Casein kinase II copurifies with yeast DNA topoisomerase II and re-activates the dephosphorylated enzyme

The eukaryotic topoisomerase II (topo II) is an abundant and essential nuclear enzyme, capable of modifying DNA topology. Studies in yeast demonstrate that topoisomerase II is required to decatenate replicated sister chromatids during both mitotic and meiotic divisions (Dinardo et al., 1984; Holm et al., 1985; Uemura and Yanagida, 1986; Rose et al., 1990), and by means of the same catalytic reaction, it is able to relax both positively and negatively supercoiled DNA in increments of two in an ATP-dependent manner (reviewed by Wang, 1985). In addition, topoisomerase II is a target of many currently used anti-tumor drugs, which block the religation event and leave the enzyme trapped in a covalent complex with DNA (reviewed by Liu, 1989).

Observations from several laboratories suggest that topoisomerase II plays an important role in chromosome organization and condensation at metaphase, in addition to its decatenation function. In the fission yeast, Schizosac - charomyces pombe, cells that are conditional mutants in both b-tubulin and topoisomerase II arrest with elongated and decondensed chromosomes (Uemura et al., 1987). Antibodies against topoisomerase II reveal a localization of the a form of the enzyme along the axis of intact metaphase chromosomes in HeLa and chicken cell lines (Earnshaw et al., 1985; Gasser et al., 1986). In addition, topoisomerase

II is the major nonhistone component of the metaphase scaf-fold, a residual chromosomal structure that resists extraction under conditions that efficiently remove histones (Lewis and Laemmli, 1982; Gasser et al., 1986; Earnshaw et al., 1985). More direct evidence that topoisomerase II is essential for chromosome condensation comes from studies in a Xenopus extract that support nuclear disassembly. Depletion of topoisomerase II was shown to arrest con- densation of chromosomes in in vitro extracts (Adachi et al., 1991).

Topoisomerase II has been shown to be a phosphoprotein in intact cells from a variety of species (Ackerman et al., 1988; Heck et al., 1989; Kroll and Rowe, 1991; Cardenas et al., 1992). Two previous reports have described kinase activities that co-purify with topoisomerase II (Sander et al., 1984; Saijo et al., 1990). Conflicting results have been published as to the character of this kinase. In one case it was proposed that topoisomerase II might be an autophosphorylating enzyme (Sander et al., 1984), while in the other a kinase was identified that had some characteristics of casein kinase II, but which differed in other aspects (Saijo et al., 1990). In this paper we characterize the properties of the copurifying kinase from Saccharomyces cere - visiae and use a casein kinase II-temperature sensitive mutant to show conclusively that the topoisomerase II-copurifying kinase is casein kinase II.

We have recently demonstrated by phosphotryptic peptide mapping that casein kinase II modifies yeast topoisomerase II in intact cells blocked in either G1 or mitotic phases of the cell cycle (Cardenas et al., 1992). The sites of modification were mapped to multiple serine and threonine residues in the C-terminal 350 amino acids of the protein. In this report, we show that after dephosphorylation the decatenation and cleavage activities of yeast topoisomerase II are greatly reduced, and that phosphorylation of the C-terminal domain by purified CKII restores that decatenation activity in vitro. Hyperphosphorylation by CKII does not, however, significantly stimulate or alter the site-specific double-stranded cleavages induced by topoisomerase II on naked DNA. The C-terminal phosphateaccepting domain appears to be a negative regulatory element that is neutralized by phosphorylation.

The S. cerevisiae strain GA-24: MATa ura3 his bar1 suc2–9 pep4-3 was provided by Dr H. Riezman. S. cerevisae strain YDH8: MATa cka1::HIS3 cka2::TRP1 ura3 pRS315(CEN/ARS H4 LEU2 cka2-8_ts) was isolated in the laboratory of Dr C. V. C. Glover (Georgia), and details of this will be presented elsewhere (D. H. and C. V. C. Glover, in preparation). The plasmid YEpTOP2-PGAL1 (Giaever et al., 1988) was kindly provided by Dr J. C. Wang.

Topoisomerase II overexpression and purification from strains GA-24 or YDH8 bearing YEpTOP2-PGAL1 was as described by Worland and Wang (1989), except that buffer 1 contained 50 Mm HEPES-KOH, pH 7.7, 1 mM EDTA, 1 mM EGTA, 1 mM dithio-threitol, 10% (v/v) glycerol and, where indicated, this buffer was supplemented with KCl instead of NaCl. In addition, all buffers contained the protease and/or phosphatase inhibitors 1 mM phenylmethylsulfonyl fluoride, 0.5% Trasylol (100 units/ml apro-tinin, Bayer), 1 mM sodium bisulfite, 1 mM NaF, 2 μg/ml pep-statin and 30 mM Na-pyrophosphate. b-glycerolphosphate was found to interfere with the phosphocellulose chromatography and was omitted. For purification of topoisomerase II from the CKII mutant strain YDH8, the culture was grown at 24°C and then induced by addition of galactose either at 24°C or 37°C. The induced topoisomerase II was purified as for wild-type cells but through the polyethyleneimine-celite step only.

The fractions from the phosphocellulose column containing topo-isomerase II were pooled, and 200 ml were loaded onto a 12 ml linear gradient of 15% to 40% (v/v) glycerol in buffer 1 containing 400 mM KCl. The gradient was centrifuged for 68 h at 39,000 revs per min in a Kontron TST41 rotor. 200 μl fractions were collected starting from the bottom of the gradient and were then assayed for DNA decatenation and protein kinase activities.

Topoisomerase II activity was tested by decatenation of kineto-plast DNA isolated from Crithidia fasciculata as described (Simp-son and Simpson, 1974). Decatenation reactions contained 10 mM HEPES-KOH, pH 7.4, 100 mM KCl, 10 mM MgCl2, 0.5 mM dithiothreitol, 0.5 mM EDTA, 30 μg/ml bovine serum albumin, 1 mM ATP and 13 μg/ml DNA. Reactions were initiated by addition of topoisomerase II, and incubations were for 15 min at 30°C. Reactions were stopped by addition of SDS to 1%, EDTA to 10 mM, and proteinase K to 25 μm/ml. The mixtures were further incubated at 60°C for 30 min and were analyzed by 1% agarose gel electrophoresis in Tris-acetate buffer.

Protein kinase assays were performed with the phosphocellulose or the glycerol gradient-purified topoisomerase II. When glycerol gradient-purified topoisomerase II was used as substrate, the kinase sources were casein kinase II purified from chicken (Nak-agawa et al., 1989) or the glycerol gradient fractions containing the peak of topoisomerase II copurifying kinase. Also where indi-cated, 10 μm casein were used as a substrate. Phosphorylation reactions using the copurifying kinase or casein kinase II were carried out in a total volume of 20 ml kinase buffer containing 50 mM Tris-HCl, pH 7.4, 150 mM KCl, 10 mM MgCl2, 0.1 mM

EDTA, 10 mM ATP and 5 mCi of [γ-₃₂P]ATP (Amersham). The mixture of the radiolabelled and unlabelled ATP will be referred to as 10 mM [γ-₃₂P]ATP throughout the text. In some assays, GTP, heparin or the RRREEETEEE peptide were included in the reactions at the indicated concentrations.

The phosphorylations by protein kinase C and p34_cdc2 kinases are described by Cardenas et al. (1992). The yeast Ca₂₊/calmod-ulin-dependent protein kinase (kindly provided by Dr. M. Prusehy), was assayed in 50 mM Tris-HCl pH 7, 10 mM MgCl2 and 10 mM [γ-₃₂P]ATP, supplemented with 1 mM CaCl2 and 2 μM calmodulin where indicated, or with 2 mM EGTA for control reactions.

In most experiments, reactions were incubated for 15 min at 30°C. However for reactions shown in Fig. 2, the enzyme was preincubated at 25°C (permissive temperature) or at 37°C (non-permissive temperature) for 30 min prior to the addition of [γ-₃₂P]ATP. The reactions were then carried out at either 25°C or 37°C for the indicated times. Reactions were stopped with SDS-sample buffer and analyzed by SDS-PAGE.

Calf intestinal alkaline phosphatase attached to agarose beads (Sigma) was washed three times in the kinase buffer containing 0.1 mM ZnCl2 and resuspended in the same buffer at 1:1 volume with respect to beads. Topoisomerase II from the glycerol gradient, which had or had not been phosphorylated with the copurifying kinase, was incubated with 0.5 units of alkaline phosphatase in 20 ml of kinase buffer containing 0.1 mM ZnCl2 for 30 min at 30°C.

These analyses were carried out as described by Beeman and Hunter (1978) and Cooper et al. (1983) with minor modifications (Peter et al., 1990).

DNA cleavage reactions were performed on the end-labelled and gel purified SphI-HinfI fragment from the Drosophila ftz scaffold attached region, in 10 mM Tris-HCl, pH 7.5, 5 mM MgCl2, 5 mM CaCl2, 50 mM KCl, with 50 ng of purified topoisomerase II for 5 min at 30°C. After this incubation the reaction was adjusted to a final concentration of 1% SDS by addition of prewarmed SDS. After digestion with proteinase K and extraction with phenol, the precipitated DNA fragment was denatured in formamide sample buffer and cleavage products were separated on a 6% sequencing gel.

Topoisomerase II was dephosphorylated prior to the DNA cleavage reaction as follows: calf alkaline phosphatase attached to agarose beads (Sigma) was washed three times in cold cleavage buffer and resuspended to a concentration of 10 units/ml. Dephosphorylation of topoisomerase II was performed in cleav-age buffer at 30°C for 30 min, at a ratio of 1 unit phosphatase per 500 ng topoisomerase II. Beads were removed by centrifugation before addition to the cleavage reaction. Phosphorylation of topo-isomerase II was performed in twice concentrated cleavage buffer including 10 mM ATP for 15 min at 30°C. The samples were then diluted to cleavage conditions prior to use. For control lanes, topoisomerase II was incubated for equivalent times in the same buffers without the addition of phosphatase beads or ATP.

Worland and Wang (1989) have shown that topoisomerase II can be overexpressed in yeast by placing the yeast TOP2 gene under the control of the GAL1 promoter, allowing a strong induction of topoisomerase II by galactose (Giaever et al., 1988). The overexpressed topoisomerase II was puri-fied by a simple and efficient procedure (Worland and Wang, 1989), which we have modified slightly in this study. The results from the purification are summarized in Fig. 1. Following overexpression, the band at 170 kDa, which cor-responds to topoisomerase II, becomes visible in the total cell homogenate on an SDS gel stained by Coomassie blue (Fig. 1, lane 1, arrow). Topoisomerase II was then purified by polyethyleneimine-celite and phosphocellulose chro-matography, resulting in a highly purified enzyme fraction (Fig. 1, lanes 10–12). When the gel is overloaded, a ladder of polypeptides smaller than the full-size topoisomerase II polypeptide is observed (lanes 10–11). On western blots these minor polypeptides react with an antibody affinity-purified against the topoisomerase II band at 170 kDa, sug-gesting that these are degradation products of the intact enzyme (data not shown). The purified fraction (Fig. 1, lanes 10-12) was used to analyze the activation of topoi-somerase II by phosphorylation.

Two previous studies have reported that a kinase activity copurifies with eukaryotic DNA topoisomerase II (Sander et al., 1984; Saijo et al., 1990). Because our studies of topoII phosphorylation require that the enzyme is free from any kinase activity, we tested the phosphocellulose peak frac-tion for an endogenous phosphorylating activity. When incubated with [γ -32P]ATP, this fraction revealed a strong kinase activity that efficiently modified topoisomerase II (Fig. 2, lanes a-e, arrowheads). This copurifying kinase could be separated from topoisomerase II by glycerol gra-dient centrifugation in 0.4 M KCl (Fig. 3). Coomassie blue staining of the gradient fractions revealed a band of 170 kDa in fractions 20–29 (data not shown), which also contain the decatenation activity (Fig. 3B). No major polypeptides are stained by Coomassie blue in fractions 40–48, in which the kinase activity is recovered (Fig. 3A). When this gradient-purified kinase was challenged with a range of substrates, we found that topoisomerase II was more efficiently phos-phorylated by this kinase than either histone H1 or casein (data not shown).

Topoisomerase II has been reported to be phosphorylated by various kinases in vitro, including protein kinase C, cAMP-dependent kinase, the Ca2+/calmodulin-dependent kinase (Sahyoun et al., 1986; Rottman et al., 1987) and casein kinase II (Ackerman et al., 1985). A casein kinase II-like activity was shown to phosphorylate Drosophila topoisomerase II in whole cell homogenates and in intact cells (Ackerman et al., 1988), and our own studies demon-strated that in yeast CKII is a major kinase modifying topoII in vivo (Cardenas et al., 1992). Since the co-purifying kinase has a high affinity for topoisomerase II as a sub-strate, we decided to characterize this kinase by testing it for casein kinase II properties.

A hallmark of casein kinase II is that GTP can serve nearly as well as ATP as the phosphate donor (Cochet et al., 1982). Accordingly, we found that transfer of 32P from [γ -32P]ATP to topoisomerase II by the copurifying kinase was effectively competed for by GTP (Fig. 4A). From the titration of cold GTP into the reaction mixture shown in Fig. 4A, we can calculate that the presence of 70 mM GTP reduces the amount of 32P transferred to topoisomerase II by 50%. In addition, 32P was readily incorporated into topoisomerase II when the kinase reaction contained [g-32P]GTP (not shown), demonstrating that GTP can serve as a substrate for the phosphorylation reaction.

While most known protein kinases are relatively insen-sitive to heparin, casein kinase II is strongly inhibited by heparin at concentrations of less than 1 μm/ml (Hathaway et al., 1980; Hathaway and Traugh, 1982). The topoiso-merase II copurifying kinase was inhibited 50% by 0.05 μm/ml heparin and completely blocked by 0.25 μm/ml heparin (Fig. 4B). This result is comparable to the I50 of 0.05 μm/ml heparin previously determined by Padmanabha and Glover (1987) for purified casein kinase II from S. cere -visiae.

The substrate specificity requirements for casein kinase II have been defined using proteins and synthetic peptides as substrates (for a review see Pinna, 1990). These studies revealed that casein kinase II phosphorylates serine or thre-onine residues located on the N-terminal side of clusters of acidic amino acids; in particular, a negatively charged residue is required three amino acids away on the C-ter-minal side of the target residue. Because the synthetic pep-tide RRREEETEEE has been shown to be the preferred sub-strate for CKII phosphorylation (Kuenzel and Krebs, 1985), it was tested as a competitor for phosphorylation of topoi-somerase II by the copurifying kinase. At 0.5 mM this pep-tide virtually blocked topoisomerase II phosphorylation by the copurifying kinase (Fig. 4C). Taken together, these results suggest that a single kinase with the properties of casein kinase II copurifies with topoisomerase II.

Further evidence that the topoisomerase II-copurifying kinase is casein kinase II, was obtained from two-dimen-sional phosphotryptic peptide maps. Kinase-free, gradient-purified topoisomerase II was phosphorylated by either the copurifying kinase or purified casein kinase II, eluted from an SDS-gel, digested with trypsin and subjected to two-dimensional phosphopeptide analysis, as described in Mate-rials and methods. As shown in Fig. 5, topoisomerase II phosphorylated by purified chicken casein kinase II yielded 7 major and a few minor phosphopeptides (Fig. 5A), and an identical pattern was obtained from topoisomerase II phos-phorylated with immunoprecipitated chicken casein kinase II (not shown). The phosphopeptide map from topoiso-merase II phosphorylated by the copurifying kinase was strikingly similar to these (Fig. 5B), and a mixture of radio-equivalent amounts (determined by Cerenkov counting) of the two tryptic digests shows that the major and the minor phosphopeptides from either phosphorylation reaction comigrate (Fig. 5C). Moreover, quantitation of two-dimensional phosphoamino acid analysis showed that both the copurifying kinase and casein kinase II phosphorylate serine and threonine residues in a similar ratio (65% Ser, 35% Thr and 70% Ser, 30% Thr, respectively (Fig. 5D and E).

Previous studies have identified these two diagonal series of phosphopeptides as multiphosphorylated forms of pep-tides derived from the C-terminal domain of yeast topoII (Cardenas et al., 1992). The significance of their location for topoisomerase II activation will be discussed below.

Casein kinase II is a tetrameric complex composed of two catalytic subunits (α and α′) and two β subunits, which are targets for modification, presumably autophosphorylation (for a review, see Pinna, 1990). The genes for the S. cere -visiae a (CKA1) and α′ (CKA2) subunits have been cloned and sequenced (Chen-Wu et al., 1988; Padmanabha et al., 1990). Disruption of either the CKA1 or the CKA2 gene alone does not affect growth or viability (Chen-Wu et al., 1988; Padmanabha et al., 1990). However, disruption of both CKA1 and CKA2 is lethal. Double disrupted cells car-rying a temperature-sensitive CKA2 allele (cka2-8ts) arrest cell division at the restrictive temperature as half unbudded and half large-budded cells (Padmanabha et al., 1990; Hanna and C.V.C. Glover, unpublished data). We antici-pated that if the topoisomerase II copurifying kinase is iden-tical to casein kinase II, the kinase activity should be markedly decreased or absent in topoisomerase II purified from a temperature sensitive CKII mutant (strain YDH8) at the restrictive temperature.

Topoisomerase II expression from the GAL1 promoter was induced for 10 hours in strain YDH8 bearing the YEpTOP2-PGAL1, and the culture was then shifted to 37°C for 4 hours. Following the temperature shift, topoisomerase II was purified and then incubated with [γ-32P]ATP. In par-allel, topoisomerase II was purified from the YDH8 cells grown at the permissive temperature. In contrast to the strong kinase activity detected in the topoisomerase II preparation from CKA1+CKA2+ wild-type cells (Fig. 2A, lanes a-e), minimal kinase activity could be detected in the preparation from the temperature-sensitive cka2-8ts cells grown at 25°C (Fig. 2A, lanes f-j), and only a trace of activity was observed in a preparation from cells placed at the restrictive temperature (Fig. 2A, lanes k-o). Coomassie blue staining shows that equivalent amounts of topoiso-merase II have been assayed in each lane (Fig. 2B), allow-ing the conclusion that the kinase copurifying with topoi-somerase II is deficient in the cka2-8ts mutant. This genetic evidence, along with the biochemical characterization pre-sented above, proves conclusively that the topoisomerase II-copurifying kinase is casein kinase II.

To examine the effects of phosphorylation on topoiso-merase II activity we were obliged first to separate topoi-somerase II from the copurifying kinase. This is achieved by glycerol gradient centrifugation, after which decatena-tion activity is measured in the presence or absence of puri-fied casein kinase II. We have allowed phosphorylation to proceed to its maximum (reached after 20 min at 30°C) and find that modification by CKII strongly enhances the gra-dient-purified topoisomerase II DNA decatenation activity (Fig. 6). Densitometric scanning of this gel suggests that the level of topoII activity is more than twenty-fold greater after phosphorylation than that detected with the unphos-phorylated enzyme after an equivalent incubation (left lane, Fig. 6). It is not clear, however, that the decatenation reac-tion proceeds in a linear fashion. Earlier workers have reported a 4-to 6-fold activation of Drosophila topoiso-merase II by casein kinase II (Ackerman et al., 1985).

This difference in degree of stimulation may reflect the fact that the yeast topoisomerase II is highly overexpressed prior to purification. Due to a disproportionate ratio of sub-strate (topoII) to kinase (CKII) in these cells, we suspect that the yeast topoisomerase II isolated in this study is underphosphorylated compared with that purified from Drosophila or calf thymus tissue. Consistently, we find that the overexpressed yeast topoisomerase II has nearly a ten-fold lower specific activity than topoisomerase II isolated from calf thymus or Drosophila (S. M. G. and P. Jensen, data not shown). We thus propose that the enzymes puri-fied from calf thymus and Drosophila are already partially activated, perhaps through modification at relatively stable phosphorylation sites. The existence of fairly stable phos-phorylated residues on topoisomerase II is suggested from a study of the human enzyme by Kroll and Rowe (1991).

Interestingly casein kinase II is not the only kinase that stimulates topoisomerase II activity in vitro. We also find that topoisomerase decatenation activity is activated in vitro by phosphorylation by protein kinase C, p34cdc2 kinase or by Ca2+/calmodulin-dependent protein kinase (Fig. 6). It is not known whether stimulation by the other kinases is phys-iologically relevant, nor whether it is achieved by the same mechanism as CKII (see Discussion). Because the two-dimensional phosphopeptide maps obtained from topoiso-merase II phosphorylated in vivo correspond to the sites modified by casein kinase II (Cardenas et al., 1992), we have characterized further the effects of phosphorylation by CKII on topoisomerase II activity in vitro.

While Fig. 6 and previous studies demonstrated that a basal level of topoisomerase II activity is enhanced following phosphorylation by casein kinase II (Ackerman et al., 1985), it was not clear whether or not CKII phosphoryla-tion was sufficient to activate topoisomerase II from a fur-ther dephosphorylated state. To demonstrate this we treated the gradient-purified topoisomerase II with alkaline phos-phatase prior to rephosphorylation by CKII and screened both states for ATP-dependent decatenation activity. Topo-isomerase II purified from the glycerol gradient shows a low DNA decatenation activity (open circles, Fig. 7). With equal amounts of topoisomerase II added, no decatenation activity is detectable by our assay after incubation with alkaline phosphatase prior to the assay (open triangles, Fig. 7). Incubation of the alkaline phosphatase-treated topoiso-merase II with the copurifying kinase (i.e. the glycerol gra-dient fraction containing the kinase peak, see Fig. 2) dra-matically enhances DNA decatenation activity (Fig. 7, filled triangles), nearly to the level of the phosphorylated enzyme which was not phosphatase treated (Fig. 7, filled circles). We conclude that modification of topoisomerase II by the copurifying casein kinase II is sufficient to restore the activity lost due to alkaline phosphatase treatment. We cannot exclude that there are some phosphorylated residues that are insensitive to alkaline phosphatase and are required for activation, but they are, in any case, not sufficient to activate topoII decatenation activity.

There are multiple steps in the decatenation reaction cat-alyzed by topoisomerase II. The enzyme must bind DNA, recognize the catenated state of the DNA, create a double strand cut, pass strands through the cleavage site, and reli-gate the double helix. For an efficient decatenation assay, the enzyme must also have a high turnover rate, i.e. leave its original site and repeat the operation at a new site. While certain of these subreactions are difficult to study isolated from the others, a preferential interaction of topoisomerase II with specific DNA sequences can be mapped by adding SDS to an incubation of topoisomerase II and DNA. Under these conditions the enzyme denatures, remaining in a cova-lent complex with the DNA, resulting in a double strand cut (for a review see Liu, 1989). This effectively isolates the first two steps of the decatenation reaction from subse-quent steps, and allows us to test whether phosphorylation by casein kinase II enhances this interaction or alters the sequence specificity of topoisomerase II.

We have mapped two strong topoisomerase II cleavage sites within a 1.2 kb fragment of DNA located upstream of the Drosophila gene fushi tarazu (R. W. and S. M. G. unpublished results). This fragment is an A-T rich region that shows specific association with histone-depleted nuclear structures, called nuclear scaffolds (Gasser and Laemmli, 1986). One of the preferred topoisomerase II cleavage sites falls within the 320 bp SphI-HinfI fragment pictured in Fig. 8B. The enzyme cleaves both strands, resulting in two end-labelled single strand fragments of 95 and 225 nucleotides from the double end-labelled 320 bp probe (see open circles, lanes 2-4, Fig. 8A). Neither the amount of complex trapped nor the choice of site is altered following phosphorylation by casein kinase II (Fig 8A, lane 5). In contrast, after dephoshorylation by alkaline phos-phatase, no cleavage of the ftz probe could be detected (Fig. 8A, lane 6). This is consistent with the result shown in Fig. 7, where the topoisomerase II decatenation activity is strik-ingly diminished after dephosphorylation by alkaline phos-phatase. These experiments demonstrate that the stimula-tion of decatenation activity does not reflect an enhanced rate of cleavage or a loss in specificity. Rather, the phos-phorylation of topoisomerase II appears to work either at the level of turnover (recycling) of the enzyme or on its association rate with DNA. Studies addressing this are in progress.

An increasing number of enzymes catalyzing reactions crit-ical for cell cycle progression are found to be regulated by phosphorylation. DNA topoisomerase II, like DNA ligase I, is an enzyme involved in DNA replication, and both are regulated by casein kinase II (Prigent et al., 1992; Carde-nas et al., 1992). In this paper we show that casein kinase II copurifies with topoisomerase II, and we present a method for the separation of the two enzymes. We show that phosphorylation of topoisomerase II by CKII signifi-cantly stimulates the decatenation activity of topoII, and that the dephosphorylated enzyme is inactive in the same assay. We have recently demonstrated that the casein kinase II target sites in yeast topoisomerase II are modified in vivo in both G1 and metaphase, with a significantly higher level of modification at metaphase (Cardenas et al., 1992). In yeast we have not observed large variations in the level of topoisomerase II between cycling and noncycling cells, nor between G1 and mitosis (Cardenas et al., 1990, 1992), thus the changes in phosphorylation are likely to help regulate this enzyme in vivo. A modulation of phosphorylation inde-pendent of the protein level was also observed in mouse 3T3 cells by Saijo et al. (1992).

Two previous reports have noted that a kinase activity copurifies with topoisomerase II through several purifica-tion columns (Sander et al., 1984; Saijo et al., 1990), and one of these proposed that topoisomerase II might itself have kinase activity (Sander et al., 1984). Our study shows clearly that yeast topoisomerase II has no endogenous kinase activity, and the copurifying kinase can be efficiently separated from topoII by glycerol gradient centrifugation in 0.4 M KCl. The copurifying kinase activity is inactive at 36°C in a temperature-sensitive casein kinase II mutant, has the biochemical characteristics typical for CKII, and phos-phorylates topoisomerase II on the same tryptic peptides as purified chicken CKII. These considerations make it extremely likely that the copurifying kinase is casein kinase II itself. It has been noted that CKII copurifies with other bona fide nuclear substrates, such as nucleolin (Caizergues-Ferrer et al., 1987). This may reflect a relatively high affin-ity between CKII and each of these substrates, rather than a fortuitous co-purification event.

Casein kinase II is highly conserved, containing a cat-alytic subunit α that complexes the regulatory b subunit in an α2β2 heterodimer configuration. The β subunit appears to be encoded by a single gene, while there are two a-sub-units encoded by different genes in yeast (Padmanabha and Glover, 1987). Previous localization studies found CKII in the cytosol, nucleolus, nucleus and possibly in a cell mem-brane fraction (e.g. Yu et al., 1991), although a more recent study finds the enzyme almost exclusively nuclear (Krek et al., 1992). This is consistent with the observation that many proteins important in controlling nuclear events are phos-phorylated by CKII (reviewed by Meisner and Czech, 1991), including c-ErbA, E1A, the E7 protein of human papilloma virus, SV-40 T antigen, the avian c-Myb, Max, c-Jun, c-Fos, p53, cAMP response element binding factor (CREB), serum response factor (SRF), mUBF, DNA ligase, and two nucleolar substrates, B23 and nucleolin (reviewed by Meisner and Czech, 1991; Hunter and Karin, 1992; see also Berberich and Cole, 1992; Lin et al., 1992; Lüscher et al., 1990; Voit et al., 1992; Prigent et al., 1992; Meier and Blobel, 1992; Ackerman et al., 1988; Cardenas et al., 1992). The effects of CKII on these various substrates is diverse, as discussed by Meisner and Czech (1991) and Hunter and Karin (1992).

In our hands, the topoisomerase II purified from over-expressing yeast cells has a lower specific activity than the enzyme isolated from calf thymus or Drosophila (P. Jensen and S. M. G., unpublished), and the yeast enzyme shows a much more significant stimulation of decatenation activity by casein kinase II than reported for higher eukaryotic enzymes. Our interpretation of these results is that the over-expressed yeast topoII is in an underphosphorylated state, and therefore a more dramatic stimulation of activity is pos-sible. Phosphorylation by CKII is also sufficient to re-acti-vate alkaline phosphatase-treated topoisomerase II, which has a very low decatenation activity and is unable to cleave DNA efficiently in vitro (Figs 7 and 8). We cannot exclude that some stable phosphate groups added by other kinases are insensitive to the phosphatase treatment, and are required for activity, yet our results strongly suggest that CKII modification of the C terminus alone greatly stimu-lates the decatenation activity of the protein.

Eukaryotic type II topoisomerases contain roughly 1400 amino acids, and of these the first 900 N-terminal amino acids show a high degree of sequence homology with bac-terial gyrase. Regions of significant homology can also be aligned up to amino acid 1170 in the S. cerevisiae gene, but thereafter homodogy with the bacterial gyrase disap-pears, and even among eukaryotic species the region from amino acid 1167 to the end of the protein is highly diver-gent (Caron and Wang, 1992). It is remarkable, therefore, to find that all the CKII target sites are located in the extreme C terminus, in one of the least conserved domains among eukaryotes, and a region absent in prokaryotic topoi-somerases. A priori, this C-terminal, phosphate-accepting domain could represent a part of the protein involved in regulating the aspects of topoisomerase II function that are unique to eukaryotic cells. Since bandshift data in which both the dephoshorylated and phosphorylated topoiso-merase II produce complexes of similar mobility (Q. Dang, G-C. Alghisi and S. M. G., unpublished results), the addi-tional negative charge in the C terminus is unlikely to be involved in the dimerization of topoisomerase II.

Analyses of truncated forms of yeast topoisomerase II have provided insights into the importance of the C-termi-nal domain and its phosphorylation sites. If the C-terminal domain of topoisomerase II is truncated in the S. cerevisiae TOP2 gene to amino acid 979, and of the S. pombe top2+ gene to amino acid 922, neither can rescue top2-mutant cells (Shiozaki and Yanagida, 1991; M. E. C. and S. M. G., unpublished). In contrast, a C-terminal truncation of the S. pombe topoisomerase II to amino acid 1198 can partially complement S. pombe top2 mutations, producing a condi-tional mutant that will not grow at elevated or reduced tem-peratures (Shiozaki and Yanagida, 1991). Surprisingly, however, protease digestion studies show that full topoiso-merase II decatenation activity is maintained by a prote-olytic core of the enzyme that extends only to amino acid 1220. Thus, the C-terminally truncated topoisomerase II is active in vitro, while the full-length, dephosphorylated enzyme is inactive in vitro, suggesting that the C-terminal 259 amino acids serve as a negative regulatory domain that in a dephoshorylated state somehow masks the active sites of the enzyme (see model in review by Cardenas and Gasser, this issue). Upon phosphorylation, the C terminus would become more extended, unmasking the active site exposing the DNA binding regions. If this phosphate-acc-cepting, inhibitory domain is absent, topoisomerase II should no longer require phosphorylation to be active, and should indeed become insensitive to the stimulatory effects of the kinases.

A precedent for this type of regulation has recently been described for human DNA ligase I (Prigent et al., 1992), where it has been shown that the full-length human ligase I gene is unable to rescue conditional lethal ligase mutant E. coli cells, presumably due to its inactivity. However, expression of an N-terminally truncated DNA ligase I allows the mutant E. coli cells to grow at the restrictive temperature (Kodama et al., 1991). Intriguingly, mam-malian ligase l is a phosphoprotein in intact cells and in vitro its N terminus is phosphorylated by CKII. The full-length, E. coli-made form of DNA ligase l is inactive in vitro, but phosphorylation by CKII at N-terminal acceptor site(s), renders the enzyme active (Prigent et al., 1992). The mode of activation by CKII may therefore be very similar for DNA topoisomerase I and DNA ligase I, as reviewed by Cardenas and Gasser (this issue).

Why should the eukaryotic DNA topoisomerase II have evolved a mechanism for negative regulation of its activity? As long as there is transcription or replication proceeding in the cell, the relaxation activity of either type I or type II topoisomerase is needed. In the absence of growth, how-ever, as in serum-starved cells, the cell may need to inac-tivate topoisomerase II, which may be achieved by a drop in casein kinase II activity (Yu et al., 1991). The positive regulation by enhanced phosphorylation may also provide a rapid stimulation of activity at the end of DNA replica-tion. Alternatively, the phosphorylated C-terminal domain may have a secondary function that involves interaction with other proteins or with other molecules of topoiso-merase II. This possibility is currently under study.

Several highly purified kinases, each with a unique and defined substrate sequence specificity, stimulate the topoi-somerase II decatenation activity in vitro. This activation might be explained either by an overlap of the major sites of phosphorylation by the different kinases, or by a common effect of phosphorylation in the same domain, such as the C terminus. To examine the question of whether the vari-ous kinases modify common serine or threonine residues, we have generated 2-dimensional tryptic phosphopeptide maps of topoisomerase II phosphorylated by the different kinases in vitro (Cardenas et al., 1992). Phosphorylation by each of the different kinases gives a unique tryptic map, with the exception of one phosphopeptide that appears to be a common target for both protein kinase C and p34cdc2 (see Cardenas et al., 1992). Perhaps the most likely expla-nation is that the C terminus of topoisomerase II, which is serine and threonine-rich and contains stretches of alter-nating basic and acidic residues, provides target sites for most commonly studied kinases. The precise mapping of modification sites for other activating kinases will establish whether the stimulation of topoisomerase II decatenation activity in these cases also involves modification of the C-terminal domain.

We would like to thank Drs M. Peter, V. Simanis and C. V. C. Glover for helpful suggestions throughout this work; Drs P. Parker, J. Nakagawa and M. Doreé for kinase samples; Dr W. Krek for anti-chicken casein kinase II; and Dr J. C. Wang for the TOP2 overexpression plasmid. Research was supported by grants from the Swiss Cancer League, the Swiss National Science Foun-dation and the International Human Frontier Science Program to S. M. G., and by a Roche Foundation Fellowship to M. E. C.