient fractions 15–16 together with DNA fragments of about 900-bp lengths (Fig. 6 C). A second peak, also comprising all Mcm proteins, sedimented at a sedimentation rate of 11 S, together with mononucleosomes (Fig. 6 C,fractions 8–9). In addition, a considerable amount of Mcm proteins in the 105-kDa size range remained close to the top of the gradient (Fig. 6 C, fractions 3–5). This group of apparently uncomplexed proteins included MCM3p and MCM6p as shown by specific antibodies in Western blot experiments (Fig. 6, D and E). Therefore, MCM3p and MCM6p can be recovered from nuclease-treated chromatin as components of two different fast sedimenting larger Mcm complexes but also as apparently free molecules close to the top of a glycerol gradient. Are Mcm proteins in the larger complexes bound to chromatin fragments? This point was investigated by treating chromatin more extensively with micrococcal nuclease. As expected, more extensive nuclease treatment mainly produced mono- and dinucleosomal chromatin fragments (Fig. 7, A and B). The distribution of Mcm proteins in this gradient was significantly different from that obtained after short nuclease digestion times since the larger Mcm-carrying structures were converted to a major 11 S peak (Fig. 7 C, fractions 9–11) with a shoulder extending into faster moving regions. This leading shoulder of the Mcm peak disppeared upon further incubation with micrococcal nuclease (not shown). The 11 S peak of Mcm proteins only partially overlapped with the peak of mononucleosomes. In fact, an analysis of the 11 S peak material by non-denaturing polyacrylamide gel electrophoresis revealed that Mcm proteins migrated much slower than mononucleosomes (35Schlaeger E.J. Biochemistry. 1982; 21: 3167-3174Crossref PubMed Scopus (16) Google Scholar) (not shown), strongly suggesting that Mcm proteins in the 11 S peak were not bound to nucleosomes. Furthermore, the peak of free MCM3p close to the top of the glycerol gradient (Fig. 7 D, fraction 5) appeared to be more prominent and symmetric after longer nuclease digestion times, suggesting that residual pieces of DNA which remained bound to MCM3p after short digestion times (Fig. 6 D) were degraded after longer treatment with micrococcal nuclease. The distribution of MCM6p in the glycerol gradient was significantly different from that of MCM3p. Most MCM6p sedimented faster than free MCM3p (Fig. 7 E), and a considerable amount of MCM6p appeared in leading gradient fractions suggesting that MCM6p, released from chromatin, tends to form fast sedimenting aggregates. It has yet to be determined whether these aggregates are composed only of MCM6p or whether they also contain chromatin proteins other than MCM6p. Using MCM6-specific antibodies as a tool, we could unambiguously detect MCM6p in protein extracts from HeLa cell nuclei. This has not been possible in earlier experiments because MCM6p could not be well separated from MCM3p by standard polyacrylamide gel electrophoresis. We can show now that MCM6p, like other Mcm proteins, partitions between two nuclear compartments and occurs either as free protein in the nucleosol or bound to chromatin. Nucleosolic MCM6p participates in complex formation. The largest of these complexes has a sedimentation coefficient of 14–15 S in glycerol gradients and most likely includes all six Mcm proteins (18Musahl C. Schulte D. Burkhart R. Knippers R. Eur. J. Biochem. 1995; 230: 1096-1101Crossref PubMed Scopus (80) Google Scholar) although its precise stoichiometry has yet to be determined and may vary with the cell cycle (see Fig. 4 in Ref. 25Richter A. Knippers R. Eur. J. Biochem. 1997; 247: 136-141Crossref PubMed Scopus (24) Google Scholar). A smaller complex, including MCM6p together with MCM4p and MCM7p, was also detected in the glycerol gradient of nucleosolic proteins (Fig. 3). It is not known whether this subcomplex exists in vivo because high molecular weight nucleosolic Mcm protein complexes tend to disintegrate under certain extraction conditions. For example, it has been noticed that high salt conditions induce a dissociation of the large nucleosolic Mcm protein complex into more stable subcomplexes including a trimeric complex of MCM6p with MCM4p and MCM7p (18Musahl C. Schulte D. Burkhart R. Knippers R. Eur. J. Biochem. 1995; 230: 1096-1101Crossref PubMed Scopus (80) Google Scholar), and sedimentation through sucrose gradients, rather than glycerol gradients, also causes a disintegration of high molecular weight complexes. 2H. P. Holthoff, unpublished data. We add that the Mcm proteins in Xenopus oocytes as well as in the nucleosol of a variety of cells from different eukaryotes also form large complexes that tend to dissociate into subcomplexes, of which one contains MCM4p, MCM6p, and MCM7p (33 and references therein). A novel and probably more important aspect of the work reported above concerns Mcm proteins on isolated chromatin. For reasons explained above, we have used the classic Hancock procedure for the isolation of chromatin from HeLa cells. This procedure requires the use of the detergent Nonidet P-40 in a buffer with very low ionic strength. We show that most Mcm proteins remain associated with isolated chromatin. An exception is MCM2p which is specifically lost during chromatin isolation although previous experiments in vivo had clearly shown that MCM2p can bind to chromatin just like other Mcm proteins (5Todorov I.T. Attaran A. Kearsey S.E. J. Cell Biol. 1995; 129: 1433-1445Crossref PubMed Scopus (204) Google Scholar). One reason for the loss of MCM2p could be that MCM2p requires ATP for tight chromatin binding (32Fujita M. Kiyono T. Hayashi Y. Ishibashi M. J. Biol. Chem. 1997; 272: 10928-10935Abstract Full Text Full Text PDF PubMed Scopus (76) Google Scholar). We could indeed show that an addition of ATP to the extraction buffer prevents the dissociation of MCM2p from Hancock chromatin. 3A. Richter, unpublished observations. In any case, MCM6p as well as other Mcm proteins remain rather firmly bound to isolated chromatin which clearly excludes the possibility that Mcm proteins are tethered to chromatin by MCM2p (36Ishimi Y. Ichinose S. Omori A. Sato K. Kimura H. J. Biol. Chem. 1996; 271: 24115-24122Abstract Full Text Full Text PDF PubMed Scopus (102) Google Scholar). MCM6p as well as other Mcm proteins appear to be bound to isolated chromatin in a physiologically meaningful manner because their amount was found to be highest in chromatin, prepared from cells arrested at the G1/S phase boundary, but the amount decreases in chromatin from cells at later stages of the cell cycle and rapidly reassociates with chromatin at the end of mitosis. This cycle of dissociation and reassociation mirrors the behavior of Mcm proteins in HeLa cells as investigated by in vivo immunofluorescence studies (6Krude T. Musahl C. Laskey R.A. Knippers R. J. Cell Sci. 1996; 109: 309-318Crossref PubMed Google Scholar). To learn more about the chromatin binding of MCM6p, we treated isolated Hancock chromatin with micrococcal nuclease and could show that MCM6p is released during chromatin digestion at a rate that was similar to the rate by which other Mcm proteins were mobilized. This could probably indicate that all detectable Mcm proteins are located in chromatin regions of similar general organization. Chromatin fragments, obtained by micrococcal nuclease digestion, were analyzed by glycerol gradient centrifugation. A major observation was that MCM6p, together with other Mcm proteins, were not evenly distributed among chromatin fragments. The data show instead that, after short exposure to nuclease, a considerable amount of Mcm proteins were enriched in glycerol gradient fractions with DNA fragments of about 900-bp lengths as well as in slower moving fractions with a sedimentation coefficient of about 11 S. Mcm proteins in the faster moving complexes were bound to DNA because these complexes disappeared upon longer nuclease digestion times and were at least partially converted into 11 S forms. The 11 S forms of Mcm complexes cosedimented, but were not associated, with mononucleosomes and seem to contain little, if any, DNA because their sedimentation properties did not change even after extensive treatment with nucleases. The presence of five different Mcm proteins in the 11 S forms of chromatin-digestion products is consistent with the possibility that Mcm proteins on chromatin interact and constitute multiprotein complexes (10Kubota Y. Mimura S. Nishimoto S. Masuda T. Nojima H. Zakisawa H. EMBO J. 1997; 16: 3320-3331Crossref PubMed Scopus (94) Google Scholar, 32Fujita M. Kiyono T. Hayashi Y. Ishibashi M. J. Biol. Chem. 1997; 272: 10928-10935Abstract Full Text Full Text PDF PubMed Scopus (76) Google Scholar, 33Thömmes P. Kubota Y. Takisawa H. Blow J.J. EMBO J. 1997; 16: 3312-3319Crossref PubMed Scopus (117) Google Scholar). However, only one part of the recovered MCM3p occurs in complexes with other Mcm proteins, whereas another part of MCM3p appears in a free uncomplexed form close to the top of glycerol gradients, even after short digestion times (Fig. 7). It will be interesting to investigate whether free MCM3p has a function distinct from the multiprotein Mcm complex and whether MCM3p is transiently released from the multiprotein complex during chromatin replication. Similar to MCM3p, a fraction of MCM6p, mobilized by nuclease treatment of chromatin, sedimented independently of other Mcm proteins (Fig. 7). In fact, much of the recovered MCM6p was widely distributed over fractions in the leading part of glycerol gradients and seems to occur in aggregates, possibly as a consequence of the release of MCM6p from chromatin sites. Aggregation is specific for chromatin-derived MCM6p since a nucleosolic MCM6p was never found to aggregate (see Fig. 3). An interesting possibility is that the tendency to aggregate is determined by the state of Mcm protein phosphorylation because chromatin-bound Mcm proteins are generally less highly phosphorylated than nucleosolic Mcms (4Kimura H. Nozaki N. Sugimoto K. EMBO J. 1994; 13: 4311-4320Crossref PubMed Scopus (147) Google Scholar, 5Todorov I.T. Attaran A. Kearsey S.E. J. Cell Biol. 1995; 129: 1433-1445Crossref PubMed Scopus (204) Google Scholar, 18Musahl C. Schulte D. Burkhart R. Knippers R. Eur. J. Biochem. 1995; 230: 1096-1101Crossref PubMed Scopus (80) Google Scholar, 21Schulte D. Burkhart R. Musahl C. Hu B. Schlatterer C. Hameister H. Knippers R. J. Cell Sci. 1995; 108: 1381-1389Crossref PubMed Google Scholar, 37Sato N. Arai K. Masai H. EMBO J. 1997; 16: 4340-4351Crossref PubMed Scopus (121) Google Scholar). In sum, our chromatin digestion experiments suggest that the five recovered Mcm proteins interact with chromatin in the form of 11 S complexes. In addition, two proteins, MCM3p and MCM6p, may independently occur on chromatin as monomeric units not tightly linked to other Mcm proteins. We could also show that Mcm proteins reside on chromatin sites that appear to be more vulnerable to nuclease attack than bulk chromatin. Further experiments are necessary to define the Mcm-carrying chromatin sites with respect to the underlying DNA structure and sequence and to additional proteins in the vicinity of bound Mcms. If Mcm proteins change their location from origins to replication forks, as has recently been suggested (38Aparicio O.M. Weinstein D.M. Bell S.P. Cell. 1997; 91: 59-69Abstract Full Text Full Text PDF PubMed Scopus (640) Google Scholar), we may expect to find different chromatin environments and different protein neighbors before and after initiation of chromatin replication.
Referência(s)