Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Multiple domains are involved in the targeting of the mouse DNA methyltransferase to the DNA replication foci

Multiple domains are involved in the targeting of the mouse DNA methyltransferase to the DNA... 1038–1045 Nucleic Acids Research, 1998, Vol. 26, No. 4  1998 Oxford University Press Multiple domains are involved in the targeting of the mouse DNA methyltransferase to the DNA replication foci + § Yongliang Liu , Edward J. Oakeley, Lijie Sun and Jean-Pierre Jost* Friedrich Miescher-Institut, P.O. Box 2543, Basel, Switzerland Received October 3, 1997; Revised and Accepted December 19, 1997 shorter C-terminal catalytic domain of 500 aa. An additional 118 aa ABSTRACT peptide at the N-terminus of this enzyme has been described by It has been shown that, during the S-phase of the cell Yoder et al. (6) and Tucker et al. (7). The start codon of this newly cycle, the mouse DNA methyltransferase (DNA MTase) defined N-terminus was the only one present in the murine DNA is targeted to sites of DNA replication by an amino acid MTase purified from MEL cells (8). Within the N-terminal domain, sequence (aa 207–455) lying in the N-terminal domain there is a DNA replication foci-targeting sequence (aa 207–455) that of the enzyme [Leonhardt, H., Page, A. W., Weier, H. U. targets the DNA MTase to the sites where its preferred substrate, and Bestor, T. H. (1992) Cell, 71, 865–873]. In this paper hemimethylated DNA, is being synthesized (1). There is a major it is shown, by using enhanced green fluorescent phosphorylation site (Ser396) lying in this motif which was protein (EGFP) fusions, that other peptide sequences identified in the DNA MTase from the MEL cells (8). Downstream of DNA MTase are also involved in this targeting. The of this targeting sequence is a short sequence, that has been shown work focuses on a sequence, downstream of the 2+ to bind Zn , which is homologous to the zinc-binding motif of reported targeting sequence (TS), which is homologous ALL/TRX proteins (4). This is followed by a stretch of sequence to the Polybromo-1 protein. This motif (designated as that is homologous to the Polybromo-1 protein (4,9). This PBHD) is separated from the reported targeting sequence designated as Polybromo-1 protein homologous domain sequence by a zinc-binding motif [Bestor , T. H. (1992) (PBHD) has 23% identity in a 270 aa overlap with the EMBO J, 11, 2611–2617]. Primed in situ extension using Polybromo-1 protein (Liu, unpublished FASTA search). The centromeric-specific primers was used to show that function of this sequence motif has not yet been defined. both the host DNA MTase and EGFP fusion proteins The distribution of DNA MTase changes dynamically in a cell containing the targeting sequences were localized to cycle-dependent manner (10,11). By immunostaining of NIH3T3 centromeric, but not telomeric, regions during late cells with specific antibodies, it has been shown that the mouse S-phase and mitosis. Also found was that, in ~ 10% of DNA MTase forms toroidal structures in middle and late S-phase the S-phase cells, the EGFP fusions did not co-localize cells. In G1 and early S-phase cells the enzyme showed a diffused with the centromeric regions. Mutants containing distribution (1). These S-phase toroidal structures were shown, by either, or both, of these targeting sequences could act in situ hybridization to γ-satellite DNA probes, to be at the as dominant negative mutants against the host DNA position of centromeric heterochromatin (1). MTase. EGFP fusion proteins, containing the reported The B1 sequence (aa 202–369) (containing DNA binding TS (aa 207–455), were targeted to centromeric regions motifs DB1 and AZn) from the N-terminal domain of human throughout the mitotic stage which lead to the DNA MTase has both zinc and DNA-binding activity (12) and its discovery of a similar behavior of the endogenous murine homologue (aa 201–377) is included in its known DNA DNA MTase although the host MTase showed much replication-targeting sequence (12). Constructs over expressing less intense staining than in S-phase cells. The this targeting sequence may therefore have dominant negative biological role of the centromeric localization of DNA effects by competing with the endogenous DNA MTase for the MTase during mitosis is currently unknown. DNA substrate or other factors that recruit the DNA MTase to the DNA replication foci. Measuring changes in the level of genomic INTRODUCTION DNA methylation could assess this effect. Such a dominant The mouse DNA methyltransferase (MTase) (3–5) contains a long negative mutant could be a useful tool for reducing the level of N-terminal regulatory domain of >1000 amino acids (aa) and a genomic DNA methylation, particularly as most human tumor *To whom correspondence should be addressed. Tel: +41 61 697 6688; Fax: +41 61 697 6687; Email: [email protected] Present addresses: Department of Cell and Developmental Biology, University of Pennsylvania School of Medicine, 608 Stellar Chance Labs, 422 Curie Boulevard, Philadelphia, PA 19104-6100, USA and Department of Dermatology, University of Pennsylvania School of Medicine, 230 Clinical Research Building, 422 Curie Boulevard, Philadelphia, PA 19104-6100, USA Downloaded from https://academic.oup.com/nar/article-abstract/26/4/1038/2902024 by Ed 'DeepDyve' Gillespie user on 02 February 2018 1039 Nucleic Acids Research, 1998, Vol. 26, No. 4 1039 Nucleic Acids Research, 1994, Vol. 22, No. 1 cells show elevated levels of DNA MTase (13), a dominant negative mutant of DNA MTase could be of therapeutic value. In this study, the tetracycline-regulated system (14) was used to control the expression of different mouse DNA MTase deletion mutants. These mutants were made as N-terminal fusions with the reporter enhanced green fluorescent protein (EGFP) (Clontech) to provide a fluorescent marker for studying their subcellular localizations. The effects of the expression of these mutants on the DNA methylation level of C3H10T were studied by using a modified SssI methyl-accepting assay (15). MATERIALS AND METHODS Materials and plasmids The EGFP N-fusion series of vectors were purchased from Clontech. pMC1NeoPolyA which contained the neomycin resistance gene (neoR) was obtained from Stratagene. Dr K. Ballmer-Hoffer provided the plasmid pX343 that contained a hygromycin resistance selectable marker. The autoregulatory tetracycline regulated expression system (14) was a gift from Dr D. Pollays (Life Technologies Inc., USA). Oligonucleotides were synthesized in-house by the FMI oligonucleotide Synthesis Lab. G418 sulfate was purchased from GibcoBRL. Restriction enzymes and the Klenow fragment of DNA polymerase I were obtained from Biofinex. The rapid DNA ligation kit, Taq polymerase and Texas Red isothiocyanate (TRITC)-labeled dUTP were purchased from Boehringer Mannheim. Calf intestine alkaline phosphatase was obtained from NEB. Collagen R solutions were purchased from SERVA. The secondary anti- bodies used for indirect antibody staining were purchased from Sigma and Boehringer Mannheim. BrdU, anti-BrdU monoclonal antibodies and Hygromycin B were obtained from CalBiochem. DAPI was purchased from Fluka. All plasmids were prepared with QIAGEN plasmid Midi kit (Qiagen GmbH). Figure 1. DNA MTase-EGFP fusion construct maps. The map at the top represents the full-length wild-type mouse DNA MTase. ‘NN’ refers to the new N-terminal sequence that has recently been described for the mouse MTase Construction of EGFP fusion protein expression vectors (6,7) with the numbers underneath in the parenthesis to indicate the start and the end of this sequence; ‘TS’ is the DNA replication foci-targeting sequence that Part of the newly discovered N-terminal coding sequence (6–8) was described by (1); ‘Zn’ is the zinc/DNA binding motif (2); PBHD is the was amplified by PCR using two overlapping oligo- Polybromo-1 protein homologous domain; ‘PC’ is the catalytic site; ‘GK’ is the linker between the N- and C-terminal domains of DNA MTase and ‘EGFP’ nucleotides and cloned upstream of the EcoRI site of the reported refers to the enhanced green fluorescent protein sequence used as a marker in cDNA sequence to generate the full length of DNA MTase coding the fusion proteins. The black box stands for the intrinsic nuclear localization sequence (Fig. 1, WT). The primers used were: signal (aa 72–92) and the green box stands for the NLS of SV40 large T antigen. 5′-TCTGTCGCTCGAGTCGCCACCATGCCAGCGCGAACA- The numbers written against the wild-type protein sequence refer to amino acid GCTCCAGCCCGAGTGCCTGCGCTTGCCTCCCCGGCAGG- positions. The old numbering of the positions of the amino acids is still used here in order to make comparison with the previous report (2). The numbers CTCGCTCCCGGACCATGTCCGCAGGCGGC-3′ (forward written on the right-hand side of the figure indicate the construct names for each primer) and 5′-TTTGCAGGAATTCATGCAGTAAGTTTAATT- map and the capital letters in parenthesis correspond to the panels in Figure 2. TTCCCTCACACACTCCTTTTCTGTTAAGCCATCTCTTTC- Constructs 1–8 were based on the tetracycline-regulated expression system in CAAGTCTTTGAGCCGCCTGCGGACATGGTCCGGGA- which the promoter (P ) is controlled by the availability of tetracycline (14). tet Constructs 9–11 were based on the EGFP fusion vectors (Clontech) using the GC-3′ (reverse primer). The complementary sequence of the above constitutive, human cytomegalovirus immediate early gene (hCMV) promoter. oligonucleotides and the designed XhoI and EcoRI sites are in bold. The solutions for PCR amplification consisted of 5 μl 10× PCR buffer (Boehringer Mannheim), 5 μl 2 mM dNTPs, 2 μl of each of the above primers (10 μM), 35.5 μl ddH O and 0.5 μl Vent overhangs without changing any of the surrounding amino acids: polymerase (10 U/μl, New England Biolabs). Amplification was 5′-TAGGAAGCTGGTCTATCA-3′ (upper strand) and 5′-ACG- performed in a Stratagene Robocycler for 30 cycles (94C 1 min, TATCCTTCGACCAGATAGTCTAG-3′ (lower strand). 58C 1 min and 72C 40 s). The amplified fragment (183 bp) was Different fragments of the cDNA encoding the N-terminal digested with XhoI and EcoRI and ligated to the XhoI–EcoRI domains of DNA MTase were cloned in the N-fusion EGFP digested construct 2 (Fig. 1). vector (Clontech) (Fig. 1). Those constructs that lacked intrinsic The deletion of the earlier reported targeting sequence was nuclear localization signals (NLS) had the SV40 T antigen NLS achieved by replacing the NsiI–BglII fragment with the following added, in-frame, to their N-termini. The expression cassettes were short double stranded oligonucleotides containing NsiI and BglII then excised and inserted into the plasmid pTet-Splice down- Downloaded from https://academic.oup.com/nar/article-abstract/26/4/1038/2902024 by Ed 'DeepDyve' Gillespie user on 02 February 2018 1040 Nucleic Acids Research, 1998, Vol. 26, No. 4 stream of the tetracycline-regulated promoter (P ). The constructs Conventional and confocal microscopy tet for transient transfection used the constitutive hCMV promoter Laser confocal microscopic scanning was performed using a rather than the tetracycline-regulated promoter. Leica confocal microscope and the SCANWare sofware (Leica). Conventional microscopic observations for the co-localization of EGFP fusion proteins and DAPI staining were performed using Preparation of antiserum against the C-terminal domain of a conventional Leica fluorescence microscope. the mouse DNA MTase DAPI staining The whole C-terminal domain coding-region of the mouse DNA MTase was excised as an XhoI–SpeI fragment, blunted by Fixed cells were permeablized with 1% Triton X100 in PBS at Klenow and inserted into the blunted BamHI site of the vector room temperature for 5 min and then incubated with 1 μg/ml pQE31 (Qiagen) downstream of the 6× His-tag coding sequence. DAPI in PBS at room temperature (in the dark) for 10 min and This construct was expressed in Escherichia coli and the then washed three times with PBS. expressed protein was purified using established protocols (Qiagen). This recombinant protein was used to immunize rabbits Primed in situ extension (PRINS) and the resultant serum was (NH ) SO -precipitated as described 4 2 4 Cells were cultured on collagen-coated 18 × 18 mm coverslips as by Harlow and Lane (16) to raise the polyclonal antibody against described above. Primed in situ synthesis was performed the C-terminal domain of the mouse DNA MTase (anti-MTC). according to the protocols described in the Boehringer Mannheim in situ hybridization manual (2nd Edition) using a Perkin Elmer PCR machine. The chromosomes were labeled by the incorporation Cell culture and stable transfection of TRITC-dUTP during the extension of either centromeric- or telomeric-specific primers by Taq polymerase. The sequences of Dr A. Baumeister (FMI, Basel) provided the C3H10T cell line. A the primers were: centromeric-specific (5′-ATTTAGAAATGTC- subclone called C3.3, which was stable after 15 passages, was used CACTGTAGGAC-3′) (18) and telomeric-specific (5′-TTAGG- for the transfections in this study. Transfections were performed GTTAGGGTTAGGG-3′) (19). Indirect immunofluorescence was when the cells reached 40–50% confluency. Transient transfections performed after the PRINS reaction using an anti-GFP monoclonal were performed using SuperFect (Qiagen) according to the antibody (Clontech) and an FITC-conjugated secondary antibody manufacturer’s instructions. Stable transfection of C3.3 cells was (goat anti-mouse, Sigma) to bring up the intensity of the EGFP as performed using the calcium phosphate procedure (17). The C3.3 the heating step in the PRINS reaction seriously damaged the cells were co-transfected with the transactivator expression plasmid EGFP fluorophore. pTet-tTAk and the neomycin-resistance plasmid pMC1NeoPolyA (Strategene). The cells were selected using 600 μg/ml of G418 Genomic DNA extraction and SssI methyl-group accepting sulfate for 12 days before the clones were picked up. The cloned assay cells were selected at the same concentration of G418 a week Cells that were stably transfected with the tetracycline-regulated before being tested for transgene expression. Clone TA5, which constructs (1,4,5,7,8) were passed three times with an overall stably expressed the transactivator, was used for the second round induction time of 72 h. Cells were scraped off the culture dishes of co-transfection using the mutant expression vectors and the using a rubber policeman and their genomic DNA was extracted plasmid pX343 carrying hygromycin-resistance gene. Stable transfectants were selected using 150 μg/ml hygromycin B. according to Sambrook et al. (20). To assess the dominant negative effect of the DNA MTase–EGFP fusion proteins the genome wide DNA methylation level was measured using a modified SssI methyl-group accepting assay (15). Briefly each reaction contained Indirect immunofluorescence 1 μg of genomic DNA digested to completion by EcoRI, 6 U of SssI methylase, 2 μM of [ H]S-adenosyl-methionine (SAM) and 18 μM The cells were cultured on 18 × 18 mm microscopy coverslips of cold SAM. After 2 h incubation at 37C, the reactions were (Menzel-Gläser, Germany). Cells were processed 16 h after stopped by adding 500 μl 20% TCA using 100 μg BSA as carrier. induction of expression (stable clones with tetracycline controlled The samples were vortexed for 20 s and incubated on ice for constructs) or 16 h post-transfection (transiently transfected cells 30 min, the precipitates were spun down, washed three times with with constitutive expression constructs). Prior to fixation they 20% TCA and hydrolyzed overnight with 100 μl of formic acid. were washed twice in PBS. Fixation was performed by incubating The samples were then counted for c.p.m. using a scintillation the cells on ice in 4% fresh paraformaldehyde solution for 30 min. counter. Every genomic DNA sample was tested in triplicate. The paraformaldehyde was removed by two PBS washes and the cells were then permeablized with 1% Triton-in PBS at room RESULTS temperature for 10 min. After two washes with PBS, the cells were incubated with blocking solution (10% goat serum in PBS) Subcellular localization of EGFP fusion proteins containing at room temperature for 30 min followed by incubation with the different parts of the N-terminal domain of the DNA fractionated anti-C terminal antibody (1:25 dilution) for 1 h at methyltransferase room temperature. After three washes with PBS, the cells were incubated with the secondary antibody using the concentration Stable transfection of C3.3 cells with the full DNA MTase suggested by the manufacturer. The cells were then washed three N-terminal domain, under the control of a constitutive promoter, times with PBS and mounted on glass slides in Vectorshield was not successful and may be due to the toxicity of this (Vector laboratories, Inc., USA) mounting medium. constitutive expression of the truncated DNA MTase (7). In order Downloaded from https://academic.oup.com/nar/article-abstract/26/4/1038/2902024 by Ed 'DeepDyve' Gillespie user on 02 February 2018 1041 Nucleic Acids Research, 1998, Vol. 26, No. 4 1041 Nucleic Acids Research, 1994, Vol. 22, No. 1 to obtain a high level of stable expression and to reduce any possible toxic side effects, a modified tetracycline-regulated expression system (14) was used. In this system, both the transgene and the transactivator (the tetracycline repressor-VP16 fusion protein) (21) expression were driven by the tetracycline- regulated promoter (P ) (14). The regulation of expression from tet this construct was tested by northern blot hybridization (20). The EGFP was used as a fusion partner to monitor the expression of the mutants. The screening of the clones expressing the fusion proteins was performed as described in the method section. Fluorescent clones were then cultured on collagen-coated glass cover slips. Expression was induced for 24 h by changing the culture medium to tetracycline free medium. Figure 2 shows that the mutant DNA MTase–EGFP fusion proteins, that included the DNA replication TS, were all localized to the nucleus and formed toroidal or spotty structures in S-phase cells similar to those reported by Leonhardt et al. (1) (Fig. 2C and E–G). Surprisingly, construct 5, in which two thirds of the reported replication foci targeting sequence (aa 294–506) had been deleted, also showed toroidal structures although they were much less abundant (Fig. 2D). This construct contained the reported zinc-binding motif (2) and a stretch of sequence homologous to the chicken Polybromo-1 protein (PBHD, aa574–846) (4,9). Construct 7 (aa 1–138), which did not contain any part of the TS, also showed the toroidal structures as well as a relatively high level of diffused localization (Fig. 2B). On the other hand, the nuclear-localized EGFP alone showed only a diffused nuclear distribution (Fig. 2A) Figure 2. EGFP fluorescence from the MTase fusion constructs. Cells stably transfected with the constructs using the tetracycline-regulated expression vectors and the wild-type EGFP showed strong fluorescence in both the were cultured on glass coverslips and induced for expression for 24 h. Confocal cytoplasm and the nucleus (Fig. 2H). The above results indicated scanning microscopy was performed as described in the Materials and Methods. clearly that, whilst the reported DNA replication foci-targeting The equatorial section for each construct is shown. (A) Construct 8; (B) construct sequence was sufficient to target the EGFP fusion proteins to sites 7; (C) construct 4; (D) construct 5; (E) construct 3; (F) construct 2; (G) construct 1 and (H) construct 12 (wild-type EGFP, controlled by P ). The scale bars are of DNA replication, other sequences within the protein might also tet 5 μm in length. able to perform this function. Co-localization of the GFP fusion proteins with the host DNA showed mutual exclusivity (Fig. 3B, construct 11), that is to say MTase showed that the PBHD sequence could also target the cells either had strong intensities of EGFP or host MTase staining enzyme to DNA replication foci but not both (Fig. 3B, constructs 5, 7, 11). The more EGFP expression observed the less host MTase was found. Construct 9 To test whether or not construct 5 was indeed targeted to sites of showed a similar behavior to construct 11 (data not shown). This DNA replication, cells were pulse-labeled with BrdU after 16 h indicated that the PBHD was responsible for the DNA replication of induction of expression, and stained with an anti-BrdU antibody foci targeting. Construct 1, which contained the full length of as described by Leonhardt et al. (1). Figure 3A (top row) shows N-terminal domain of the DNA MTase containing both the TS that, in the wild-type cells, host DNA MTase always co-localized and PBHD, showed in general a good co-localization with the with the incorporated BrdU. Construct 5 co-localized with the host MTase but in some regions it showed more red (Host MTase) incorporated BrdU (Fig. 3A, bottom row) but was observed in the than green (EGFP) or vice versa (Fig. 3B, construct 1). Construct same cell as the host DNA MTase at an extremely low frequency 4 in some cases showed co-localization with the host MTase and (Fig. 3B, construct 5*) and often showed mutual localization some times showed mutual exclusivity (Fig. 3B, construct 4). All exclusivity with the host MTase (Fig. 3B, construct 5). While of the above results indicate that these MTase–EGFP fusions could construct 6 showed a diffused nuclear localization, the host DNA be, in principle, acting as competitors for the host protein on the MTase in the same cell showed toroidal or granular structures DNA binding sites. (Fig. 3B). This indicated that the sequence responsible for the DNA replication foci targeting in construct 5 was lying between DNA MTase was found to be targeted to chromosomal aa 506 and 846, which contains the zinc binding motif and the regions other than just the centromeric regions PBHD. To further restrict the sequence required for this localization, three other constructs 9, 10 and 11 (Fig. 1) were We determined the localization of the host MTase and the mutant made under the control of the constitutive hCMV promoter. Cells MTase–GFP fusions (construct 1, 4, 5 and 9) by immuno- were transiently transfected and stained with anti-MTC. The fluorescence and compared their distributions to those of the construct 10 showed a diffused nuclear distribution (Fig. 3B) centromeric and telomeric sequences in interphase and S-phase whereas construct 11 which contained the PBHD sequence cells. The localization of the centromeres and telomeres was sometimes showed co-localization with the host DNA MTase made by PRINS reactions using specific primer sets. The similar to construct 5 (Fig. 3B, construct 5*) and sometimes specificity of the PRINS was tested in the following ways: Downloaded from https://academic.oup.com/nar/article-abstract/26/4/1038/2902024 by Ed 'DeepDyve' Gillespie user on 02 February 2018 1042 Nucleic Acids Research, 1998, Vol. 26, No. 4 (i) single-strand DNA breaks could be blocked by a pre-treatment with ddNTPs without preventing the PRINS from working; (ii) the pattern of spots obtained was primer dependent [i.e. no spots were obtained from cells incubated without primers (data not shown)]; (iii) PRINS experiments, performed on condensed chromosomes, showed that the spots obtained with the primers were localized to the telomeric or centromeric chromosomal regions (Fig. 4A). The chromosomal localization of the DNA MTase appeared to be cell cycle dependent, in that ~ 10% of the cells failed to show co-localization with the centromeric regions (Fig. 4B, constructs 1*, 4*and 5*) whereas the other 90% did localize (Fig. 4B, constructs 1, 4, 5 and 9). We never observed telomeric localization of either the host MTase or the EGFP constructs (data not shown). To determine the distribution of the EGFP-fusion protein along the condensed chromosomes, DAPI staining of these cells was performed. The DAPI-stained cells were observed by conventional fluorescence microscopy. Figure 4C shows composite images of green (EGFP) and blue (DAPI) mitotic cells (constructs 1 and 4). It was clear that the EGFP fusions containing the TS localized to the structures reminiscent of centromeres during mitosis. Given the apparent EGFP localization, the question about the location of the host MTase was addressed by staining untransfected cells with anti-MTC. DAPI staining was used to visualize the mitotic chromosomes. Mitotic cells showed a lower staining level of host MTase than their S-phase counterparts, however, it was clear that the host MTase did indeed localize to centromeric regions preferentially (data not shown). The truncated DNA MTase–GFP fusion proteins can affect the action of the host DNA MTase in a dominant negative manner The mutual exclusivity between the GFP fusion proteins and the host DNA MTase indicated that the mutants might exert a dominant negative effect based on a competition mechanism. The potential effect of the mutants was assessed by the changes in the genome-wide DNA methylation of clones stably expressing the mutants using the SssI methyl-group accepting assay (15). The assay was designed to carry the reactions to completion. The more [ H]SAM the DNA samples incorporate, the less methylated they are. The differences between the incorporation into the control DNA (isolated from the cells that only express the nuclear-localized EGFP; Fig. 1, construct 8) and the other constructs tested are summarized in Table 1. A negative value indicates that the DNA methylation level is higher than the control and a positive value Figure 3. Co-localization of host DNA MTase with replication foci and the MTase–EGFP fusion protein. (A) Shows Br-dUTP incorporation into the replication foci, in red, of either untransfected (top row) or construct 5-expressing cells (bottom row). Host MTase staining is shown in blue (top row) and EGFP fluorescence, in green, from construct 5 (bottom row). The-right-hand column shows composite images of the left and middle columns. (B) Shows confocal images of the co-localization between the host MTase and the EGFP fusion constructs. Numbers on left-hand side of the panels stand for the names of the constructs. 5* shows that, under rare circumstances, construct 5 which contains mainly the polybromo domain PBHD and a small part of the targeting sequence TS, indeed co-localized with the host DNA MTase whereas in most cases it showed mutual exclusivity with the host MTase (Fig. 3B, construct 5). For constructs 5 and 11 there are two nuclei per panel whereas in all the other panels there is only one nucleus and as a consequence the magnification of these two panels is reduced. The scale bars are 5 μm in length. Downloaded from https://academic.oup.com/nar/article-abstract/26/4/1038/2902024 by Ed 'DeepDyve' Gillespie user on 02 February 2018 1043 Nucleic Acids Research, 1998, Vol. 26, No. 4 1043 Nucleic Acids Research, 1994, Vol. 22, No. 1 significantly increased the CH accepting capacity, indicating a decrease in the level of genomic DNA methylation. The most extreme case was construct 4, which resulted in the largest degree of demethylation and also produced the highest and most stable levels of EGFP fusion protein expression (data not shown). The delta values were converted into percentage changes of genomic DNA methylation according to the formulae described in the legend for Table 1. These changes were 3.8 ± 0.3%, 5.7 ± 0.6%, 2.9 ± 1.7% and 1.0 ± 0.9% for the constructs 1, 2, 4 and 7, respectively (Table 1, right hand column). DISCUSSION Multiple domains are involved in the targeting of the mouse DNA MTase to the DNA replication foci DNA MTase–EGFP fusion proteins, which contained the DNA replication foci TS of the mouse DNA MTase, behaved as suggested by Leonhardt et al. (1). However, construct 5 (aa 138–294; 507–846), which contained only the first 87 aa of the TS, and construct 7 (aa 1–138), which did not contain any part of the TS, also formed the toroidal structures typical for MTase staining of S-phase cells (Fig. 2B and D). Further analysis showed that construct 5 also co-localized with the incorporated BrdU (Fig. 3A, bottom row) and construct 7 showed co-localization with the host DNA MTase (Fig 3B, construct 7) while the host DNA MTase always co-localized with the BrdU incorporation (Fig. 3A, top row). This indicated that, while the TS may be sufficient for targeting, other sequences in the N-terminal domain are also involved in this targeting. However, this targeting was not always equal in efficiency. For example, the fusion proteins expressed from construct 7 showed both diffused distributions, reminiscent of the nuclear-localized EGFP control and toroidal structures similar to those observed from the constructs containing the published TS. Neither the nuclear-localized EGFP (construct 8) nor the wild-type EGFP (construct 12) controls ever showed granular structures (Fig. 2A and H). Construct 10 (aa 138–294) showed a similar diffused distribution as construct 8 while the host DNA MTase showed granular structures in the same cell. These data indicate that EGFP targeting to the replication foci may be performed by the PBHD and the sequence between aa 1–138, albeit less efficiently than the TS. The precise localization of the EGFP-constructs to replication foci was confirmed by their co-localization with BrdU (Fig. 3A and B). While these data show that there is more than one sequence involved in the DNA replication foci targeting they also show that the TS sequence is sufficient to fulfil this targeting on its own. Given the lack of an obvious DNA-binding motif in construct Figure 4. Primed in situ extension (PRINS) detection of centromeric and 7, it is possible that it might be targeted to DNA replication sites telomeric sequences. (A) Shows two control PRINS experiments using either via an indirect process such as an interaction with other proteins the centromeric or telomeric primers to demonstrate their specificity. (B) Shows or nucleic acids that are themselves targeted to the replication EGFP expression and centromere-specific PRINS. The numbers on the left are the construct names. Those marked with a ‘*’ failed to co-localize with the foci. While as yet we are unable to support or reject this centromeric PRINS (~ 10% of S-phase cells). In (B)9 there are two nuclei hypothesis, it should be noted that this protein region is highly whereas all other panels have only one. (C) Shows EGFP (constructs 1 and 4, charged and part of the peptide sequence shows high similarity to green) localized to the chromosomal structures reminiscent of centromeres. The the 70 kDa U1 snRNP and to the Su (w ) protein (1). The right hand panels shows the composite images of the EGFP (green) and DAPI construct 5 contained a sequence from the N-terminal part of the staining (blue). The scale bars are 5 μm in length. known targeting sequence (aa 207–294, the full length of the TS is from aa 207–455) which includes the sequences corresponding indicates lower methylation. Construct 7 caused a non-significant to the reported DNA binding motifs DB1 and AZn from the (P > 0.05) decrease in the methyl group (CH ) accepting capacity, human DNA MTases (12). However, construct 6 (aa 138–294), whereas the other three constructs (constructs 1, 4 and 5) containing the homologous sequences of these motifs, did not Downloaded from https://academic.oup.com/nar/article-abstract/26/4/1038/2902024 by Ed 'DeepDyve' Gillespie user on 02 February 2018 1044 Nucleic Acids Research, 1998, Vol. 26, No. 4 show any targeting (Fig. 3B) indicating that the DNA binding Chromatin association of the DNA MTase–GFP fusion activity alone was not sufficient for the targeting. In this paper, we proteins have restricted the sequence responsible for DNA replication It has been shown that DNA MTase can be associated with targeting in construct 5 to the PBHD sequence. chromatin (22–24). An extensive study of the association of the MTase–EGFP fusion proteins to the chromatin was performed Table 1. Dominant negative effect of the mutants tested using SssI using the PRINS techniques with either centromeric or telomeric methyl-group accepting assay specific primers. Consistent with the previous report (1), it was found that ~ 90% of the centromeric PRINS co-localized with the Constructs Δpmol CH /μg DNA EGFP fusions containing any of either the TS or PBHD (Fig 4B). 8 (reference) 0 However, it was also found that, in ~ 10% of the cells examined, the EGFP fusions formed granular structures that failed to 1 1.7  0.14 co-localize with the centromeric PRINS (Fig. 4B, constructs 1*, 4 2.6  0.28 4* and 5*). The satellite DNA sequences located near the 5 1.3  0.77 centromeric regions of the mouse chromosomes are heavily methylated at CpG sites (24) and are replicated at the very end of 7 –0.45  0.39 S-phase (25–29). After DNA replication, there would be a large number of hemimethylated sites, therefore, it is not surprising that One microgram of genomic DNA of each clone was used for the SssI methyl-group (CH ) accepting assay. The incorporated c.p.m. were converted into pmol CH a mechanism existing to recruit the DNA MTase to a high 3 3 accepted per μg DNA. The value of the control sample (Construct 8, the nuclear concentration to ensure the methylation pattern faithfully passed localized EGFP) was taken as the basal level (18.1 pmol/μg DNA) and the differ- from mother cells to daughter cells. The other 10% of cases may ences in the incorporation between other clones to the control were calculated. The have reflected genomic regions other than centromeric regions left column shows the constructs tested. The middle column shows the differences that also have high concentration of methylated CpG sites. in the incorporation of [ H]CH compared with the control (construct 8) ± S.E. The Although subtelomeric satellite DNA has also been shown to be right hand column shows an estimate of the changes in genomic DNA methylation heavily methylated in mouse somatic cells (30), we never assuming that the control cells have 60% of their total CG dinucleotides methylated observed the EGFP fusion proteins co-localize with the telomeric at C5 (34). The formulae used for this calculation were as follows: X / Z = 1 – 0.6 PRINS (data not shown). and C = (Y – X)/ Z. Where ‘X’ is the amount of [ H]CH incorporated into 1 μg of EGFP-fusion proteins were also found to localize to the the control genomic DNA (18.1 pmol in this case). ‘Z’ is the total number of CG dinucleotides in 1 μg of genomic DNA (45.25 pmol in this case). ‘Y’ is the amount centromeric regions in mitotic cells (Fig. 4C, constructs 4 and 8) of [ H]CH incorporated into the listed constructs. ‘C’ is the percentage change in 3 while the nuclear-localized EGFP showed a diffused distribution genomic DNA methylation. at this stage (data not shown). The host DNA MTase showed lower staining in mitotic cells than in S-phase cells, but it was also preferentially localized to the centromeres. Although it has been The apparent exclusivity between the host MTase and the reported that DNA methylation occurs several hours after ‘targeting sequence’ containing EGFP fusion proteins indicated replication (31), it was striking that both the EGFP fusion proteins a competition between these two. Because of this competition, the and the host MTase were localized to heterochromatic regions relative intensity of the red (host MTase) and green (EGFP-fusion during mitosis. The biological role of this heterochromatic protein) was dependent on the time at which the transgene was localization of the DNA needs to be further investigated. induced. If it was induced late in S-phase then some of the sites were presumably already occupied by the host MTase prior to The toxicity of the mutants may be due to the sequestration transgene expression. Thus, when the transgene was expressed it of the DNA replication machinery rather than a change in had much less opportunity to access these sites. However, if the genomic DNA methylation transgene was expressed at a high level in G1 phase, where the host MTase is at its lowest level (11), the EGFP fusion protein Initial attempts to express mutants, containing the entire N-terminal would occupy more of the DNA replication foci as the cell cycle domain of MTase, constitutively resulted in no stable-expressing progressed towards S-phase. While this may explain the observed clones (data not shown). This could be due to the toxic effects of exclusivity between the host MTase and the EGFP fusion these DNA MTase mutants as has been discussed by Tucker et al. proteins, potential qualitative differences between these two (7). Subsequent experiments made use of the autoregulatory could not be excluded. The EGFP fusion proteins that contained tetracycline-controlled expression system (14) to permit us to the PBHD and the zinc finger motif, in the absence of the TS, study these potentially toxic proteins. showed similar localization patterns to construct 5 which Because the mutants were targeted to the DNA replication foci, it contained only a small part of the TS and the zinc finger motif, is possible that the high levels of their expression might titrate out in addition to the PBHD (Fig. 3B, construct 5). They do not, several critical components of the DNA replication machinery or however, always co-localize with the endogenous DNA MTase else upset the regulation of DNA replication through the direct (Fig. 3B, constructs 5 and 11). In many cases the exclusivity binding of the mutants to genomic DNA. The work of Chuang et al. between these PBHD-containing constructs and the host MTase (12) revealed the presence of multiple DNA binding motifs in the was stronger than for those containing either the TS alone N-terminal domain of the human DNA MTase. Tucker et al. (7) (Fig 3B, construct 4) or the TS and PBHD together (Fig 3B, proposed that the toxicity of MTase overexpression might be construct 1). These differences could possibly be explained by the attributed to the reported TS. Our construct 4, which contained the intrinsic property of PBHD to favor stronger protein–protein TS and a very short sequence (aa 138–206) upstream of it, was interactions and thus show a stronger sequestration of the able to stably overexpress this sequence at a very high level and replication machinery. could be passed for prolonged periods (2 weeks) without obvious Downloaded from https://academic.oup.com/nar/article-abstract/26/4/1038/2902024 by Ed 'DeepDyve' Gillespie user on 02 February 2018 1045 Nucleic Acids Research, 1998, Vol. 26, No. 4 1045 Nucleic Acids Research, 1994, Vol. 22, No. 1 5 Jost, J. P. and Saluz, H. P. (1993) DNA Methylation: Molecular Biology toxic effects whereas any clone containing the PBHD would be lost and Biological Significance. Birkhäuser Verlag, Basel, Boston, Berlin, pp.572. in less than a week of passing with a significant loss of fluorescent 6 Yoder, J. A., Yen, R. W. C., Vertino, P. M., Bestor, T. H. and Baylin, S. B. cells 24–48 h post-induction (Liu, unpublished observations). Thus (1996) J. Biol. Chem., 271, 31092–31097. we hypothesize that this apparent toxic effect is most probably due 7 Tucker, K. L., Talbot, D., Lee, M. A., Leonhardt, H. and Jaenisch, R. (1996) Proc. Natl. Acad. Sci. USA, 93, 12920–12925. to PBHD sequence and that its over-expression sequesters critical 8 Glickman, J. F., Pavlovich, J. G. and Reich, N. O. (1997) J. Biol. Chem., components of the DNA replication machinery through protein- 272, 17851–17857. protein interaction resulting in either the loss of the construct or 9 Nicolas, R. H. and Goodwin, G. H. (1996) Gene, 175, 233–240. cell death. This idea of protein–protein interactions is consistent 10 Adams, R. L. P. and Burdon, R. H. (eds) (1985) Molecular Biology of with the observations that DNA MTase can be tightly bound to the DNA Methylation. Springer Verlag, New York. 11 Vogel, M. C., Papadopoulos, T., Muller-Hermelink, H. K., Drahovsky, D. nuclear matrix (32), possibly for enzyme storage, that it can be and Pfeifer, G. P. (1988) FEBS Lett., 236, 9–13. associated with chromatin (22–24) and that it can also interact 12 Chuang, L. S., Ng, H. H., Chia, J. N. and Li, B. F. (1996) J. Mol. Biol., with histone H1 strongly (Suetake et al., Poster, FASEB, 257, 935–948. Biological methylation, 1997). 13 Szyf, M. (1996) Pharmacol. Ther., 70, 1–37. The constructs containing the TS and PBHD could serve as 14 Shockett, P., Difilippantonio, M., Hellman, N. and Schatz, D. G. (1995) Proc. Natl. Acad. Sci. USA, 92, 6522–6526. dominant negative competitors for the host DNA MTase and 15 Schmitt, F., Oakeley, E. J. and Jost, J. P. (1997) J. Biol. Chem., 272, cause hypomethylation of the genome (Table 1). Construct 4 1534–1540. which expressed the highest level of the fusion protein, showed 16 Harlow, E. and Lane, D. (1988) Antibodies, A Laboratory Guide. the highest effect on the DNA methylation level, but was less Cold Spring Harbor Laboratory University Press, Cold Spring Harbor, NY. 17 Chen, C. and Okayama, H. (1987) Mol. Cell. Biol., 7, 2745–2752. toxic than the others (data not shown) indicating the toxic effect 18 Vissel, B. and Choo, K. H. (1989) Genomics, 5, 407–414. of the mutants might not be due to the changes in the DNA 19 Kipling, D., Wilson, H. E., Thomson, E. J. and Cooke, H. J. (1995) methylation level. This was supported by the fact that embryonic Hum. Mol. Genet., 4, 1007–1014. stem cells that had only 1/3 of the genomic DNA methylation 20 Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989) Molecular Cloning: level of the wild-type cells could grow normally (33). A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 21 Gossen, M. and Bujard, H. (1992) Proc. Natl. Acad. Sci. USA, 89, ACKNOWLEDGEMENTS 5547–5551. 22 Creusot, F., Christman, J. K. (1981) Nucleic Acids Res 9, 5359–5381. We are grateful to Dr D. Pollays of Life Technologies Inc., USA 23 Caiafa, P., Mastrantonio, S., Cacace, F., Attina, M., Rispoli, M. and Strom, R. for providing the tetracycline-regulated expression system and (1988) Biochim. Biophys. Acta, 951, 191–200. Prof. T. Bestor for his original DNA methyltransferase construct. 24 Caiafa, P., Mastrantonio, S., Attina, M., Rispoli, M., Reale, A. and Strom, R. (1988) Biochem. Int., 17, 863–875. We would like to thank Dr J. Hagmann for assisting confocal 25 Selig, S., Ariel, M., Goitein, R., Marcus, M. and Cedar, H. (1988) EMBO J., scanning microscopy, Drs B. Ludin and S. Kaech for helping with 7, 419–426. the conventional fluorescence microscopy. We would also like to 26 Church, K. (1965) Genetics, 52, 843–849. thank Dr M. Sinnreich for providing the GFP antibody. We 27 Dev, V. G., Grewal, M. S., Miller, D. A., Kouri, R. E., Hutton, J. J. and Miller, O. J. (1971) Cytogenetics, 10, 436–451. greatly appreciate Drs J. Paskowsky and S. Schwarz for critical 28 Hsu, T. C. and Markvong, A. (1975) Chromosoma, 51, 311–322. reading of this manuscript. 29 Miller, O. J. (1976) Chromosoma, 55, 165–170. 30 de Lange, T., Shiue, L., Myers, R. M., Cox, D. R., Naylor, S. L., Killery, A. M. and Varmus, H. E. (1990) Mol. Cell. Biol., 10, 518–527. REFERENCES 31 Woodcock, D. M., Adams, J. K. and Cooper, I. A. (1982) 1 Leonhardt, H., Page, A. W., Weier, H. U. and Bestor, T. H. (1992) Cell, 71, Biochim. Biophys. Acta, 696, 15–22. 865–873. 32 Hubscher, U., Pedrali-Noy, G., Knust-Kron, B., Doerfler, W., Spadari, S. 2 Bestor, T. H. (1992) EMBO J., 11, 2611–2617. (1985) Anal. Biochem., 150, 442–448. 3 Adams, R. L. P. (1995) BioEssays, 17, 139–145. 33 Li, E., Bestor, T. H. and Jaenisch, R. (1992) Cell, 69, 915–926. 4 Bestor, T. H. (1996) In Riggs, A. D. Matienssen, R. A. and Russo, V. E. A. 34 Leonhardt, H. and Bestor, T.H. (1993) In Jost and Saluz (ed.) DNA (ed.), Epigenetics. Cold Spring Harbor Laboratory Press, Methylation: Molecular Biology and Biological Significance. Birkhäuser Cold Spring Harbor, NY, pp. 61–76. Verlag, Basel, Boston, Berlin, pp. 109–119. Downloaded from https://academic.oup.com/nar/article-abstract/26/4/1038/2902024 by Ed 'DeepDyve' Gillespie user on 02 February 2018 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nucleic Acids Research Oxford University Press

Multiple domains are involved in the targeting of the mouse DNA methyltransferase to the DNA replication foci

Download PDF
8 pages

Loading next page...
 
/lp/oxford-university-press/multiple-domains-are-involved-in-the-targeting-of-the-mouse-dna-dLdj4dfKa9

References (2)

Publisher
Oxford University Press
Copyright
© 1998 Oxford University Press
ISSN
0305-1048
eISSN
1362-4962
DOI
10.1093/nar/26.4.1038
Publisher site
See Article on Publisher Site

Abstract

1038–1045 Nucleic Acids Research, 1998, Vol. 26, No. 4  1998 Oxford University Press Multiple domains are involved in the targeting of the mouse DNA methyltransferase to the DNA replication foci + § Yongliang Liu , Edward J. Oakeley, Lijie Sun and Jean-Pierre Jost* Friedrich Miescher-Institut, P.O. Box 2543, Basel, Switzerland Received October 3, 1997; Revised and Accepted December 19, 1997 shorter C-terminal catalytic domain of 500 aa. An additional 118 aa ABSTRACT peptide at the N-terminus of this enzyme has been described by It has been shown that, during the S-phase of the cell Yoder et al. (6) and Tucker et al. (7). The start codon of this newly cycle, the mouse DNA methyltransferase (DNA MTase) defined N-terminus was the only one present in the murine DNA is targeted to sites of DNA replication by an amino acid MTase purified from MEL cells (8). Within the N-terminal domain, sequence (aa 207–455) lying in the N-terminal domain there is a DNA replication foci-targeting sequence (aa 207–455) that of the enzyme [Leonhardt, H., Page, A. W., Weier, H. U. targets the DNA MTase to the sites where its preferred substrate, and Bestor, T. H. (1992) Cell, 71, 865–873]. In this paper hemimethylated DNA, is being synthesized (1). There is a major it is shown, by using enhanced green fluorescent phosphorylation site (Ser396) lying in this motif which was protein (EGFP) fusions, that other peptide sequences identified in the DNA MTase from the MEL cells (8). Downstream of DNA MTase are also involved in this targeting. The of this targeting sequence is a short sequence, that has been shown work focuses on a sequence, downstream of the 2+ to bind Zn , which is homologous to the zinc-binding motif of reported targeting sequence (TS), which is homologous ALL/TRX proteins (4). This is followed by a stretch of sequence to the Polybromo-1 protein. This motif (designated as that is homologous to the Polybromo-1 protein (4,9). This PBHD) is separated from the reported targeting sequence designated as Polybromo-1 protein homologous domain sequence by a zinc-binding motif [Bestor , T. H. (1992) (PBHD) has 23% identity in a 270 aa overlap with the EMBO J, 11, 2611–2617]. Primed in situ extension using Polybromo-1 protein (Liu, unpublished FASTA search). The centromeric-specific primers was used to show that function of this sequence motif has not yet been defined. both the host DNA MTase and EGFP fusion proteins The distribution of DNA MTase changes dynamically in a cell containing the targeting sequences were localized to cycle-dependent manner (10,11). By immunostaining of NIH3T3 centromeric, but not telomeric, regions during late cells with specific antibodies, it has been shown that the mouse S-phase and mitosis. Also found was that, in ~ 10% of DNA MTase forms toroidal structures in middle and late S-phase the S-phase cells, the EGFP fusions did not co-localize cells. In G1 and early S-phase cells the enzyme showed a diffused with the centromeric regions. Mutants containing distribution (1). These S-phase toroidal structures were shown, by either, or both, of these targeting sequences could act in situ hybridization to γ-satellite DNA probes, to be at the as dominant negative mutants against the host DNA position of centromeric heterochromatin (1). MTase. EGFP fusion proteins, containing the reported The B1 sequence (aa 202–369) (containing DNA binding TS (aa 207–455), were targeted to centromeric regions motifs DB1 and AZn) from the N-terminal domain of human throughout the mitotic stage which lead to the DNA MTase has both zinc and DNA-binding activity (12) and its discovery of a similar behavior of the endogenous murine homologue (aa 201–377) is included in its known DNA DNA MTase although the host MTase showed much replication-targeting sequence (12). Constructs over expressing less intense staining than in S-phase cells. The this targeting sequence may therefore have dominant negative biological role of the centromeric localization of DNA effects by competing with the endogenous DNA MTase for the MTase during mitosis is currently unknown. DNA substrate or other factors that recruit the DNA MTase to the DNA replication foci. Measuring changes in the level of genomic INTRODUCTION DNA methylation could assess this effect. Such a dominant The mouse DNA methyltransferase (MTase) (3–5) contains a long negative mutant could be a useful tool for reducing the level of N-terminal regulatory domain of >1000 amino acids (aa) and a genomic DNA methylation, particularly as most human tumor *To whom correspondence should be addressed. Tel: +41 61 697 6688; Fax: +41 61 697 6687; Email: [email protected] Present addresses: Department of Cell and Developmental Biology, University of Pennsylvania School of Medicine, 608 Stellar Chance Labs, 422 Curie Boulevard, Philadelphia, PA 19104-6100, USA and Department of Dermatology, University of Pennsylvania School of Medicine, 230 Clinical Research Building, 422 Curie Boulevard, Philadelphia, PA 19104-6100, USA Downloaded from https://academic.oup.com/nar/article-abstract/26/4/1038/2902024 by Ed 'DeepDyve' Gillespie user on 02 February 2018 1039 Nucleic Acids Research, 1998, Vol. 26, No. 4 1039 Nucleic Acids Research, 1994, Vol. 22, No. 1 cells show elevated levels of DNA MTase (13), a dominant negative mutant of DNA MTase could be of therapeutic value. In this study, the tetracycline-regulated system (14) was used to control the expression of different mouse DNA MTase deletion mutants. These mutants were made as N-terminal fusions with the reporter enhanced green fluorescent protein (EGFP) (Clontech) to provide a fluorescent marker for studying their subcellular localizations. The effects of the expression of these mutants on the DNA methylation level of C3H10T were studied by using a modified SssI methyl-accepting assay (15). MATERIALS AND METHODS Materials and plasmids The EGFP N-fusion series of vectors were purchased from Clontech. pMC1NeoPolyA which contained the neomycin resistance gene (neoR) was obtained from Stratagene. Dr K. Ballmer-Hoffer provided the plasmid pX343 that contained a hygromycin resistance selectable marker. The autoregulatory tetracycline regulated expression system (14) was a gift from Dr D. Pollays (Life Technologies Inc., USA). Oligonucleotides were synthesized in-house by the FMI oligonucleotide Synthesis Lab. G418 sulfate was purchased from GibcoBRL. Restriction enzymes and the Klenow fragment of DNA polymerase I were obtained from Biofinex. The rapid DNA ligation kit, Taq polymerase and Texas Red isothiocyanate (TRITC)-labeled dUTP were purchased from Boehringer Mannheim. Calf intestine alkaline phosphatase was obtained from NEB. Collagen R solutions were purchased from SERVA. The secondary anti- bodies used for indirect antibody staining were purchased from Sigma and Boehringer Mannheim. BrdU, anti-BrdU monoclonal antibodies and Hygromycin B were obtained from CalBiochem. DAPI was purchased from Fluka. All plasmids were prepared with QIAGEN plasmid Midi kit (Qiagen GmbH). Figure 1. DNA MTase-EGFP fusion construct maps. The map at the top represents the full-length wild-type mouse DNA MTase. ‘NN’ refers to the new N-terminal sequence that has recently been described for the mouse MTase Construction of EGFP fusion protein expression vectors (6,7) with the numbers underneath in the parenthesis to indicate the start and the end of this sequence; ‘TS’ is the DNA replication foci-targeting sequence that Part of the newly discovered N-terminal coding sequence (6–8) was described by (1); ‘Zn’ is the zinc/DNA binding motif (2); PBHD is the was amplified by PCR using two overlapping oligo- Polybromo-1 protein homologous domain; ‘PC’ is the catalytic site; ‘GK’ is the linker between the N- and C-terminal domains of DNA MTase and ‘EGFP’ nucleotides and cloned upstream of the EcoRI site of the reported refers to the enhanced green fluorescent protein sequence used as a marker in cDNA sequence to generate the full length of DNA MTase coding the fusion proteins. The black box stands for the intrinsic nuclear localization sequence (Fig. 1, WT). The primers used were: signal (aa 72–92) and the green box stands for the NLS of SV40 large T antigen. 5′-TCTGTCGCTCGAGTCGCCACCATGCCAGCGCGAACA- The numbers written against the wild-type protein sequence refer to amino acid GCTCCAGCCCGAGTGCCTGCGCTTGCCTCCCCGGCAGG- positions. The old numbering of the positions of the amino acids is still used here in order to make comparison with the previous report (2). The numbers CTCGCTCCCGGACCATGTCCGCAGGCGGC-3′ (forward written on the right-hand side of the figure indicate the construct names for each primer) and 5′-TTTGCAGGAATTCATGCAGTAAGTTTAATT- map and the capital letters in parenthesis correspond to the panels in Figure 2. TTCCCTCACACACTCCTTTTCTGTTAAGCCATCTCTTTC- Constructs 1–8 were based on the tetracycline-regulated expression system in CAAGTCTTTGAGCCGCCTGCGGACATGGTCCGGGA- which the promoter (P ) is controlled by the availability of tetracycline (14). tet Constructs 9–11 were based on the EGFP fusion vectors (Clontech) using the GC-3′ (reverse primer). The complementary sequence of the above constitutive, human cytomegalovirus immediate early gene (hCMV) promoter. oligonucleotides and the designed XhoI and EcoRI sites are in bold. The solutions for PCR amplification consisted of 5 μl 10× PCR buffer (Boehringer Mannheim), 5 μl 2 mM dNTPs, 2 μl of each of the above primers (10 μM), 35.5 μl ddH O and 0.5 μl Vent overhangs without changing any of the surrounding amino acids: polymerase (10 U/μl, New England Biolabs). Amplification was 5′-TAGGAAGCTGGTCTATCA-3′ (upper strand) and 5′-ACG- performed in a Stratagene Robocycler for 30 cycles (94C 1 min, TATCCTTCGACCAGATAGTCTAG-3′ (lower strand). 58C 1 min and 72C 40 s). The amplified fragment (183 bp) was Different fragments of the cDNA encoding the N-terminal digested with XhoI and EcoRI and ligated to the XhoI–EcoRI domains of DNA MTase were cloned in the N-fusion EGFP digested construct 2 (Fig. 1). vector (Clontech) (Fig. 1). Those constructs that lacked intrinsic The deletion of the earlier reported targeting sequence was nuclear localization signals (NLS) had the SV40 T antigen NLS achieved by replacing the NsiI–BglII fragment with the following added, in-frame, to their N-termini. The expression cassettes were short double stranded oligonucleotides containing NsiI and BglII then excised and inserted into the plasmid pTet-Splice down- Downloaded from https://academic.oup.com/nar/article-abstract/26/4/1038/2902024 by Ed 'DeepDyve' Gillespie user on 02 February 2018 1040 Nucleic Acids Research, 1998, Vol. 26, No. 4 stream of the tetracycline-regulated promoter (P ). The constructs Conventional and confocal microscopy tet for transient transfection used the constitutive hCMV promoter Laser confocal microscopic scanning was performed using a rather than the tetracycline-regulated promoter. Leica confocal microscope and the SCANWare sofware (Leica). Conventional microscopic observations for the co-localization of EGFP fusion proteins and DAPI staining were performed using Preparation of antiserum against the C-terminal domain of a conventional Leica fluorescence microscope. the mouse DNA MTase DAPI staining The whole C-terminal domain coding-region of the mouse DNA MTase was excised as an XhoI–SpeI fragment, blunted by Fixed cells were permeablized with 1% Triton X100 in PBS at Klenow and inserted into the blunted BamHI site of the vector room temperature for 5 min and then incubated with 1 μg/ml pQE31 (Qiagen) downstream of the 6× His-tag coding sequence. DAPI in PBS at room temperature (in the dark) for 10 min and This construct was expressed in Escherichia coli and the then washed three times with PBS. expressed protein was purified using established protocols (Qiagen). This recombinant protein was used to immunize rabbits Primed in situ extension (PRINS) and the resultant serum was (NH ) SO -precipitated as described 4 2 4 Cells were cultured on collagen-coated 18 × 18 mm coverslips as by Harlow and Lane (16) to raise the polyclonal antibody against described above. Primed in situ synthesis was performed the C-terminal domain of the mouse DNA MTase (anti-MTC). according to the protocols described in the Boehringer Mannheim in situ hybridization manual (2nd Edition) using a Perkin Elmer PCR machine. The chromosomes were labeled by the incorporation Cell culture and stable transfection of TRITC-dUTP during the extension of either centromeric- or telomeric-specific primers by Taq polymerase. The sequences of Dr A. Baumeister (FMI, Basel) provided the C3H10T cell line. A the primers were: centromeric-specific (5′-ATTTAGAAATGTC- subclone called C3.3, which was stable after 15 passages, was used CACTGTAGGAC-3′) (18) and telomeric-specific (5′-TTAGG- for the transfections in this study. Transfections were performed GTTAGGGTTAGGG-3′) (19). Indirect immunofluorescence was when the cells reached 40–50% confluency. Transient transfections performed after the PRINS reaction using an anti-GFP monoclonal were performed using SuperFect (Qiagen) according to the antibody (Clontech) and an FITC-conjugated secondary antibody manufacturer’s instructions. Stable transfection of C3.3 cells was (goat anti-mouse, Sigma) to bring up the intensity of the EGFP as performed using the calcium phosphate procedure (17). The C3.3 the heating step in the PRINS reaction seriously damaged the cells were co-transfected with the transactivator expression plasmid EGFP fluorophore. pTet-tTAk and the neomycin-resistance plasmid pMC1NeoPolyA (Strategene). The cells were selected using 600 μg/ml of G418 Genomic DNA extraction and SssI methyl-group accepting sulfate for 12 days before the clones were picked up. The cloned assay cells were selected at the same concentration of G418 a week Cells that were stably transfected with the tetracycline-regulated before being tested for transgene expression. Clone TA5, which constructs (1,4,5,7,8) were passed three times with an overall stably expressed the transactivator, was used for the second round induction time of 72 h. Cells were scraped off the culture dishes of co-transfection using the mutant expression vectors and the using a rubber policeman and their genomic DNA was extracted plasmid pX343 carrying hygromycin-resistance gene. Stable transfectants were selected using 150 μg/ml hygromycin B. according to Sambrook et al. (20). To assess the dominant negative effect of the DNA MTase–EGFP fusion proteins the genome wide DNA methylation level was measured using a modified SssI methyl-group accepting assay (15). Briefly each reaction contained Indirect immunofluorescence 1 μg of genomic DNA digested to completion by EcoRI, 6 U of SssI methylase, 2 μM of [ H]S-adenosyl-methionine (SAM) and 18 μM The cells were cultured on 18 × 18 mm microscopy coverslips of cold SAM. After 2 h incubation at 37C, the reactions were (Menzel-Gläser, Germany). Cells were processed 16 h after stopped by adding 500 μl 20% TCA using 100 μg BSA as carrier. induction of expression (stable clones with tetracycline controlled The samples were vortexed for 20 s and incubated on ice for constructs) or 16 h post-transfection (transiently transfected cells 30 min, the precipitates were spun down, washed three times with with constitutive expression constructs). Prior to fixation they 20% TCA and hydrolyzed overnight with 100 μl of formic acid. were washed twice in PBS. Fixation was performed by incubating The samples were then counted for c.p.m. using a scintillation the cells on ice in 4% fresh paraformaldehyde solution for 30 min. counter. Every genomic DNA sample was tested in triplicate. The paraformaldehyde was removed by two PBS washes and the cells were then permeablized with 1% Triton-in PBS at room RESULTS temperature for 10 min. After two washes with PBS, the cells were incubated with blocking solution (10% goat serum in PBS) Subcellular localization of EGFP fusion proteins containing at room temperature for 30 min followed by incubation with the different parts of the N-terminal domain of the DNA fractionated anti-C terminal antibody (1:25 dilution) for 1 h at methyltransferase room temperature. After three washes with PBS, the cells were incubated with the secondary antibody using the concentration Stable transfection of C3.3 cells with the full DNA MTase suggested by the manufacturer. The cells were then washed three N-terminal domain, under the control of a constitutive promoter, times with PBS and mounted on glass slides in Vectorshield was not successful and may be due to the toxicity of this (Vector laboratories, Inc., USA) mounting medium. constitutive expression of the truncated DNA MTase (7). In order Downloaded from https://academic.oup.com/nar/article-abstract/26/4/1038/2902024 by Ed 'DeepDyve' Gillespie user on 02 February 2018 1041 Nucleic Acids Research, 1998, Vol. 26, No. 4 1041 Nucleic Acids Research, 1994, Vol. 22, No. 1 to obtain a high level of stable expression and to reduce any possible toxic side effects, a modified tetracycline-regulated expression system (14) was used. In this system, both the transgene and the transactivator (the tetracycline repressor-VP16 fusion protein) (21) expression were driven by the tetracycline- regulated promoter (P ) (14). The regulation of expression from tet this construct was tested by northern blot hybridization (20). The EGFP was used as a fusion partner to monitor the expression of the mutants. The screening of the clones expressing the fusion proteins was performed as described in the method section. Fluorescent clones were then cultured on collagen-coated glass cover slips. Expression was induced for 24 h by changing the culture medium to tetracycline free medium. Figure 2 shows that the mutant DNA MTase–EGFP fusion proteins, that included the DNA replication TS, were all localized to the nucleus and formed toroidal or spotty structures in S-phase cells similar to those reported by Leonhardt et al. (1) (Fig. 2C and E–G). Surprisingly, construct 5, in which two thirds of the reported replication foci targeting sequence (aa 294–506) had been deleted, also showed toroidal structures although they were much less abundant (Fig. 2D). This construct contained the reported zinc-binding motif (2) and a stretch of sequence homologous to the chicken Polybromo-1 protein (PBHD, aa574–846) (4,9). Construct 7 (aa 1–138), which did not contain any part of the TS, also showed the toroidal structures as well as a relatively high level of diffused localization (Fig. 2B). On the other hand, the nuclear-localized EGFP alone showed only a diffused nuclear distribution (Fig. 2A) Figure 2. EGFP fluorescence from the MTase fusion constructs. Cells stably transfected with the constructs using the tetracycline-regulated expression vectors and the wild-type EGFP showed strong fluorescence in both the were cultured on glass coverslips and induced for expression for 24 h. Confocal cytoplasm and the nucleus (Fig. 2H). The above results indicated scanning microscopy was performed as described in the Materials and Methods. clearly that, whilst the reported DNA replication foci-targeting The equatorial section for each construct is shown. (A) Construct 8; (B) construct sequence was sufficient to target the EGFP fusion proteins to sites 7; (C) construct 4; (D) construct 5; (E) construct 3; (F) construct 2; (G) construct 1 and (H) construct 12 (wild-type EGFP, controlled by P ). The scale bars are of DNA replication, other sequences within the protein might also tet 5 μm in length. able to perform this function. Co-localization of the GFP fusion proteins with the host DNA showed mutual exclusivity (Fig. 3B, construct 11), that is to say MTase showed that the PBHD sequence could also target the cells either had strong intensities of EGFP or host MTase staining enzyme to DNA replication foci but not both (Fig. 3B, constructs 5, 7, 11). The more EGFP expression observed the less host MTase was found. Construct 9 To test whether or not construct 5 was indeed targeted to sites of showed a similar behavior to construct 11 (data not shown). This DNA replication, cells were pulse-labeled with BrdU after 16 h indicated that the PBHD was responsible for the DNA replication of induction of expression, and stained with an anti-BrdU antibody foci targeting. Construct 1, which contained the full length of as described by Leonhardt et al. (1). Figure 3A (top row) shows N-terminal domain of the DNA MTase containing both the TS that, in the wild-type cells, host DNA MTase always co-localized and PBHD, showed in general a good co-localization with the with the incorporated BrdU. Construct 5 co-localized with the host MTase but in some regions it showed more red (Host MTase) incorporated BrdU (Fig. 3A, bottom row) but was observed in the than green (EGFP) or vice versa (Fig. 3B, construct 1). Construct same cell as the host DNA MTase at an extremely low frequency 4 in some cases showed co-localization with the host MTase and (Fig. 3B, construct 5*) and often showed mutual localization some times showed mutual exclusivity (Fig. 3B, construct 4). All exclusivity with the host MTase (Fig. 3B, construct 5). While of the above results indicate that these MTase–EGFP fusions could construct 6 showed a diffused nuclear localization, the host DNA be, in principle, acting as competitors for the host protein on the MTase in the same cell showed toroidal or granular structures DNA binding sites. (Fig. 3B). This indicated that the sequence responsible for the DNA replication foci targeting in construct 5 was lying between DNA MTase was found to be targeted to chromosomal aa 506 and 846, which contains the zinc binding motif and the regions other than just the centromeric regions PBHD. To further restrict the sequence required for this localization, three other constructs 9, 10 and 11 (Fig. 1) were We determined the localization of the host MTase and the mutant made under the control of the constitutive hCMV promoter. Cells MTase–GFP fusions (construct 1, 4, 5 and 9) by immuno- were transiently transfected and stained with anti-MTC. The fluorescence and compared their distributions to those of the construct 10 showed a diffused nuclear distribution (Fig. 3B) centromeric and telomeric sequences in interphase and S-phase whereas construct 11 which contained the PBHD sequence cells. The localization of the centromeres and telomeres was sometimes showed co-localization with the host DNA MTase made by PRINS reactions using specific primer sets. The similar to construct 5 (Fig. 3B, construct 5*) and sometimes specificity of the PRINS was tested in the following ways: Downloaded from https://academic.oup.com/nar/article-abstract/26/4/1038/2902024 by Ed 'DeepDyve' Gillespie user on 02 February 2018 1042 Nucleic Acids Research, 1998, Vol. 26, No. 4 (i) single-strand DNA breaks could be blocked by a pre-treatment with ddNTPs without preventing the PRINS from working; (ii) the pattern of spots obtained was primer dependent [i.e. no spots were obtained from cells incubated without primers (data not shown)]; (iii) PRINS experiments, performed on condensed chromosomes, showed that the spots obtained with the primers were localized to the telomeric or centromeric chromosomal regions (Fig. 4A). The chromosomal localization of the DNA MTase appeared to be cell cycle dependent, in that ~ 10% of the cells failed to show co-localization with the centromeric regions (Fig. 4B, constructs 1*, 4*and 5*) whereas the other 90% did localize (Fig. 4B, constructs 1, 4, 5 and 9). We never observed telomeric localization of either the host MTase or the EGFP constructs (data not shown). To determine the distribution of the EGFP-fusion protein along the condensed chromosomes, DAPI staining of these cells was performed. The DAPI-stained cells were observed by conventional fluorescence microscopy. Figure 4C shows composite images of green (EGFP) and blue (DAPI) mitotic cells (constructs 1 and 4). It was clear that the EGFP fusions containing the TS localized to the structures reminiscent of centromeres during mitosis. Given the apparent EGFP localization, the question about the location of the host MTase was addressed by staining untransfected cells with anti-MTC. DAPI staining was used to visualize the mitotic chromosomes. Mitotic cells showed a lower staining level of host MTase than their S-phase counterparts, however, it was clear that the host MTase did indeed localize to centromeric regions preferentially (data not shown). The truncated DNA MTase–GFP fusion proteins can affect the action of the host DNA MTase in a dominant negative manner The mutual exclusivity between the GFP fusion proteins and the host DNA MTase indicated that the mutants might exert a dominant negative effect based on a competition mechanism. The potential effect of the mutants was assessed by the changes in the genome-wide DNA methylation of clones stably expressing the mutants using the SssI methyl-group accepting assay (15). The assay was designed to carry the reactions to completion. The more [ H]SAM the DNA samples incorporate, the less methylated they are. The differences between the incorporation into the control DNA (isolated from the cells that only express the nuclear-localized EGFP; Fig. 1, construct 8) and the other constructs tested are summarized in Table 1. A negative value indicates that the DNA methylation level is higher than the control and a positive value Figure 3. Co-localization of host DNA MTase with replication foci and the MTase–EGFP fusion protein. (A) Shows Br-dUTP incorporation into the replication foci, in red, of either untransfected (top row) or construct 5-expressing cells (bottom row). Host MTase staining is shown in blue (top row) and EGFP fluorescence, in green, from construct 5 (bottom row). The-right-hand column shows composite images of the left and middle columns. (B) Shows confocal images of the co-localization between the host MTase and the EGFP fusion constructs. Numbers on left-hand side of the panels stand for the names of the constructs. 5* shows that, under rare circumstances, construct 5 which contains mainly the polybromo domain PBHD and a small part of the targeting sequence TS, indeed co-localized with the host DNA MTase whereas in most cases it showed mutual exclusivity with the host MTase (Fig. 3B, construct 5). For constructs 5 and 11 there are two nuclei per panel whereas in all the other panels there is only one nucleus and as a consequence the magnification of these two panels is reduced. The scale bars are 5 μm in length. Downloaded from https://academic.oup.com/nar/article-abstract/26/4/1038/2902024 by Ed 'DeepDyve' Gillespie user on 02 February 2018 1043 Nucleic Acids Research, 1998, Vol. 26, No. 4 1043 Nucleic Acids Research, 1994, Vol. 22, No. 1 significantly increased the CH accepting capacity, indicating a decrease in the level of genomic DNA methylation. The most extreme case was construct 4, which resulted in the largest degree of demethylation and also produced the highest and most stable levels of EGFP fusion protein expression (data not shown). The delta values were converted into percentage changes of genomic DNA methylation according to the formulae described in the legend for Table 1. These changes were 3.8 ± 0.3%, 5.7 ± 0.6%, 2.9 ± 1.7% and 1.0 ± 0.9% for the constructs 1, 2, 4 and 7, respectively (Table 1, right hand column). DISCUSSION Multiple domains are involved in the targeting of the mouse DNA MTase to the DNA replication foci DNA MTase–EGFP fusion proteins, which contained the DNA replication foci TS of the mouse DNA MTase, behaved as suggested by Leonhardt et al. (1). However, construct 5 (aa 138–294; 507–846), which contained only the first 87 aa of the TS, and construct 7 (aa 1–138), which did not contain any part of the TS, also formed the toroidal structures typical for MTase staining of S-phase cells (Fig. 2B and D). Further analysis showed that construct 5 also co-localized with the incorporated BrdU (Fig. 3A, bottom row) and construct 7 showed co-localization with the host DNA MTase (Fig 3B, construct 7) while the host DNA MTase always co-localized with the BrdU incorporation (Fig. 3A, top row). This indicated that, while the TS may be sufficient for targeting, other sequences in the N-terminal domain are also involved in this targeting. However, this targeting was not always equal in efficiency. For example, the fusion proteins expressed from construct 7 showed both diffused distributions, reminiscent of the nuclear-localized EGFP control and toroidal structures similar to those observed from the constructs containing the published TS. Neither the nuclear-localized EGFP (construct 8) nor the wild-type EGFP (construct 12) controls ever showed granular structures (Fig. 2A and H). Construct 10 (aa 138–294) showed a similar diffused distribution as construct 8 while the host DNA MTase showed granular structures in the same cell. These data indicate that EGFP targeting to the replication foci may be performed by the PBHD and the sequence between aa 1–138, albeit less efficiently than the TS. The precise localization of the EGFP-constructs to replication foci was confirmed by their co-localization with BrdU (Fig. 3A and B). While these data show that there is more than one sequence involved in the DNA replication foci targeting they also show that the TS sequence is sufficient to fulfil this targeting on its own. Given the lack of an obvious DNA-binding motif in construct Figure 4. Primed in situ extension (PRINS) detection of centromeric and 7, it is possible that it might be targeted to DNA replication sites telomeric sequences. (A) Shows two control PRINS experiments using either via an indirect process such as an interaction with other proteins the centromeric or telomeric primers to demonstrate their specificity. (B) Shows or nucleic acids that are themselves targeted to the replication EGFP expression and centromere-specific PRINS. The numbers on the left are the construct names. Those marked with a ‘*’ failed to co-localize with the foci. While as yet we are unable to support or reject this centromeric PRINS (~ 10% of S-phase cells). In (B)9 there are two nuclei hypothesis, it should be noted that this protein region is highly whereas all other panels have only one. (C) Shows EGFP (constructs 1 and 4, charged and part of the peptide sequence shows high similarity to green) localized to the chromosomal structures reminiscent of centromeres. The the 70 kDa U1 snRNP and to the Su (w ) protein (1). The right hand panels shows the composite images of the EGFP (green) and DAPI construct 5 contained a sequence from the N-terminal part of the staining (blue). The scale bars are 5 μm in length. known targeting sequence (aa 207–294, the full length of the TS is from aa 207–455) which includes the sequences corresponding indicates lower methylation. Construct 7 caused a non-significant to the reported DNA binding motifs DB1 and AZn from the (P > 0.05) decrease in the methyl group (CH ) accepting capacity, human DNA MTases (12). However, construct 6 (aa 138–294), whereas the other three constructs (constructs 1, 4 and 5) containing the homologous sequences of these motifs, did not Downloaded from https://academic.oup.com/nar/article-abstract/26/4/1038/2902024 by Ed 'DeepDyve' Gillespie user on 02 February 2018 1044 Nucleic Acids Research, 1998, Vol. 26, No. 4 show any targeting (Fig. 3B) indicating that the DNA binding Chromatin association of the DNA MTase–GFP fusion activity alone was not sufficient for the targeting. In this paper, we proteins have restricted the sequence responsible for DNA replication It has been shown that DNA MTase can be associated with targeting in construct 5 to the PBHD sequence. chromatin (22–24). An extensive study of the association of the MTase–EGFP fusion proteins to the chromatin was performed Table 1. Dominant negative effect of the mutants tested using SssI using the PRINS techniques with either centromeric or telomeric methyl-group accepting assay specific primers. Consistent with the previous report (1), it was found that ~ 90% of the centromeric PRINS co-localized with the Constructs Δpmol CH /μg DNA EGFP fusions containing any of either the TS or PBHD (Fig 4B). 8 (reference) 0 However, it was also found that, in ~ 10% of the cells examined, the EGFP fusions formed granular structures that failed to 1 1.7  0.14 co-localize with the centromeric PRINS (Fig. 4B, constructs 1*, 4 2.6  0.28 4* and 5*). The satellite DNA sequences located near the 5 1.3  0.77 centromeric regions of the mouse chromosomes are heavily methylated at CpG sites (24) and are replicated at the very end of 7 –0.45  0.39 S-phase (25–29). After DNA replication, there would be a large number of hemimethylated sites, therefore, it is not surprising that One microgram of genomic DNA of each clone was used for the SssI methyl-group (CH ) accepting assay. The incorporated c.p.m. were converted into pmol CH a mechanism existing to recruit the DNA MTase to a high 3 3 accepted per μg DNA. The value of the control sample (Construct 8, the nuclear concentration to ensure the methylation pattern faithfully passed localized EGFP) was taken as the basal level (18.1 pmol/μg DNA) and the differ- from mother cells to daughter cells. The other 10% of cases may ences in the incorporation between other clones to the control were calculated. The have reflected genomic regions other than centromeric regions left column shows the constructs tested. The middle column shows the differences that also have high concentration of methylated CpG sites. in the incorporation of [ H]CH compared with the control (construct 8) ± S.E. The Although subtelomeric satellite DNA has also been shown to be right hand column shows an estimate of the changes in genomic DNA methylation heavily methylated in mouse somatic cells (30), we never assuming that the control cells have 60% of their total CG dinucleotides methylated observed the EGFP fusion proteins co-localize with the telomeric at C5 (34). The formulae used for this calculation were as follows: X / Z = 1 – 0.6 PRINS (data not shown). and C = (Y – X)/ Z. Where ‘X’ is the amount of [ H]CH incorporated into 1 μg of EGFP-fusion proteins were also found to localize to the the control genomic DNA (18.1 pmol in this case). ‘Z’ is the total number of CG dinucleotides in 1 μg of genomic DNA (45.25 pmol in this case). ‘Y’ is the amount centromeric regions in mitotic cells (Fig. 4C, constructs 4 and 8) of [ H]CH incorporated into the listed constructs. ‘C’ is the percentage change in 3 while the nuclear-localized EGFP showed a diffused distribution genomic DNA methylation. at this stage (data not shown). The host DNA MTase showed lower staining in mitotic cells than in S-phase cells, but it was also preferentially localized to the centromeres. Although it has been The apparent exclusivity between the host MTase and the reported that DNA methylation occurs several hours after ‘targeting sequence’ containing EGFP fusion proteins indicated replication (31), it was striking that both the EGFP fusion proteins a competition between these two. Because of this competition, the and the host MTase were localized to heterochromatic regions relative intensity of the red (host MTase) and green (EGFP-fusion during mitosis. The biological role of this heterochromatic protein) was dependent on the time at which the transgene was localization of the DNA needs to be further investigated. induced. If it was induced late in S-phase then some of the sites were presumably already occupied by the host MTase prior to The toxicity of the mutants may be due to the sequestration transgene expression. Thus, when the transgene was expressed it of the DNA replication machinery rather than a change in had much less opportunity to access these sites. However, if the genomic DNA methylation transgene was expressed at a high level in G1 phase, where the host MTase is at its lowest level (11), the EGFP fusion protein Initial attempts to express mutants, containing the entire N-terminal would occupy more of the DNA replication foci as the cell cycle domain of MTase, constitutively resulted in no stable-expressing progressed towards S-phase. While this may explain the observed clones (data not shown). This could be due to the toxic effects of exclusivity between the host MTase and the EGFP fusion these DNA MTase mutants as has been discussed by Tucker et al. proteins, potential qualitative differences between these two (7). Subsequent experiments made use of the autoregulatory could not be excluded. The EGFP fusion proteins that contained tetracycline-controlled expression system (14) to permit us to the PBHD and the zinc finger motif, in the absence of the TS, study these potentially toxic proteins. showed similar localization patterns to construct 5 which Because the mutants were targeted to the DNA replication foci, it contained only a small part of the TS and the zinc finger motif, is possible that the high levels of their expression might titrate out in addition to the PBHD (Fig. 3B, construct 5). They do not, several critical components of the DNA replication machinery or however, always co-localize with the endogenous DNA MTase else upset the regulation of DNA replication through the direct (Fig. 3B, constructs 5 and 11). In many cases the exclusivity binding of the mutants to genomic DNA. The work of Chuang et al. between these PBHD-containing constructs and the host MTase (12) revealed the presence of multiple DNA binding motifs in the was stronger than for those containing either the TS alone N-terminal domain of the human DNA MTase. Tucker et al. (7) (Fig 3B, construct 4) or the TS and PBHD together (Fig 3B, proposed that the toxicity of MTase overexpression might be construct 1). These differences could possibly be explained by the attributed to the reported TS. Our construct 4, which contained the intrinsic property of PBHD to favor stronger protein–protein TS and a very short sequence (aa 138–206) upstream of it, was interactions and thus show a stronger sequestration of the able to stably overexpress this sequence at a very high level and replication machinery. could be passed for prolonged periods (2 weeks) without obvious Downloaded from https://academic.oup.com/nar/article-abstract/26/4/1038/2902024 by Ed 'DeepDyve' Gillespie user on 02 February 2018 1045 Nucleic Acids Research, 1998, Vol. 26, No. 4 1045 Nucleic Acids Research, 1994, Vol. 22, No. 1 5 Jost, J. P. and Saluz, H. P. (1993) DNA Methylation: Molecular Biology toxic effects whereas any clone containing the PBHD would be lost and Biological Significance. Birkhäuser Verlag, Basel, Boston, Berlin, pp.572. in less than a week of passing with a significant loss of fluorescent 6 Yoder, J. A., Yen, R. W. C., Vertino, P. M., Bestor, T. H. and Baylin, S. B. cells 24–48 h post-induction (Liu, unpublished observations). Thus (1996) J. Biol. Chem., 271, 31092–31097. we hypothesize that this apparent toxic effect is most probably due 7 Tucker, K. L., Talbot, D., Lee, M. A., Leonhardt, H. and Jaenisch, R. (1996) Proc. Natl. Acad. Sci. USA, 93, 12920–12925. to PBHD sequence and that its over-expression sequesters critical 8 Glickman, J. F., Pavlovich, J. G. and Reich, N. O. (1997) J. Biol. Chem., components of the DNA replication machinery through protein- 272, 17851–17857. protein interaction resulting in either the loss of the construct or 9 Nicolas, R. H. and Goodwin, G. H. (1996) Gene, 175, 233–240. cell death. This idea of protein–protein interactions is consistent 10 Adams, R. L. P. and Burdon, R. H. (eds) (1985) Molecular Biology of with the observations that DNA MTase can be tightly bound to the DNA Methylation. Springer Verlag, New York. 11 Vogel, M. C., Papadopoulos, T., Muller-Hermelink, H. K., Drahovsky, D. nuclear matrix (32), possibly for enzyme storage, that it can be and Pfeifer, G. P. (1988) FEBS Lett., 236, 9–13. associated with chromatin (22–24) and that it can also interact 12 Chuang, L. S., Ng, H. H., Chia, J. N. and Li, B. F. (1996) J. Mol. Biol., with histone H1 strongly (Suetake et al., Poster, FASEB, 257, 935–948. Biological methylation, 1997). 13 Szyf, M. (1996) Pharmacol. Ther., 70, 1–37. The constructs containing the TS and PBHD could serve as 14 Shockett, P., Difilippantonio, M., Hellman, N. and Schatz, D. G. (1995) Proc. Natl. Acad. Sci. USA, 92, 6522–6526. dominant negative competitors for the host DNA MTase and 15 Schmitt, F., Oakeley, E. J. and Jost, J. P. (1997) J. Biol. Chem., 272, cause hypomethylation of the genome (Table 1). Construct 4 1534–1540. which expressed the highest level of the fusion protein, showed 16 Harlow, E. and Lane, D. (1988) Antibodies, A Laboratory Guide. the highest effect on the DNA methylation level, but was less Cold Spring Harbor Laboratory University Press, Cold Spring Harbor, NY. 17 Chen, C. and Okayama, H. (1987) Mol. Cell. Biol., 7, 2745–2752. toxic than the others (data not shown) indicating the toxic effect 18 Vissel, B. and Choo, K. H. (1989) Genomics, 5, 407–414. of the mutants might not be due to the changes in the DNA 19 Kipling, D., Wilson, H. E., Thomson, E. J. and Cooke, H. J. (1995) methylation level. This was supported by the fact that embryonic Hum. Mol. Genet., 4, 1007–1014. stem cells that had only 1/3 of the genomic DNA methylation 20 Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989) Molecular Cloning: level of the wild-type cells could grow normally (33). A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 21 Gossen, M. and Bujard, H. (1992) Proc. Natl. Acad. Sci. USA, 89, ACKNOWLEDGEMENTS 5547–5551. 22 Creusot, F., Christman, J. K. (1981) Nucleic Acids Res 9, 5359–5381. We are grateful to Dr D. Pollays of Life Technologies Inc., USA 23 Caiafa, P., Mastrantonio, S., Cacace, F., Attina, M., Rispoli, M. and Strom, R. for providing the tetracycline-regulated expression system and (1988) Biochim. Biophys. Acta, 951, 191–200. Prof. T. Bestor for his original DNA methyltransferase construct. 24 Caiafa, P., Mastrantonio, S., Attina, M., Rispoli, M., Reale, A. and Strom, R. (1988) Biochem. Int., 17, 863–875. We would like to thank Dr J. Hagmann for assisting confocal 25 Selig, S., Ariel, M., Goitein, R., Marcus, M. and Cedar, H. (1988) EMBO J., scanning microscopy, Drs B. Ludin and S. Kaech for helping with 7, 419–426. the conventional fluorescence microscopy. We would also like to 26 Church, K. (1965) Genetics, 52, 843–849. thank Dr M. Sinnreich for providing the GFP antibody. We 27 Dev, V. G., Grewal, M. S., Miller, D. A., Kouri, R. E., Hutton, J. J. and Miller, O. J. (1971) Cytogenetics, 10, 436–451. greatly appreciate Drs J. Paskowsky and S. Schwarz for critical 28 Hsu, T. C. and Markvong, A. (1975) Chromosoma, 51, 311–322. reading of this manuscript. 29 Miller, O. J. (1976) Chromosoma, 55, 165–170. 30 de Lange, T., Shiue, L., Myers, R. M., Cox, D. R., Naylor, S. L., Killery, A. M. and Varmus, H. E. (1990) Mol. Cell. Biol., 10, 518–527. REFERENCES 31 Woodcock, D. M., Adams, J. K. and Cooper, I. A. (1982) 1 Leonhardt, H., Page, A. W., Weier, H. U. and Bestor, T. H. (1992) Cell, 71, Biochim. Biophys. Acta, 696, 15–22. 865–873. 32 Hubscher, U., Pedrali-Noy, G., Knust-Kron, B., Doerfler, W., Spadari, S. 2 Bestor, T. H. (1992) EMBO J., 11, 2611–2617. (1985) Anal. Biochem., 150, 442–448. 3 Adams, R. L. P. (1995) BioEssays, 17, 139–145. 33 Li, E., Bestor, T. H. and Jaenisch, R. (1992) Cell, 69, 915–926. 4 Bestor, T. H. (1996) In Riggs, A. D. Matienssen, R. A. and Russo, V. E. A. 34 Leonhardt, H. and Bestor, T.H. (1993) In Jost and Saluz (ed.) DNA (ed.), Epigenetics. Cold Spring Harbor Laboratory Press, Methylation: Molecular Biology and Biological Significance. Birkhäuser Cold Spring Harbor, NY, pp. 61–76. Verlag, Basel, Boston, Berlin, pp. 109–119. Downloaded from https://academic.oup.com/nar/article-abstract/26/4/1038/2902024 by Ed 'DeepDyve' Gillespie user on 02 February 2018

Journal

Nucleic Acids Research – Oxford University Press

Published: Feb 1, 1998

There are no references for this article.