Modulation of Human DNA Topoisomerase II a Function by Interaction with 14-3-3 e *

, Human DNA topoisomerase II a (topo II), a ubiquitous nuclear enzyme, is essential for normal and neoplastic cellular proliferation and survival. Several common an-ticancer drugs exert their cytotoxic effects through interaction with topo II. In experimental systems, altered topo II expression has been associated with the appearance of drug resistance. This mechanism, however, does not adequately account for clinical cases of resistance to topo II-directed drugs. Modulation by protein-protein interactions represents one mechanism of topo II regulation that has not been extensively defined. Our laboratory has identified 14-3-3 e as a topo II-interacting protein. In this study, glutathione S -transferase co-pre-cipitation, affinity column chromatography, and immu-noprecipitations confirm the authenticity of these interactions. Three assays evaluate the impact of 14-3-3 e on distinct topo II functional properties. Using both a modified alkaline comet assay and a DNA cleavage assay, we demonstrate that 14-3-3 e negatively affects the ability of the chemotherapeutic, etoposide, to trap topo II in cleavable complexes with DNA, thereby preventing DNA strand breaks. By electrophoretic mobility shift assay, this appears to be due to reduced DNA binding activity. The association of topo II with 14-3-3 proteins does not extend to all 14-3-3 isoforms. No protein interaction or disruption of topo II function was cDNA Expression Library Screening for TIPs— To identify unknown protein interactive partners of topo II, approximately 1 3 10 6 plaques of a HeLa cDNA library in l gt11 were plated at a density of 20,000 plaque-forming units/150-mm plate. Protein induction was accomplished by overlaying plates with 132-mm nitro- cellulose filters in 10 m M isopropylthio- b -galactoside. Filter lifts were subjected to a protein denaturing/renaturing protocol blocked and probed as described previously but using 125 I-human topo II (amino 857–1448) as the interaction probe. Expression and puri- fication of this recombinant protein have been detailed elsewhere prior to autoradiography. Putative positive clones were plated for two additional cycles of screening until plaque purification was achieved. Selected cDNA clones were obtained by polymerase chain reaction using l gt11 primers and the PCR prod-ucts cloned in the T-tailed sequencing vector, pCRII (Invitro- gen, Carlsbad, CA). One TIP clone was identical to the DNA sequence to amino acids 126–255 (terminus) of the epsilon form of human 14-3-3 protein. Full-length 14-3-3 e cDNA was obtained by polymerase chain reaction using a dilution of HeLa cDNA library as the template and was subcloned in frame with glutathione S -transferase (GST) in pGEX-2TK (Amersham Pharmacia Biotech) to enable prokary- otic expression and affinity purification of the resulting chimeric protein (GST14-3-3 e ). GST Co-precipitation Assays— The co-precipitation method of Conklin et al. (38) was employed. Briefly, 10–200 m g of purified GST14- 3-3 e peptides or GST alone was incubated with 20 m g of topo II peptide (amino acids 857–1448) in PBS containing 0.5% Triton X-100 and 0.1% BSA for 16 h at 4 °C. In experiments where interactions of full-length proteins were investigated, CCRF/CEM nuclear extract was used as the topo II source. Portions of the nuclear extract were treated with calf intestinal alkaline phosphatase (Life Technologies, Inc.) (1 unit/6 m g of extract) prior to incubation with GST fusion proteins in a buffer containing 40 m M Hepes, 100 m M NaCl, 0.5 m M EDTA, 5 m M MgCl 2 , and 0.05% Nonidet P-40. Complexes were subsequently precipitated by the addition of 10 m l of a 50% slurry of glutathione-Sepharose (Amersham Pharmacia Biotech) and incubated on ice on a rotating shaker. Beads were allowed to settle, washed three times in binding buffer, then heated to 95 °C in 2 3 Laemmli SDS sample buffer. Following SDS-PAGE and electrotransfer, blots were probed with a polyclonal anti- serum to human topo II a . Co-immunoprecipitation and Affinity Chromatography— CCRF-CEM extracts were prepared by sonication in radioimmunoprecipitation assay buffer (RIPA: PBS 1 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS with Complete y protease inhibitor (Roche Molecular Bio-chemicals), 1 m M phenylmethylsulfonyl

tems, changes in the expression of topo II have been associated with the appearance of antitumor drug resistance (4 -6). This mechanism, however, is unable to account for clinical cases of resistance to topo II-directed drugs, implying that alternative topo II regulatory mechanisms require consideration (7,8).
Modulation by protein-protein interactions represents one mechanism of topo II regulation that has been largely underappreciated. Interaction of the transcription factors CREB, ATF-2, and c-Jun with topo II stimulates topo II DNA decatenation activity in vitro (9), while association of wild-type retinoblastoma protein with topo II decreases this DNA decatenation activity (10). Yeast topo II, in addition to associating with nuclear scaffold proteins and proteins involved in chromosomal segregation (11)(12)(13), has been demonstrated to interact functionally with the Sgs1 helicase (14). Cells lacking the Sgs1 protein (Sgs1p) exhibit a slow growth phenotype due to the inability to completely decatenate DNA. The finding that Sgs1p is required in concert with topo II for completely faithful chromosomal segregation now appears to be significant in human neoplasia. A defect in the human homolog of Sgs1p, BLM, appears to be the genetic basis of Bloom's syndrome, a human disease of genetic instability (15). Clearly, continued characterization of topo II protein-protein interactions will yield new and important roles for the enzyme and may reveal implications for these interactions in cellular sensitivity to topo II-directed chemotherapeutics.
Previous work in the laboratory led to the isolation of several topo II-interacting proteins (TIPs) (16). We have now identified one of these as the epsilon isoform of the human 14-3-3 protein (14-3-3⑀) and independently isolated a 14-3-3⑀ cDNA from screening a human HeLa cell cDNA expression library with a radiolabeled, 80-kDa C-terminal fragment of human topo II protein. The 14-3-3 proteins belong to a well conserved family of seven isoforms (17). In addition to binding cruciform DNA structures (18), 14-3-3 proteins have been characterized as preferentially binding serine phosphorylated proteins and are implicated in the regulation of signal transduction pathways and cell cycle-associated proteins (reviewed in Refs. 17 and 19). Among their reported functions, 14-3-3 proteins have been demonstrated to: (a) bind the protein kinases Raf1 (20 -22), protein kinase C (23,24), and BCR (25,26) modulating their activities; (b) serve an apoptosis checkpoint function by binding to the apoptotic agonist BAD, thereby preventing the association of BAD with Bcl-2 and Bcl-X L (27,28); (c) interact with p53, thereby enhancing p53 DNA binding activity (29) and modulating G 2 /M progression (30,31); and (d) modulate the intracellular localization of cdc25 (32)(33)(34)(35) and protein kinase U␣ (36) during phases of the cell cycle and in response to DNA damage. This present study characterizes the interaction between topo II and 14-3-3⑀ and presents evidence suggesting that this association modulates topo II-DNA interactions and drug-stabilized cleavable complex formation.

EXPERIMENTAL PROCEDURES
Mammalian Cell Culture-HeLa human cervical carcinoma cells, HL-60 human promyelocytic leukemia cells, and CCRF/CEM human acute lymphoblastic leukemia cells were obtained from the American Type Culture Collection (Manassas, VA). HeLa cells were maintained as monolayer cultures in Dulbecco's modified Eagle's medium (Life Technologies, Inc.), and HL-60 and CCRF/CEM cells were maintained as suspension cultures in RPMI 1640 medium (Life Technologies, Inc.). Both media were supplemented with 10% fetal bovine serum, 50 units/ml penicillin G, and 50 g/ml streptomycin sulfate. All cells were maintained at 37°C in a humidified atmosphere containing 5% CO 2 .
cDNA Expression Library Screening for TIPs-To identify unknown protein interactive partners of topo II, approximately 1 ϫ 10 6 plaques of a HeLa cDNA library in gt11 (CLONTECH, Palo Alto, CA) were plated at a density of 20,000 plaque-forming units/150-mm plate. Protein induction was accomplished by overlaying plates with 132-mm nitrocellulose filters soaked in 10 mM isopropylthio-␤-galactoside. Filter lifts were subjected to a protein denaturing/renaturing protocol (37), blocked and probed as described previously (9), but using 125 I-human topo II (amino acids 857-1448) as the interaction probe. Expression and purification of this recombinant protein have been detailed elsewhere (9). Filters were washed extensively prior to autoradiography. Putative positive clones were plated for two additional cycles of screening until plaque purification was achieved. Selected cDNA clones were obtained by polymerase chain reaction using gt11 primers and the PCR products directly cloned in the T-tailed sequencing vector, pCRII (Invitrogen, Carlsbad, CA). One TIP clone was identical to the DNA sequence corresponding to amino acids 126 -255 (terminus) of the epsilon form of human 14-3-3 protein. Full-length 14-3-3⑀ cDNA was obtained by polymerase chain reaction using a dilution of HeLa cDNA library as the template and was subcloned in frame with glutathione S-transferase (GST) in pGEX-2TK (Amersham Pharmacia Biotech) to enable prokaryotic expression and affinity purification of the resulting chimeric protein (GST14-3-3⑀).
GST Co-precipitation Assays-The co-precipitation method of Conklin et al. (38) was employed. Briefly, 10 -200 g of purified GST14-3-3⑀ peptides or GST alone was incubated with 20 g of topo II peptide (amino acids 857-1448) in PBS containing 0.5% Triton X-100 and 0.1% BSA for 16 h at 4°C. In experiments where interactions of full-length proteins were investigated, CCRF/CEM nuclear extract was used as the topo II source. Portions of the nuclear extract were treated with calf intestinal alkaline phosphatase (Life Technologies, Inc.) (1 unit/6 g of extract) prior to incubation with GST fusion proteins in a buffer containing 40 mM Hepes, 100 mM NaCl, 0.5 mM EDTA, 5 mM MgCl 2 , and 0.05% Nonidet P-40. Complexes were subsequently precipitated by the addition of 10 l of a 50% slurry of glutathione-Sepharose (Amersham Pharmacia Biotech) and incubated on ice on a rotating shaker. Beads were allowed to settle, washed three times in binding buffer, then heated to 95°C in 2ϫ Laemmli SDS sample buffer. Following SDS-PAGE and electrotransfer, blots were probed with a polyclonal antiserum to human topo II␣.
Co-immunoprecipitation and Affinity Chromatography-CCRF-CEM extracts were prepared by sonication in radioimmunoprecipitation assay buffer (RIPA: PBS ϩ 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS with Complete protease inhibitor (Roche Molecular Biochemicals), 1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, and the phosphatase inhibitors 10 mM NaF and 10 mM Na 3 VO 4 ). Extracts (25-500 g of total protein) were brought up to 500 l of RIPA and pre-cleared with 20 l of a 50% slurry of Protein A/Protein Gagarose beads with rocking for 1 h on ice. Cleared extracts were then incubated with 5 l of polyclonal antiserum (Santa Cruz Biotechnology, Santa Cruz, CA) to 14-3-3⑀ (T-16), a broad spectrum 14-3-3 antiserum that recognizes all mammalian isoforms (K-19), or a c-Myb monoclonal antibody as a control. Following a 2-h incubation on ice, complexes were precipitated by a 30-min incubation with another 20 l of Protein A/Protein G-agarose. Complexes were washed three times with 500 l of RIPA, then released by boiling for 5 min in 2ϫ Laemmli SDS sample buffer, followed by immunoblotting for topo II␣.
For affinity chromatography, the large C-terminal fragment of topo II (amino acids 857-1448) was expressed in Escherichia coli as a polyhistidine fusion protein as described previously (9) and purified by nickel-agarose affinity chromatography. One aliquot of purified protein (250 g) was phosphorylated with 100 units of recombinant, human casein kinase II (New England Biolabs, Beverly, MA) in 7 ml of 20 mM Tris-HCl, pH 7.5, 50 mM KCl, 10 mM MgCl 2 , 0.2 mM ATP for 30 min at 30°C. Control topo II protein was incubated under similar conditions but without the addition of kinase. Separate 400-l bed volumes of fresh nickel-agarose (Probond) (Invitrogen, Carlsbad, CA) were loaded into individual minicolumns and bound with 250 g of BSA, control topo II protein, or casein kinase II-phosphorylated topo II protein.
Columns were then washed with 10 bed volumes of 20 mM sodium phosphate, pH 7.8, 500 mM NaCl. A RIPA extract of CCRF-CEM cells (5 mg total protein) was then applied to the column, diluted 1:7 in PBS, and reapplied to the column three more times. The columns were washed extensively with 40 bed volumes of PBS, and bound proteins were eluted with multiple 500-l aliquots of 0.5% SDS in 12.5 mM Tris-HCl, pH 6.8. Aliquots of each fraction were supplemented with 2-mercaptoethanol, boiled, and resolved on a 12% SDS-PAGE prior to immunoblotting with a broad spectrum 14-3-3 antiserum (K- 19).
Alkaline Naked Comet Assay-A modified alkaline comet assay using isolated HL-60 nuclei was performed to assess the effect of 14-3-3⑀ on etoposide-mediated DNA damage (39). HL-60 nuclei were prepared from logarithmically growing cultures by standard techniques and incubated in PBS in the presence or absence of 30 M etoposide and GST, GST14-3-3⑀, or GST14-3-3 protein (0.16 -20 g/ml). Nuclei were incubated on ice for 90 min and then subjected to a standard electrophoretic comet protocol (40). After staining with ethidium bromide (20 g/ml), nuclei were quantified by fluorescence microscopy for total nuclear and tail length.
Electrophoretic Mobility Shift Assay (EMSA)-Oligonucleotides corresponding to residues 87-126 of pBR322, a strong topo II binding site (41), were annealed and end-labeled with [␣-32 P]dCTP. Nuclear protein extracts (500 mM NaCl extraction) were prepared from logarithmically growing HeLa cells as described previously (16) and using Complete protease inhibitor (Roche Molecular Biochemicals). Incubations of the binding site with nuclear protein extract were carried out on ice in a 25-l reaction volume containing 50 mM KCl, 20 mM Tris (pH 7.6), 2 mM MgCl 2 , 1 mM EDTA, 10% glycerol, 2.5 g of BSA, and 0.1 g of poly(dI-dC)⅐(dI-dC) (Amersham Pharmacia Biotech). Free and bound oligonucleotide were separated by electrophoresis through a 4% nondenaturing polyacrylamide gel in 0.25ϫ Tris borate-EDTA buffer. For competition EMSA, molar excesses of unlabeled competitor oligonucleotide were premixed with the 32 P-labeled binding site oligonucleotide before the addition of reaction buffer and nuclear protein extract. For supershift EMSA, antiserum was added 30 min after the initiation of the binding reaction and the mixture was incubated on ice for an additional 30 min prior to electrophoresis.
DNA Cleavage Assay-DNA cleavage assays were carried out essentially as described in Ref. 42. Typically, reactions contained 300 ng of pBR322 and were initiated by the addition of 6 units of purified human topo II␣ (Topogen, Columbus, OH) and incubated at 37°C for 8 min. Reactions were terminated by the addition of SDS (0.5% final concentration) and incubated with proteinase K (200 g/ml) at 56°C for 1 h. Reaction components were resolved at 20 V for 16 h on a 1.3% agarose gel containing 0.7 g/ml ethidium bromide to allow the relaxed closedcircular plasmid to run with greater mobility than supercoiled DNA so as to distinguish it from open, nicked-circular DNA (42).

14-3-3⑀ is a Topo II-interacting Protein-
To clone and identify TIPs, a gt11-HeLa cell cDNA library was screened with radioiodinated topo II␣-(857-1448). Four open reading frames were identified that survived three rounds of screening and plaque purification. Three clones corresponded to amino acids 280 -623 of XPR1/X3, a membrane-bound receptor for xenotropic and polytropic murine leukemia viruses (43,44), and are not considered further here. The fourth open reading frame clone was 100% identical to the DNA sequence encoding amino acids 126 -255 of the epsilon isoform of human 14-3-3 protein.
14-3-3⑀ Interacts with Topo II in Solution-To verify the authenticity of the 14-3-3⑀/topo II interaction, expression vectors encoding GST in frame with the N terminus of 14-3-3⑀ (or fragment thereof) were generated and tested for the ability to co-precipitate topo II from solution (Fig. 1). Both the full-length GST14-3-3⑀ and a C-terminal fragment (amino acids 126 -255) interacted with topo II-(857-1448), while GST alone did not precipitate any of the recombinant topo II peptide (Fig. 1A). Conversely, the topo II-(857-1448) fragment, containing a polyhistidine tag, was immobilized on a nickel-chelating column and was tested for the ability to adsorb endogenous 14-3-3 proteins using CCRF-CEM RIPA extract as the source. When compared with a BSA control column, the topo II column selectively co-purified a protein of 30 -32 kDa that reacted with a broad spectrum antiserum to all isoforms (Fig. 1B). Phosphorylation of the topo II-(857-1448) fragment with casein kinase II prior to immobilization on the nickel-chelating column greatly enhanced the co-purification of the 14-3-3-immunoreactive 30 -32-kDa protein (Fig. 1B), relative to either BSA or unphosphorylated topo II-(857-1448).
As the above experiments were carried out using a recombinant topo II peptide, we examined whether GST14-3-3⑀ was able to interact with full-length 170-kDa topo II from nuclear protein extracts. Using nuclear protein extracts from CCRF/ CEM cells as a source of 170-kDa topo II, we determined that GST14-3-3⑀, but not GST, could interact with full-length topo II (Fig. 1C). This association of topo II with 14-3-3 proteins does not extend to all isoforms, as no interaction was observed with a GST14-3-3 fusion protein even though comparable amounts of protein were used (Fig. 1C). Using an alkaline phosphatase treatment previously demonstrated to remove more than 80% of the phospholinkages on topo II (45), we demonstrated that, while the GST14-3-3⑀ fusion protein preferentially interacts with a phosphorylated form of topo II (Fig. 1C), dephosphoryl-FIG. 1. Human topo II interacts in solution with human 14-3-3⑀. Panel A, interaction between immobilized GST14-3-3⑀ proteins and purified C-terminal topo II peptide. Increasing amount of GST, GST14-3-3⑀FL (full-length human 14-3-3⑀), or GST14-3-3⑀CT (14-3-3⑀ C terminus, amino acids 126 -255) were combined with 20 g of topo II-(857-1448) prior to incubation with GSH-Sepharose beads. Proteins liberated from the beads by boiling were resolved by SDS-PAGE (9% acrylamide) and processed for immunoblotting with anti-topo II␣ antiserum. Panel B, affinity column. A 400-l bed volume of nickel-chelating resin (Probond, Invitrogen) was used to bind BSA, polyhistidine-tagged topo II-(857-1448), or casein kinase II-phosphorylated topo II-(857-1448). CCRF-CEM RIPA extracts corresponding to 5 mg of total protein were applied four times to each column, then washed with 40 bed volumes of PBS. Specifically bound proteins were eluted with multiple 500-l aliquots of 0.5% SDS in 12.5 mM Tris-HCl, pH 6.8 and the fractions concentrated 10-fold under vacuum prior to electrophoresis on a 12% SDS-PAGE gel and immunoblotting with a pan-14-3-3 antiserum. The arrow denotes the primary, topo II-specific band of 14-3-3 reactivity migrating at 30 -32 kDa. Panel C, interaction between immobilized GST14-3-3⑀ and full-length topo II␣ from CCRF/CEM nuclear extract. Approximate molar equivalents of GST (50 g), GST14-3-3⑀ (100 g), or GST14-3-3 (100 g) were incubated with a nuclear extract of log phase CCRF/CEM cells (125 or 250 g of protein) prior to incubation with GSH-Sepharose beads. Where indicated, the nuclear extract was treated with calf intestinal alkaline phosphatase (1 unit/6 g of extract) prior to co-precipitation. Samples were boiled, electrophoresed, and processed for topo II detection as in panel A. Panel D, coimmunoprecipitation of topo II with 14-3-3 antisera. The indicated amounts of CCRF-CEM RIPA cell extracts were incubated for 2 h on ice with 5 l of antisera raised to recognize either all 14-3-3 isoforms (K- 19) or the ⑀ form only (T-16). Following immobilization on Protein A/Protein G-agarose, complexes were washed under high stringency with RIPA and immunoprecipitated complexes were liberated by the addition of 2ϫ Laemmli SDS-PAGE sample buffer and boiling. Complexes were resolved on a 7.5% polyacrylamide-SDS gel and immunoblotted as in panel A. ation abrogates the interaction.
Finally, we investigated the ability of 14-3-3 antisera to co-immunoprecipitate 170-kDa topo II from CCRF-CEM cell extracts, using standard methods but with the addition of high stringency washing conditions with RIPA (instead of the low salt, detergent-lacking washes used commonly) (Fig. 1D). Using antisera recognizing either all 14-3-3 isoforms or only the ⑀ isoform, a topo II immunoreactive band in immunoprecipitates migrating at 170 kDa was observed when using 500 g of cell extract. The authenticity of this interaction is emphasized by the fact that this experiment was performed without overexpression of either protein, but rather using the endogenous levels of each protein normally present in CCRF-CEM cells. Despite the fact that 14-3-3 is capable of binding several other proteins in mammalian cells, the interaction with topo II was of enough abundance to be detected in this assay.
14-3-3⑀ Abrogates Topo II DNA Binding Activity-To evaluate the effect of 14-3-3⑀ on topo II DNA binding activity, EMSAs were carried out using a radiolabeled double-stranded oligonucleotide containing a strong topo II binding site. Competition EMSA, using increasing molar excesses of unlabeled oligonucleotide representing the topo II binding site and an oligonucleotide of unrelated sequence, demonstrated that the predominant observed DNA-protein interaction is sequencespecific (Fig. 3A). Addition of an anti-topo II␣ antiserum produced a supershifted complex, indicating the involvement of topo II in the protein-DNA complex (Fig. 3B). Incubation of the nuclear extract with a nonspecific antiserum produced neither a supershifted complex nor a reduction in the intensity of the specific DNA-protein complex (Fig. 3B). GST14-3-3⑀ was added to the binding reaction to evaluate its effect on topo II DNA binding activity. Addition of increasing amounts of GST14-3-3⑀ abrogated topo II DNA binding activity (Fig. 3C). This effect was specific for 14-3-3⑀, as the addition of increasing amounts of GST alone or GST14-3-3 had no effect on the topo II-DNA interaction (Fig. 3C).
14-3-3⑀ Decreases Etoposide-stabilized Cleavable Complex Formation-To examine whether 14-3-3⑀ would alter the ability of etoposide to trap topo II in cleavable complexes, a plasmid DNA cleavage assay was performed (Fig. 4). This assay, designed to evaluate the catalytic activity of topo II, demonstrates that 14-3-3⑀ negatively affects the ability of etoposide to stabilize topo II in covalent complexes with DNA, as demonstrated by the reduction in intensity of the linear DNA band (Fig. 4). The inhibitory effect observed was specific for GST14-3-3⑀, as neither GST alone nor GST14-3-3 had an effect on the cleavage-religation equilibrium. DISCUSSION Through a HeLa cDNA expression library screen, we identified 14-3-3⑀ as a topo II-interacting protein. We have verified the authenticity of this interaction by demonstrating that (a) recombinant GST14-3-3⑀ is able to interact in solution with either a recombinant topo II peptide (857-1448) or full-length topo II in nuclear protein extracts, (b) recombinant topo II-(857-1448) immobilized on a column interacts with 14-3-3 proteins in a whole cell extract, and (c) two antisera recognizing 14-3-3 proteins are able to precipitate full length topo II from cellular extract without the need for overexpression of either protein. Consistent with reports of the preferential binding of 14-3-3 to phosphorylated proteins, the interaction between 14-3-3⑀ and topo II is significantly enhanced by prior phosphorylation of a recombinant topo II fragment and attenuated by phosphatase pretreatment of the nuclear protein extract.
It is well established that topo II, both in vitro and in vivo, is associated with and phosphorylated by casein kinase II (46 -48), although the precise residues phosphorylated in vivo remain to be defined. In addition to phosphorylating topo II, Equilibrium reactions were terminated by the addition of SDS (0.5%) and incubated with proteinase K (200 g/ml) at 56°C for 1 h. Reaction components were resolved at 20 V for 16 h on a 1.3% agarose gel containing 0.7 g/ml ethidium bromide (42) to allow the relaxed closedcircular plasmid to run with greater mobility than open, nicked-circular DNA. Stabilized cleavable complexes are indicated by the intensity of the linear DNA band.
FIG. 3. EMSA using a topo II binding site: effects of 14-3-3⑀. Oligonucleotides, containing a topo II binding site from positions 87 to 126 of pBR322, were annealed and end-labeled with [␣-32 P]dCTP. Panel A, nuclear extract (1 g of protein) from HeLa cells was assayed for binding activity to the 32 P-labeled topo II binding site in the presence of 0.1 g of poly(dI-dC)⅐(dI-dC). The DNA-protein complex (bound) was separated from free probe (free) by electrophoresis through a non-denaturing, 4% polyacrylamide gel. The specificity of the DNA-protein interaction was determined by competition EMSA. HeLa nuclear extract (1 g of protein) was incubated with 32 P-labeled binding site and increasing molar excesses (20-, 100-, and 250-fold) of unlabeled competitor oligonucleotide (self) or an unlabeled competitor oligonucleotide of unrelated sequence (nonspecific). Panel B, HeLa nuclear extract (1 g of protein) was incubated with 32 P-labeled topo II binding site and increasing amounts of a topo II␣ antiserum (topo II␣ Ab) or a nonspecific antiserum (nonspecific Ab). The supershifted complex produced with the addition of antiserum is indicated (supershift). Panel C, the effect of 14-3-3⑀ protein on modulating specific topo II binding activity was evaluated by including in each incubation the indicated amounts of GST14-3-3⑀ fusion protein, GST14-3-3 fusion protein, or GST protein alone.
casein kinase II has been demonstrated to stabilize topo II decatenation activity in a phosphorylation-independent manner (49). Topo II appears to be phosphorylated in a cell cycle phase-specific manner, with the highest levels present at G 2 /M. It has been suggested that phosphorylation of topo II may regulate the activity of this enzyme, although this remains a subject of continued scrutiny. While phosphorylation of topo II by casein kinase II has been shown to increase the intrinsic ATPase activity of the enzyme (47), this has not been a consistent finding (50). The interpretation of these studies is complicated by the use of topo II from different species and whether or not cell cycle variations in the level of site-specific phosphorylation are taken into consideration.
Although phosphorylation of its protein partner is not an absolute requirement (51), the interaction of 14-3-3 with its targets is most often mediated by the recognition of a phosphoserine motif. Two consensus sequences for 14-3-3 binding have been identified (52,53). Examination of the topo II sequence reveals a single potential 14-3-3 binding site ( 1080 RGYDSDP 1088 ) that also contains within it a casein kinase II consensus phosphorylation site (54). Whether this site is phosphorylated by casein kinase II or contributes to the interaction with 14-3-3⑀ is under investigation.
To establish the functional consequences of these interactions, we have used three independent assays to evaluate the impact of 14-3-3⑀ on distinct topo II activities. Using both a modified alkaline comet assay with isolated HL-60 nuclei and a plasmid DNA cleavage assay, we demonstrated that 14-3-3⑀ negatively affects the ability of etoposide to trap topo II in cleavable complexes, thereby preventing the induction of DNA strand breaks. By EMSA, it is demonstrated that this appears to be due to reduced DNA binding activity. Whether 14-3-3⑀ induces a change in the preferred DNA binding sequence for topo II remains to be tested. Taken together, these findings suggest that the effect of 14-3-3⑀ on topo II activity be evaluated in intact cells, since the consequence of this interaction is reminiscent of topo II inhibition as a result of physical association with the tumor suppressor gene product, Rb (10).
Although an extensive range of studies investigating 14-3-3 proteins have been published, few papers characterize more than one isoform, while even fewer reports experimentally exclude specific isoforms as candidate interacting proteins. While the amino acid sequences of the seven identified 14-3-3 isoforms are highly conserved (55), the retention of multiple isoforms throughout evolution suggests that they may have evolved distinct, non-redundant functions. Although some proteins, such as Raf1, may be promiscuous in their association with multiple 14-3-3 isoforms (20,22,56), other 14-3-3 protein partners demonstrate a degree of selectivity (31,57). This report demonstrates that 14-3-3⑀, but not 14-3-3, interacts with topo II, thereby modulating several in vitro measures of topo II function. 14-3-3⑀ is an atypical isoform, being more closely related to yeast and plant 14-3-3 proteins than to other mammalian isoforms (55,58). This early divergence may have given rise to distinct roles for 14-3-3⑀. It should be considered that the limited distribution of 14-3-3 to cells of epithelial origin, in contrast with the ubiquitous distribution of 14-3-3⑀, may also play a role in influencing their functions (19,59). Although it is clear that topo II interacts with 14-3-3⑀ (and 14-3-3), 2 but not with 14-3-3, it is not currently known whether topo II interacts with other 14-3-3 isoforms.
There are several hypotheses as to the mechanism of 14-3-3⑀-mediated disruption of topo II function. First, dimerization is essential for topo II DNA-binding and function. Recently, two sequences in human topo II (spanning amino acids 1053-1069 and 1124 -1143) were identified to be crucial for this dimerization, with disruption of either sequence leading to a loss of capacity to dimerize (60). It is of note that the candidate 14-3-3⑀ binding site falls between these two sequences. Second, numerous 14-3-3 isoforms, including 14-3-3⑀, have been recently described as mediating subcellular localization of their protein partners, including protein kinase U␣ (36), and, in response to DNA damage, the cell cycle regulatory protein cdc25, thereby inducing cell cycle arrest (32)(33)(34)(35). Hypothetically, the interaction between 14-3-3⑀ and topo II could modulate the subcellular localization of topo II, either during phases of the cell cycle, or in response to DNA damage thereby modulating cellular sensitivity to topo II poisons. Whether the binding of 14-3-3⑀ disrupts topo II homodimerization, modulates topo II localization, or functions through a mechanism yet to be considered remains the focus of our ongoing investigation.
In summary, we have identified 14-3-3⑀ as a topo II-interacting protein. Our finding that this interaction with topo II does not extend to all 14-3-3 isoforms, as demonstrated by the lack of interaction with 14-3-3, is one of a small, but growing, number of reports that hint at functional specificity within this well conserved protein family. This observation suggests that evolutionary retention of multiple isoforms may be attributable to hereto unappreciated distinct and non-redundant functions for individual isoforms or subsets thereof. The demonstration that 14-3-3⑀ disrupts topo II/DNA interaction and reduces etoposide-mediated cleavable complex formation and DNA strand breaks has significant implications for the cellular sensitivity to topo II-directed chemotherapeutics.