Cell cycle–regulated phosphorylation of p220^NPAT by cyclin E/Cdk2 in Cajal bodies promotes histone gene transcription

p220 foci are associated with CBs and with chromosomes 1 and 6

The size and number of p220 foci observed in S-phase cells are reminiscent of those displayed by CBs, as detected by antibodies against a component p80^coilin. CBs are present in variable numbers in tissue culture cell lines (three to eight CBs/cell) (Frey and Matera 1995; Almeida et al. 1998). Because CBs typically are difficult to detect in nontransformed cells, we used antigen retrieval to examine whether p220 might be associated with CBs in normal human dermal fibroblasts. As shown in Figure ​Figure3a,3a, p220 foci coincide with coilin-containing CBs. In contrast to the colocalization of p220 and coilin throughout most of the cell cycle observed with three lines of fibroblasts (diploid dermal fibroblasts, WI38 lung fibroblasts, and bjTERT fibroblasts), transformed cells, including HeLa, Caski, SiHa, and MCF7 cells, displayed a larger and more variable number of p220 foci, ranging from three to >12 (data not shown). In HeLa cells, most if not all of the p220 foci are associated with CBs, but only a subset of CBs is associated with p220 foci (Fig. ​(Fig.33c).

Because CBs have previously been demonstrated to associate with histone gene clusters on chromosomes 1 and 6, it follows that one or more p220 foci may be expected to associate with these chromosomal domains. By using interphase chromosome painting in normal dermal fibroblasts, we found that chromosome 6 signals were closely associated with both p220 foci in 100% of cells containing two p220 foci. In 87% of cells containing four foci, the signals were associated with two foci, whereas the remaining 13% had more than two associated foci (Fig. ​(Fig.3d;3d; Table ​Table1).1). In contrast, chromosome 1 signals typically were not associated with p220 foci in cells containing two foci, but 93% of cells with four foci had two associated foci (Fig. ​(Fig.3e;3e; Table ​Table1).1). In our experience, false positive association occurs at a frequency of ∼10% or lower. Because a subset of CBs is physically associated with an snRNA U2 gene loci at 17q21 (Frey and Matera 1995), we painted chromosome 17 as well as chromosomes 5 and Y as additional controls. These chromosomal domains displayed only rare association with p220 foci (Fig. ​(Fig.3e).3e). For example, out of 400 cells, two were found to have one p220 focus associated with chromosome Y. Taken together, these data indicate that p220 foci are intimately linked with chromosome 1– and chromosome 6–associated CBs and that the chromosome 6 domain is associated with p220 foci throughout the cell cycle, whereas association with chromosome 1 occurs during S phase and coincides with the increase in p220 foci from two to four. These data also imply the existence of mechanisms that restrict association of p220 with particular chromosomes to particular points in the cycle. The increased number of p220 foci observed in some transformed cells (data not shown) likely reflects at least in part an increased ploidy in chromosomes 1 and 6.

p220 is specifically phosphorylated by cyclin E/Cdk2 on sites near the cyclin E interaction domain

Cyclin E and p220 co-immunoprecipitate from cell extracts, and cyclin E/Cdk2 can phosphorylate associated p220 (Zhao et al. 1998). To elucidate the role of cyclin E/Cdk2 in p220 regulation, we sought to determine the specificity of phosphorylation. Initially we examined the ability of p220 to bind to various cyclin/Cdk complexes. Flag-tagged p220 was expressed in insect cells and cell lysates used in binding assays with immobilized cyclin/Cdk complexes (Fig. ​(Fig.4).4). Although p220 associated efficiently with the cyclin E/Cdk2 complex (lane 16), it did not associate with the cyclin D1/Cdk4, cyclin A/Cdk1, or cyclin B/Cdk1 complex (lanes 4, 10, and 13, respectively) and bound only weakly with cyclin A/Cdk2 (lane 7). Thus, p220 displays specificity for cyclin E/Cdk2.

Open in a separate window

Figure 4

p220 preferentially associates with and is phosphorylated by cyclin E/Cdk2 in vitro. Immobilized cyclin/Cdk complexes were incubated with control insect cell lysates or insect cell lysates containing Flag-p220, as described in Materials and Methods. Complexes were washed with lysis buffer followed by 10 mM MgCl₂ and 20 mM Tris-HCl. Some samples were supplemented with γ-[³²P]ATP for 20 min before SDS–PAGE and visualization of proteins by immunoblotting or autoradiography. Flag-p220 was detected by anti-flag antibodies. The quantities of GST–cyclin/Cdk complexes were similar, as determined by immunoblotting with GST antibodies. Cdk2 complexes associated with an insect cell protein migrating slightly faster than p220 that was also a substrate for the kinase (indicated by an asterisk). An anti-flag immunoprecipitate of Flag-p220 (lane 19) was included as a control.

As expected, p220 was readily phosphorylated by the associated cyclin E/Cdk2 complex, and this phosphorylation was accompanied by reduced mobility of p220 (Fig. ​(Fig.4,4, lanes 16 and 17). Although p220 bound weakly to cyclin A/Cdk2, the associated protein underwent a similar mobility shift in the presence of ATP (lanes 7 and 8), suggesting that cyclin A/Cdk2 can also phosphorylate p220 when bound. We note that control reactions employing control insect cell lysates revealed the presence of a cyclin E/Cdk2–associated substrate (indicated by the asterisk) that migrated slightly faster than did p220 (Fig. ​(Fig.4,4, lanes 9 and 18). It can be distinguished from the human p220, because it was also phosphorylated to similar levels by cyclin A/Cdk2.

We next sought to determine the sites of p220 phosphorylation in vitro by using mass spectrometry (Zhang et al. 1998). p220 contains 18 potential Cdk phosphorylation sites (Thr/Ser followed by Pro). Four tryptic p220 phosphopeptides containing five phosphorylation sites were identified in recombinant p220 phosphorylated by associated cyclin E/Cdk2 in vitro (Fig. ​(Fig.5a,b;5a,b; Table ​Table2).2). Three singly phosphorylated peptides were sequenced by liquid chromatography/mass spectrometry/mass spectrometry (LC/MS/MS) to identify the phosphorylation sites as S1100 (site 3), T1270 (site 4), and T1350 (site 5), respectively (Table ​(Table2).2). The sequence of a doubly phosphorylated peptide encompassing residues 742–788 could not be determined because of its large size (m/z 5077.5, average mass; see Fig. ​Fig.5a).5a). However, this peptide contains two consensus Cdk substrates at S775 and S779 (sites 1 and 2) (Table ​(Table2),2), allowing a tentative assignment as sites of modification by cyclin E/Cdk2. Indeed, we found that p220 mutated in both of these serine residues was resistant to a cyclin E/Cdk2–induced shift in mobility (Fig. ​(Fig.5d).5d). In contrast, p220 proteins mutated in one or more of the other identified phosphorylation sites still underwent a mobility shift in response to cyclin E/Cdk2 treatment (Fig. ​(Fig.5d).5d). These data are consistent with the assignment of S775 and S779 as in vitro substrates and indicate that they are primarily responsible for the mobility shift observed upon phosphorylation by cyclin E/Cdk2.

Open in a separate window

Figure 5

Identification of phosphorylation sites in p220. (a,b) A portion of the matrix-assisted laser desorption/ionization mass spectrometry (MALDI/TOF) mass spectra before (upper spectra) and after (lower spectra) treatment with calf intestinal phosphatase (CIP), showing the doubly phosphorylated peptide encompassing the sequence of 742–788 from recombinant (a) and endogenous (b) p220. To generate cyclin E/Cdk2–phosphorylated p220, 2 μg of Flag-p220 immobilized on anti-flag agarose was allowed to associate with 1 μg of cyclin E/Cdk2 and washed complexes incubated with 1 mM ATP (50 min at 25°C). (c) Schematic diagram of p220 phosphorylation sites and the cyclin E binding domain inferred by expression cloning (Zhao et al. 1998; Ma et al. 1999). (d) Altered mobility of p220 in response to cyclin E/Cdk2–mediated phosphorylation requires S775 and S779. In vitro translation products were incubated in the presence or absence of 20 nM cyclin E/Cdk2 for 20 min at 30°C (top) or with 20 and 50 nM cyclin E/Cdk2 for 20 min at 30°C (bottom) before electrophoresis and autoradiography. (e) Specificity of phosphopeptide-specific antibodies. In vitro–translated p220 or p220^ΔCdk (50 μL) was incubated with 1 μg of GST–cyclin E/Cdk2 immobilized on GSH-Sepharose, and washed complexes were incubated in the presence or absence of 1 mM ATP (lanes 1–4). Proteins were separated by SDS–PAGE and immunoblotted. The anti-phosphoThr antibody (New England Biolabs), which reacts with a large number of different phosphoThr-Pro-containing sequences, did not recognize p220^ΔCdk. As a control, Flag-p220 from insect cells (∼100 ng) was incubated with or without 50 nM cyclin E/Cdk2 and 1 mM ATP before immunoblotting. (WT) Wild type. (f) Anti-p220 immune complexes from 293T cells were separated by SDS–PAGE and immunoblotted with the indicated antibodies. (NRS) Normal rabbit serum.

To examine phosphorylation in vivo, we purified p220 from a 293T cell nuclear lysate by using immunoprecipitation (Fig. ​(Fig.1a)1a) and subjected it to mass spectral analysis. p220 is present at low levels in this cell line, and from 44 mg of nuclear extract, we obtained 200 ng of p220. We were able to identify one peptide with the mass expected for a doubly phosphorylated peptide spanning Val742–Lys788 whose quantity was consistent with the level of protein available for analysis (Fig. ​(Fig.5b;5b; Table ​Table2).2). Importantly, this peptide is absent from spectra after treatment with a phosphatase (Fig. ​(Fig.5b),5b), indicating that S775 and S779 are phosphorylated in vivo. Signals for the three singly phosphorylated peptides observed in vitro were not evident, possibly because of the small amounts of material available for analysis.

p220 phosphopeptide antibodies recognize phosphorylated p220 in vitro and in vivo

To investigate T1270 and T1350 phosphorylation in vivo, we generated phospho-specific antibodies against these sites. These antibodies recognize Flag-p220 purified from insect cells only after phosphorylation with cyclin E/Cdk2 (Fig. ​(Fig.5e,5e, lanes 5 and 6). To examine the specificity of these antibodies for reaction with p220, we bound either wild-type p220 or p220^ΔCdk (containing Ala substitutions at S775, S779, S1100, T1270, and T1350) to immobilized cyclin E/Cdk2 and incubated these complexes in the presence or absence of ATP (Fig. ​(Fig.5e).5e). Both antibodies recognized phosphorylated p220 (lane 2) but did not recognize p220^ΔCdk. The two proteins were present at comparable levels based on anti-p220 immunoblotting (lanes 3 and 4). The low levels of reactivity toward the in vitro–translated p220 in the absence of added ATP likely reflect phosphorylation that occurred during immunoprecipitation from ATP-containing reticulocyte lysates, because larger amounts of recombinant p220 purified from insect cells did not react with the phosphopeptide-specific antibodies in the absence of phosphorylation by cyclin E/Cdk2 (lanes 5 and 6). We also note that a general phosphothreonine-proline antibody gave similar results, indicating that other TP sequences in p220^ΔCdk are not phosphorylated by bound cyclin E/Cdk2 in vitro. We found that both phosphopeptide antibodies reacted with p220 immunoprecipitated from cycling 293T cells (Fig. ​(Fig.5f),5f), indicating that these sites are indeed phosphorylated in vivo. On the basis of the comparison with p220 in 293T cells examined in parallel (Fig. ​(Fig.5f),5f), the more slowly migrating p220 protein was preferentially detected by the phosphopeptide antibodies.

CB-associated p220 co-localizes with cyclin E and is phosphorylated on Cdk sites in a cell cycle–dependent manner

The data described thus far suggest that p220 is targeted to CBs and is a substrate of cyclin E/Cdk2. However, it is unclear whether p220 is phosphorylated while in CBs. To examine whether CB-associated p220 is phosphorylated on Cdk2 sites, we performed immunofluorescence by using antibodies against phospho-T1270 and phospho-T1350. Both recognized nuclear foci similar to the antibody against p220 (Fig. ​(Fig.6a;6a; data not shown). To demonstrate that the foci coincided with those found with anti-p220, we performed double immunofluorescence staining with antibodies against phospho-T1270 and coilin in cycling fibroblasts (Fig. ​(Fig.6).6). From 500 cells scored, 30.2% displayed primarily two foci reactive toward both antibodies, whereas 40.4% had primarily four co-localized foci (Fig. ​(Fig.6a),6a), including cells in prophase (Fig. ​(Fig.6b).6b). The remaining 29.4% of the cells lacked staining with the phospho-T1270 antibody. This is in marked contrast to staining with p220 antibodies in which cells lacking antibody reactivity were very rare except for those in metaphase and telophase (Figs. ​(Figs.2a2a and ​and3b).3b). In cells that were negative for reactivity with anti-phospho-T1270, coilin reactivity was still observed (Fig. ​(Fig.6a).6a). Quantification of relative DAPI intensity of 200 additional cells demonstrated that cells nonreactive with the phosphopeptide antibodies had a lower DNA content than did reactive cells, which is consistent with these cells being in the G1 phase, whereas phospho-T1270 antibody-positive cells had a higher DNA content consistent with S- or G2-phase cells (Fig. ​(Fig.2b).2b). In this separate experiment, a somewhat higher percentage of cells had no detectable focus (data not shown). In agreement with the results with p220 antibodies, the few cells in metaphase and telophase were negative for staining with the phosphopeptide antibody, whereas coilin signals appeared dispersed (Fig. ​(Fig.6d,e).6d,e).

Open in a separate window

Figure 6

p220 in Cajal bodies (CBs) is phosphorylated on Cdk sites in a cell cycle–specific manner. Dual detection with anti-phospho-T1270 antibodies and p80^coilin in dermal fibroblasts is shown. Focal co-localization was observed in S phase (a) and prophase (b), but antiphospho-T1270 did not detect any foci in metaphase (c) or telophase (d). a also contains three cells that display anti-coilin reactive foci but not anti-phospho-T1270 reactive foci. The DNA content of these cells is consistent with G1 phase (Fig. ​(Fig.2b).2b). Only diffused coilin signals were observed in c and d. (Green) anti-phospho-T1270; (red) anti-coilin; (blue) 4′,6-diamidino-2-phenylindole (DAPI).

If cyclin E is responsible for phosphorylation of p220 within CBs, one would predict that p220 would co-localize with cyclin E and that the timing of p220 phosphorylation would be coincident with co-localization. To examine this issue directly, we initially performed co-localization experiments using anti–cyclin E and anti-p220 antibodies. As shown in Figure ​Figure7a,7a, cyclin E was concentrated in foci that are coincident with p220 foci. At longer exposure, cyclin E was evident as faint dust throughout the nucleus as well. These results are consistent with the recent report that cyclin E is concentrated in CBs in S phase (Lui et al. 2000).

Open in a separate window

Figure 7

Co-localization of cyclin E with anti-phospho-T1270 reactive foci in a cell cycle–dependent manner. (a) Cyclin E (red) co-localizes with p220 (green) in a cell containing four p220 foci. Nuclei are in blue. (b–e) Growing diploid fibroblasts were labeled with BrdU for 1 h and subjected to immunofluorescence using anti-phospho-T1270 (red), anti-cyclin E (green), anti-BrdU (magenta), and 4′,6-diamidino-2-phenylindole (DAPI; blue) to visualize nuclei. (b) A BrdU-negative cell containing two phospho-T1270 foci that co-localize with cyclin E. This field also contains a BrdU-negative cell that is negative for both anti–cyclin E and anti-phospho-T1270 antibodies. (c) A BrdU-positive cell containing two phospho-T1270 foci that co-localize with cyclin E adjacent to a BrdU-negative cell that is negative for both anti–cyclin E and anti-phospho-T1270 antibodies. (d) BrdU-positive cells containing two or four phospho-T1270 foci that co-localize with cyclin E. (e) A BrdU-negative cell containing four phospho-T1270 foci that lacks cyclin E staining.

To examine whether phosphorylation of p220 correlates with co-localization with cyclin E, we pulse-labeled asynchronous diploid fibroblasts with BrdU and determined the presence of phospho-T1270, cyclin E, BrdU, and DAPI-stained nuclei (Fig. ​(Fig.7b–e).7b–e). Among cells containing two phospho-T1270 foci, both BrdU-positive and BrdU-negative cells were observed; in both cases, however, these foci contained cyclin E (Fig. ​(Fig.7b–d).7b–d). In contrast, among cells containing four anti-phospho-T1270 foci, those that were BrdU positive most frequently displayed co-localized cyclin E, whereas those that were BrdU negative typically lacked cyclin E staining (Fig. ​(Fig.7d,e).7d,e). Given the data presented previously, we believe the latter class of cells to be G2 cells that have lost cyclin E expression but maintain p220 in a phosphorylated form.

The in situ results with antibodies to p220, phospho-T1270, phospho-T1350, coilin, and cyclin E (Figs. ​(Figs.11–3, ​,6;6; Liu et al. 2000) indicate the following: (1) p220 is an in vivo Cdk2 substrate and can be phosphorylated on cyclin E/Cdk2 sites while present in CBs. There is a tight correlation between the appearance of cyclin E in foci and the occurrence of p220 phosphorylation such that (2) cells that are nonreactive with antibody to phospho-T1270 are in the early G1 phase before cyclin E/Cdk2 is present to phosphorylate p220 in CBs. Consistent with this, cells that lacked anti-phospho-T1270 foci also lacked detectable cyclin E. (3) Cells containing two anti-phospho-T1270 reactive foci that co-localize with coilin and cyclin E are in late G1 or early S phases when cyclin E/Cdk2 levels peak. (4) Cells that have four co-localized foci are well into S phase, and p220 remains phosphorylated (Figs. ​(Figs.11 and ​and2).2). Although the pattern of p220 phosphorylation persists into prophase, cyclin E staining is lost at some point in late S phase or G2 phase, as exemplified by the presence of BrdU-negative cells containing four phospho-p220 foci but lacking cyclin E co-localization. (5) Around the time of metaphase and later, p220 foci are not detected with either p220 or phospho-T1270 antibodies.

Mutation of cyclin E/Cdk2 sites in p220 reduces p220-mediated histone H2B promoter activation

Our data suggest that cyclin E/Cdk2 may regulate p220 during the G1/S-phase transition. To examine this question, we took advantage of the recent finding that p220 expression in tissue culture cells leads to increased expression of histone 2B (H2B) promoter– and histone 4 (H4) promoter–luciferase reporter constructs, independent of its effects on S-phase acceleration (Zhao et al. 2000). Histone gene expression is complex, involving both message stabilization (approximately sevenfold) and transcriptional activation (approximtely fivefold), which together account for an ∼35-fold increase in histone mRNA levels during S phase (Harris et al. 1991; Heintz 1991). But the process by which the histone transcriptional apparatus senses cell cycle position is unknown.

We compared the ability of a vector expressing p220 (pCMV-p220) or p220^ΔCdk (pCMV-p220^ΔCdk) to activate luciferase expression from an H2B (−200/0) promoter–luciferase reporter plasmid in transiently transfected 293T cells (Fig. ​(Fig.8a).8a). The level of induction by p220 relative to control transfections ranged from two- to 10-fold, depending on the quantity of p220 plasmid used and the level of p220 expression achieved, as determined by immunoblotting (Fig. ​(Fig.8b).8b). In contrast, p220^ΔCdk displayed a substantially reduced ability to activate the H2B reporter construct when expressed at comparable levels (Fig. ​(Fig.8a–e).8a–e). Similar results were obtained with a minimal H2B promoter (−127/−27) (data not shown). At low levels of expression, p220^ΔCdk displayed levels of H2B reporter activation comparable to control transfected cells (Fig. ​(Fig.8a,b);8a,b); however, at higher levels of expression, a twofold increase in reporter activity over control transfected cells was typically observed (Fig. ​(Fig.8c,d).8c,d). Transfected cells displayed p220 foci as well as diffuse signals throughout the nucleoplasm because of elevated levels of expression from the transfected plasmids (Fig. ​(Fig.8e).8e). Taken together, these results suggest that phosphorylation at one or more cyclin E/Cdk2 sites contributes substantially to this aspect of p220 function. However, it is possible that elevated levels of p220 can partially bypass a requirement for phosphorylation at these sites. In these experiments, p220 appeared to be primarily in a more slowly migrating phosphorylated form, while p220^ΔCdk remained in a more rapidly migrating form (Fig. ​(Fig.8b,d).8b,d). Thus, it appeared that sufficient cyclin E/Cdk2 existed in these cells to phosphorylate fully the transiently expressed p220. This may explain why co-expression of cyclin E/Cdk2 had little effect on the levels of H2B reporter activity in 293T cells (data not shown).

Open in a separate window

Figure 8

Cyclin E/Cdk2–mediated phosphorylation of p220 is important for optimal p220-induced histone H2B transcriptional activation. (a) pCMV, pCMV-p220, or pCMV-p220^ΔCdk was transfected into 293T cells (3.5-cm dish), along with 50 ng of pCMV-β-galactosidase and 50 ng of pGL-H2B-luciferase reporter plasmid, and subsequently processes for luciferase and β-galactosidase activity as described in Materials and Methods. Luciferase activities were normalized relative to β-galactosidase activities. (b) A portion of extracts used in a was immunoblotted using anti-p220 antibodies. The wild-type p220 protein comigrates with the endogenous protein found at low levels, whereas p220^ΔCdk migrates slightly faster as a result of the absence of phosphorylation. (c) The results of two independent experiments each performed using triplicate independent calcium phosphate precipitates derived from 10 μg of pCMV, pCMV-p220, or pCMV-p220^ΔCdk are shown. (−) Lacking p220; (WT) wild type. (d) and (e) Expression levels for p220 and p220^ΔCdk for experiments shown in c were similar as determined by immunoblotting (d) or immunofluorescence (e). (f) Expression of p57^KIP2 blocks activation of the H2B promoter by p220 overexpression. 293T cells were transfected with limiting amounts of pCMV-p220 or p200^ΔCdk expressing plasmids (5 μg) in the presence or absence of either 1 or 2 μg of pCMV-p57^KIP2. At this level of expression, p220 leads to a twofold increase in reporter activity. p57^KIP2 expression leads to a dramatic reduction in the level of p220-induced reporter activity in the presence or absence of p220 expression. The averages of duplicate independent transfections are shown.

Consistent with a role for cyclin E/Cdk2–mediated phosphorylation in p220 function, we found that co-expression of p57^KIP2, which can inhibit Cdk2 activity and block cells at the G1/S-phase transition (Matsuoka et al. 1995), led to a reduction in the ability of p220 to activate H2B promoter activity in transiently transfected cells (Fig. ​(Fig.8f).8f). In this experiment, the levels of p220 expression plasmid used were such that a twofold increase in H2B promoter activity was observed, but p57^KIP2 reduced luciferase levels below that obtained with control transfected cells. As expected, expression of p57^KIP2 alone also reduced the levels of promoter activity in the absence of p220 expression (Fig. ​(Fig.8f).8f). This repression likely reflects the fact that cells are blocked in G1 with low cyclin E/Cdk2 activity. Immunoblotting demonstrated comparable levels of expression of p220 and p220^ΔCdk and verified the expression of p57^KIP2 (data not shown).

Antibodies and immunofluorescence assays

Bacterial GST-p220 (residues 1054–1397) was used to generate antibodies in rabbits. Antibodies were depleted of reactivity to the GST protein and affinity-purified using immobilized GST-p220. Anti-coilin monoclonal antibodies (π-isotype) were provided by M. Carmo-Fonseca (University of Lisbon, Portugal; Almeida et al. 1998). Anti-cyclin E (HE12) came from Pharmingen. Antibodies against Thr-1270 (Asp-Leu-Pro-Val-Pro-Arg-phosphoThr-Pro-Gly-Ser-Gly-Ala-Gly-Cys) and Thr-1350 (Ser-Arg-Thr-Thr-Ser-Ala-phosphoThr-Pro-Leu-Lys-Asp-Asn-Thr-Cys) were generated in rabbits after coupling to keyhole limpet hemocyanin. For immunofluorescence, cells were fixed in either ethanol or formalin and permeabilized with 0.1% Triton X-100. Detection of rabbit antibodies was accomplished using secondary antibodies labeled with Cy3 or Alexa 488 (Molecular Probes, Eugene, OR) or with anti-rabbit horseradish peroxidase (HRP) (Roche, Indianapolis, IN) and tyramides (NEN, Boston, MA) (see below). BrdU was detected using fluorescein isothiocyanate (FITC)–conjugated anti-BrdU antibody. Nuclear DNA was revealed with DAPI staining. Microscopic analysis was performed using either an Olympus BX-60 fitted with an Optronics CCD camera (Baylor College of Medicine) or an AX-70 Olympus microscope and an Olymix digital camera (University of Alabama). Images were captured with a 100× objective lens by using either multiband pass filters or single pass filters with merging using Adobe Photoshop.

Correlation of p220 foci or phospho-T1270 foci with DNA content

Asynchronous human dermal fibroblasts were fixed with 4% paraformaldehyde, permeabilized with 0.5% Triton X-100 in phosphate-buffered saline (PBS), and then treated with 3% hydrogen peroxide. Slides were then incubated with a 1:100 dilution of anti-p220 or anti-phosphopeptide for 2 h at 37°C and washed three times in PBS containing 0.1% Tween 20. After incubation with anti-rabbit HRP (1:100), cells were washed three times, and the signals were developed using a 1:100 dilution of cyanine-3-tyramide (NEN). To detect BrdU, the cells were fixed once more with 4% paraformaldehyde, treated with 3 N HCl for 15 min, incubated with an anti-BrdU FITC antibody for 2 h at 37°C, DAPI (Sigma) stained, and mounted with antifade.

Relative nuclear DNA content was determined as follows: After staining with various antibodies, we took images of DAPI and of the p220 or phosphopeptide foci separately, but all images came from the same slide to avoid any variability in DAPI staining. Using Image Pro Plus software (Media Cybernetics, Silver Spring, MD), we identified nuclei to be measured and determined the average DAPI density individually. The average background for each image was then determined and subtracted, giving the corrected average nuclear DAPI density for each cell. Cells were then individually correlated to the foci number. For graphic representation (Fig. ​(Fig.2a,b),2a,b), the average DAPI density of all two-foci nuclei (the majority of which were BrdU negative; see Figs. ​Figs.11 and ​and2c)2c) was used to divide the density of each individual cell. This normalization procedure yielded a number between 1 (G1) and 2 (G2) and somewhere in between (S).

Interphase chromosome painting

Asynchronous primary human dermal fibroblasts were fixed and stained with antibodies to p220, and signals were developed with tyramide as described above. Slides were fixed again with 4% paraformaldehyde and treated with RNase A (100 μg/mL) in 2 × SSC (1 × SSC is 0.15 M NaCl and 0.015 M sodium citrate) for 1 h at 37°C. They were then ethanol-dehydrated and denatured in 70% formamide, 2 × SSC, pH 7.0, for 2 min at 72°C, dehydrated in ethanol, and hybridized individually with chromosome paints overnight at 37°C. The chromosome 1 and chromosome Y paints were a direct FITC conjugate or cyanine 3 conjugate, respectively, from Vysis (Downers Grove, IL). Signals were detected according to the manufacturer's protocol. Paints for chromosome 5, 6, or 17 were digoxigenin probes from Oncor (Gaithersburg, MD). They were detected using an anti-dig Texas Red antibody. The slides were DAPI stained before imaging.

Immunoprecipitation and phosphorylation

For nuclear extracts, 293T cells were lysed in 50 mM Tris-HCl, pH 7.5, containing 1 mM EDTA, 100 mM NaCl, 0.3% Nonidet P-40, and protease/phosphatase inhibitors, and nuclei pelleted by centrifugation. Nuclear proteins were solubilized using RIPA buffer (50 mM Tris-HCl at pH 7.5, 150 mM NaCl, 0.5% deoxycholate, 1% Nonidet P-40, 0.1% SDS, and protease/phosphatase inhibitors) before centrifugation. Extracts were precleared with protein A–bound normal rabbit IgG before incubation with affinity-purified anti-p220 antibodies bound to protein A–Sepharose. Beads were washed with RIPA buffer. Proteins were separated by SDS–PAGE before staining with Coomassie Brilliant Blue R 250 or transferred to nitrocellulose filters for immunoblotting with anti-p220 antibodies. Detection was accomplished using enhanced chemiluminescence (Amersham).

For expression of p220 in insect cells, the coding sequence (Imai et al. 1996) was cloned into pUNI-50, and in vitro recombination was then used to generate a pVL1392-based plasmid with p220 fused at its N terminus to the Flag tag as described (Liu et al. 1998). Viruses were made using Baculogold (Pharmingen). To examine interaction with Cdks, cyclin/Cdk complexes were purified with glutathione-Sepharose beads (Ma et al. 1999). Immobilized complexes were then incubated with insect cell extracts with or without Flag-p220. Complexes were washed with lysis buffer and with 20 mM Tris-HCL, pH 7.5, and 10 mM MgCl₂. A portion of each mixture was used for kinase assays employing 50 μM γ-[³²P]ATP. Samples were separated by SDS–PAGE and transferred to nitrocellulose before immunoblotting and autoradiography.

Matrix-assisted laser desorption/ionization mass spectrometry (MALDI/TOF) with delayed extraction (Voyager-DE, Perseptive Biosystems, Framingham, MA) was used for the identification of phosphopeptides, as described by Zhang et al. (1998). An electrospray ion trap mass spectrometer (LCQ, Finnigan, San Jose, CA) coupled on-line with a capillary high-pressure liquid chromatograph (Magic 2002, Auburn, CA) was used for identification of phosphorylation sites. A MAGICMS C18 column (5 μm particle diameter, 150 Å pore size, 0.1 × 50 mm dimension) was used for the LC/MS/MS analysis.

We thank J. Gall for discussions, J. Zhao, A.G. Matera, and E. Harlow for communicating results before publication, M. Carmo-Fonseca for anti-coilin antibodies, and Richard Atkinson, Brian Streib, and Heather Benedict-Hamilton for assistance. J.W.H. was supported by U.S. Public Health Service (USPHS) grant GM54137 and by the Welch Foundation. L.T.C. was supported by USPHS grants CA36200 and DE/CA11910. B.A.V.T was partially supported by the University of Alabama Medical Scientist Training Program. The Digital Imaging Microscopy Facility at the University of Alabama was supported by the University of Alabama Health Services Foundation and by grant DE/CA11901.

The publication costs of this article were defrayed in part by payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 USC section 1734 solely to indicate this fact.