¹Laboratoire d'Oncologie Moléculaire, FRE 2224 CNRS, Poitiers Cedex, France

²Service de cytofluorométrie, UMR 6558 CNRS, IFR de communication intercellulaire, 40 Avenue du Recteur Pineau, 86021, Poitiers Cedex, France

³Laboratoire de Biochimie des Protéines, CHU de Poitiers, France

⁴Aventis Pharma, Centre de Recherche de Vitry-Alfortville, 94403 Vitry-sur-Seine, France

Correspondence to: C-J Larsen, Laboratoire de Biochimie des Protéines, CHU de Poitiers, France

^aJ-F Riou and P Séité equally contributed to this work

The ARF gene (p19^ARF in mouse and p14^ARF in man) has become a central actor of the cell cycle regulation process as it participates to the ARF-MDM2-p53 pathway and the Rb-E2F-1 pathway. By use of immunoprecipitation and Western blotting (IP/WB), we now show that ARF physically associates with Topoisomerase I (Topo I). ARF-Topo I immune complexes were detected in SF9 insect cells infected with recombinant baculoviruses encoding the two genes as well as in 293 cells that express endogeneously these proteins. Preparations of a GST-ARF recombinant protein stimulated the DNA relaxation activity of Topo I but, in contrast, had no effect on the decatenation activity of Topo II. The Topo I stimulation was also detected in cell extracts of SF9 cells expressing both proteins. A confocal microscopy study indicated that part of ARF and Topo I colocalized in the granular component structure of the nucleolus. As a whole, our data indicate that Topo I is a new partner of ARF and suggest that ARF is involved in cell reactions that require Topo I. Oncogene (2001) 20, 836-848.

p14^ARF; Topoisomerase I; protein interactions

Introduction

The p16^INK4a locus on chromosomal band 9p21 is frequently impaired by a variety of genetic or epigenetic abnormalities in human oncogenic processes (Larsen, 1997; Ruas and Peters, 1998). The status of this locus is very unusual as it is expressed under the form of two alternative transcripts ( and ) that differ by one nucleotide change in their exon 2 open reading frame, leading to two structurally unrelated proteins: p16^INK4a and p19/p14^ARF (herein referred to as ARF). The transcript encodes the human p16^INK4a protein, the founding member of the INK4 family of cyclin D-dependent kinases (Chin et al., 1998). The transcript, first identified in human cells (Duro et al., 1995; Mao et al., 1995; Stone et al., 1995) has the capacity to encode the ARF protein whose existence was first reported in mouse (Kamijo et al., 1997) then, in human cells (Della Valle et al., 1997). Most interestingly, the murine p19^ARF was shown to block cell proliferation at the G1/S transition and also, at G2/M (Quelle et al., 1995). Since this seminal report, several important results provided important clues as to deciphering the function of ARF. Most significantly, ARF was shown to specifically interact with the mdm2 oncogenic protein (Weber et al., 1999; Zhang et al., 1998). This interaction that involves the central region of mdm2 and the N-terminal 37 amino acids of p19^ARF abrogates the capacity of mdm2 to block the transactivatory function of p53 (Stott et al., 1998). In addition the mdm2-ARF interaction stabilizes p53 by an uncompletely known mechanism that involves inhibition of its intrinsic E3 ubiquitine-ligase capacity which is required for p53 degradation (Honda and Yasuda, 1999) and cessation of its capacity to shuttle p53 from the nucleus to the cytoplasm where p53 is degraded in the proteasome (Tao and Levine., 1999). In addition, murine as well as human ARFs have been shown to possess nucleolus-addressing signal sequences (NoLS) that might be involved in sequestration of mdm2 in this organite, thus contributing to p53 stabilisation and/or prevention of p53 translocation to cytoplasm. In fact, NoLs are not located at the same position in murine and human proteins: residues 26-37 for murine ARF (Weber et al., 1999), residues 80-101 for human ARF (Zhang and Xiong, 1999; Rizos et al., 2000). A second NoLs has been identified in human ARF residues (residues 1-13, Rizos et al., 2000). This region is highly conserved in mouse and Monodelphis ARF but for these two species, the need of this domain for correct nucleolar localization has not been tested so far. In human ARF, this aminoterminal region also appears to be important for hdm2 binding (Rizos et al., 2000). To make things more complex, presence of NoLs in mdm2 and hdm2 appears to be required for efficient location of the ARF-mdm2 complex in nucleolus (Lohrum et al., 2000, Weber et al., 2000). Strikingly, the NoLS in mdm2 is normally concealed and its exposure most probably requires a conformational change induced by ARF (Lohrum et al., 2000).

Since interaction of ARF with mdm2 or (hdm2) is necessary for ARF to inhibit cell proliferation via p53, it has been postulated that ARF is part of a cell cycle regulatory pathway that contributes to stabilize p53 in response to a variety of oncogenic stresses. Indeed, ARF production appears to be stimulated by outgrowth signals generated by different oncogenic proteins such as c-myc (Zindy et al., 1998), mutated ras (Palmero et al., 1998), adenoviral E1A (De Stanchina et al., 1998), v-abl (Radfar et al., 1998), and also the transcriptional regulator E2F-1 that promotes entry in S phase by direct activation of a number of genes required for DNA replication (Bates et al., 1998). This explains that overexpression of E2F-1 which is tightly regulated under physiological conditions of growth, can activate the ARF-mdm2-p53 pathway resulting in cell growth and apoptosis (Dimri et al., 2000). Consistently, mouse embryonic fibroblasts (MEFs) from animals in which ARF has been invalidated do not stop growing anymore upon E2F-1 overexpression. Therefore, abnormal E2F-1 activity which is potentially oncogenic appears to be controlled by ARF.

The ARF-mdm2-p53 pathway appears to be distinct from the one that activates p53 in response to DNA damage (for a review, see Sherr, 1998). This observation is of interest as it suggests that ARF inactivation frequently observed in human tumors can coexist with p53 mutations, indicating that events inactivating the two genes do not proceed in a mutually exclusive fashion. Such a situation has been reported to occur in primary lung tumors in which both ARF and p53 were deleted and inactivated (Gazzeri et al., 1998). Moreover, experiments designed to modulate expression of p16^INK4a and p19^ARF in MEFs resulted in cessation of growth when ARF was reexpressed even in the absence of a functional p53 in these cells (Carnero et al., 2000). The conclusion of this study was that ARF has the capacity to negatively regulate cell growth by p53-dependent and p53-independent mechanisms.

Up to now, only mdm2 has been recognized as a true partner of ARF, leaving open the question of other partners. A first clue to answer this question was provided by the finding that ARF can shuttle from the nucleoplasm to the nucleolus, suggesting that it can be more or less tightly linked to proteins located in these two structures (Tao and Levine, 1999). A more direct indication came from a recently published study in which ARF was shown to interact with a number of discrete proteins on the basis of their coprecipitation by antibodies to ARF, but no further identification of these proteins was attempted (Kurokawa et al., 1999).

DNA topoisomerases I and II (Topo I and II) are essential enzymes in higher eukaryotes (for reviews, see Wang, 1996; Pommier et al., 1998). Topo I and II are nicking-closing enzymes that modulate DNA supercoiling along the cell cycle and relieve torsionnal stress during DNA replication, transcription and possibly repair. Topo I can also form intermolecular religation and has been associated with DNA recombination process (Christiansen and Westergaard, 1994). In addition to its catalytic activity on DNA, Topo I is also a specific kinase that phosphorylates serine-arginin rich (SR) splicing factors and therefore is able to modulate mRNA splicing (Tazi et al., 1997). Topo II is one of the major components of the nuclear matrix and act during chromosome segregation and cell cycle progression (Burden et al., 1998; Nitiss, 1996). Both Topo I and II are the target of clinically important anticancer agents (Beck, 1997).

Like many of the enzymes involved in the synthetic phase of DNA replication, DNA topoisomerases are cell cycle regulated. In mammalian cells, Topo II exists as two isoforms, (170 kDa) and (180 kDa), which are encoded by different genes (Tan et al., 1992). The expression of Topo II is cell-cycle dependent, reaching a maximum in G2/M phase (Kimura et al., 1994), through mainly post-transcriptional mechanism of change in mRNA stability, while the expression of Topo II is constant through the cell cycle (Goswami et al., 1996). Topo II plays also a major role during mitosis and promotes chromosomes individualization in G2 (Gimenez-Abian et al., 2000), and a specific Topo II-dependent G2 checkpoint, different from the G2-damage checkpoint has been demonstrated by the use of specific catalytic inhibitors of Topo II (Downes et al., 1994).

For Topo I, G1 or S phase specific transcription of the Topo I gene has been reported (Lee et al., 1995) and increase in both transcriptional and catalytic activity of Topo I were described in tissues induced for cell proliferation or in cancer cells (for a review, see Pommier et al., 1998).

The antitumor agent camptothecin specifically interfere with the catalytic cycle of Topo I by reversibly stabilizing an intermediary covalent complex between DNA and Topo I. During S phase, this ternary drug-enzyme-DNA complex produces an irreversible inhibition of DNA synthesis, double-stranded DNA breaks, cell cycle arrest in G2 phase and apoptosis (Pommier et al., 1998). The cytotoxicity is resulting from a DNA damage processing of the DNA lesions, during DNA replication. This model is supported by the fact that inhibitors of DNA synthesis abrogate campothecin-induced lethality.

On the other hand, a yeast Topo I mutant that mimics a catalytic inhibition of the enzyme also provokes a terminal G2-arrest phenotype in checkpoint-proficient cells (Megonigal et al., 1997). TRF4, a non essential yeast gene, is responsible for a synthetic lethal phenotype with top 1 null mutation. Double TRF4/top1 mutation produces a defect of chromosome condensation in the rDNA at mitosis that are independent from the DNA damage checkpoint (Castano et al., 1996). These studies further support the function of Topo I at a critical step during mitosis.

Finally, Topo I is a recombinase which can mediate illegitimate recombination during its critical reaction intermediate step corresponding to the DNA-Topo I complex (cleavable complex). Cleavable complexes could be formed either in the presence of the antitumor drug camptothecin or in the vicinity of DNA lesions (for a review see Larsen and Gobert, 1999). While formation of cleavable complexes are necessary for the onset of the DNA damage response, they are also potentially dangerous for the cell due to their recombinational properties, and need to be tightly regulated. This could be achieved by maintaining the enzyme level relatively constant or by limiting the stability of cleavable complex. Physical interaction between Topo I and the oncogene suppressor p53 and its modulation of Topo I catalytic activity corresponds to a double-swedge pathway able to stimulate Topo I activity (and its recombinant properties) and to degradate Topo I by the proteasome system (Gobert et al., 1996; Larsen and Gobert, 1999). In cancer cells with mutant p53, Topo I and p53 are constitutively associated, in contrast to wild-type p53 where association only take place during brief periods of genotoxic stress (Gobert et al., 1999).

Physical interaction between Topo II and wtp53 was recently reported (Yuwen et al., 1997; Cowell et al., 2000). Association between Topo II and p53 resulted in the stimulation of the catalytic activity of Topo II by enhancing the rate of ATP hydrolysis (Kwon et al., 2000). Unlike Topo II, both clivage and religation steps in the catalytic cycle of Topo I were stimulated by p53 (Gobert et al., 1996). P53 mutants presented a decreased or abolished stimulation of the in vivo activity of Topo II, compared to wtp53 (Cowell et al., 2000; Kwon et al., 2000). Therefore, wtp53 may contribute to the proper regulation of Topo II levels required in the G2 Topo II-dependent checkpoint.

Because initial observations had indicated that murine ARF stopped cycling cells at G1/S and G2/M, we decided to investigate proteins involved in DNA synthesis and mitosis. DNA topoisomerases are potential candidates for such interaction, because of their implication during G2/M and during DNA synthesis. In addition, Topo I and Topo II were found involved in the p53 pathway, as does ARF through its interaction with mdm2. We report here a specific interaction of ARF with human Topo I. This interaction takes the form of a physical association between ARF and Topo I and a stimulation of enzyme ability to relax supertwisted DNA substrates.

Results

ARF enhances the activity of Topo I but not that of Topo II

In view of the inhibitory effects of ARF on cell growth, we decided to investigate possible interactions of the protein on Topo I and II, two enzymes whose activities are required for proliferating cells to go through S phase to mitosis. In order to determine whether p14^ARF could modulate the function of Topo I, reconstitution experiments were carried out in the presence of purified human Topo I and recombinant GST-p14^ARF. As shown in Figure 1a, serial dilutions of Topo I induced a dose-dependent relaxation of supercoiled plasmid DNA, starting from complete DNA relaxation (lane 2, 60 ng Topo I) to nearly undetectable relaxation levels (lane 5, 2.2 ng enzyme). Upon addition of 250 ng of GST-p14^ARF to the reaction, stimulation of enzymatic activity could be detected at Topo I concentrations under which the enzyme was almost or totally inactive (compare lanes 4 and 5 to lanes 9 and 10 in Figure 1a). Under the same conditions, incubation of GST-p14^ARF in the absence of Topo I appeared to have no effect on the topological state of plasmid DNA (Figure 1a, lane 6). Control experiments were also performed with GST (Figure 1b). Neither GST alone (Figure 1b, lane 15) nor GST plus Topo I (Figure 1b, lanes 10-13) were found to stimulate the DNA relaxation catalysed by Topo I as did GST-p14^ARF (Figure 1b, lanes 6-9).

Since the stimulatory effect of p14^ARF was detected at 5-15-fold molar excess of the recombinant protein relative to Topo I, increasing amounts (3-250 ng) of GST-p14^ARF were incubated with constant concentrations (6.6 and 20 ng of Topo I (Figure 1c). By comparison with unstimulated Topo I (Figure 1c, lanes 2,3), the addition of GST-p14^ARF resulted in a dose-dependent stimulation of the DNA relaxation process (lanes 4-13). This effect was detectable with 9.2 ng of GST-p14^ARF added to 20 ng of Topo I (lane 10) and 3 ng added to 6.6 ng of enzyme (lane 13). In these conditions, the molar ratio between the two proteins was approximately equal to 1.

These data indicate that GST-p14^ARF stimulates the catalytic activity of Topo I and that this effect could be achieved for equimolar amounts of the two proteins.

A possible effect of GST-p14^ARF on Topo II was also examined on the specific DNA decatenation reaction catalysed by this enzyme (Figure 1d). Trypanosoma cruzi kinetoplast DNA (kDNA) was utilised for this assay as this substrate is constituted by a huge network of catenated DNA minicircles (mc) that do not migrate into a 2% agarose gel. Addition of GST-p14^ARF induced no effect on the decatenation reaction (Figure 1d, compare lanes 2-4 and 5-7). In view of these data, the stimulatory effect exhibited by ARF appeared to be limited to Topo I among these two types of Topoisomerases.

ARF has the ability to bind Topo I

The previous results raised immediately the possibility of a direct association between ARF and Topo I in order for ARF to exert its stimulatory activity. This question was addressed by infecting separately SF9 cells with recombinant baculoviruses carrying a p14^ARF cDNA insert (BacHARF) or a human Topo I cDNA insert (BacHTopo I). Because, in preliminary tests using non ionic detergents for cell lysis, ARF was found principally in non soluble material, sodium dodecylsulphate (SDS) was systematically added at final concentrations of 0.05 or 0.1% in order to release the protein in the soluble phase (Figure 2, compare lanes 1, 3, 4 to lanes 2, 5).

These conditions being fixed, SF9 cells were co-infected with both BacHARF and BacHTopo I recombinant viruses (Figure 3a,b). Immunoprecipitation was carried out with anti-ARF serum and the resulting immune complexes were analysed by Western blot for their Topo I content (IP/Western). The immunoblots were revealed by either a polyclonal rabbit serum to human Topo I (Topogen) or by an autoimmune serum from a patient with scleroderma (anti-hSc). A band running at the position of true Topo I was reproducibly found in extracts from cells co-infected with BacHARF and BacHTopo I baculoviruses (Figure 3a). This band was not detected in uninfected cells, in cells infected with wild type baculovirus (AcNPV), neither in cells infected separately with BacHARF or BacHTopo I. This result was observed whatever the nature of the anti-Topo I antibodies used.

This protocol was repeated by inversing the order of antibodies for performing IP/Western blot (Figure 3b). By using anti-hSc serum for the immunoprecipitation step, a band running at the position of ARF was found in the extracts from cells co-infected with both recombinant virus but not in extracts from cells infected by BacHARF alone or BacHTopo I alone.

The same approach was repeated with monoinfected SF9 cell lysates. Equal amounts of proteins from BacHARF and BacHTopo I infected cells were mixed and IP/Western blot experiments were carried out. Again, consistent with the previous results, each serum specific for one of the two proteins immunoprecipitated the other arguing in favor of a physical association between ARF and Topo I (Figure 3c,d).

Evidence for in vivo interaction between ARF and Topo I in human cells

293 cells (human kidney cells transformed with Adenovirus 5 DNA) contain abundant levels of endogenous ARF. This gave us the opportunity of looking for an interaction between the endogenous 293 ARF and Topo I proteins. When ARF immunoprecipitate from 293 cells were immunoblotted with antibody against Topo I, co-precipitation of both proteins succeeded indicating that the interaction occurred also between the two endogenous proteins (Figure 4a).

This experiment was repeated with HeLa cells that also express abundant amounts of ARF most presumably because of the absence of p53 in these cells. We looked for an interaction between the endogenous HeLa ARF protein and exogenous Topo I resulting from transient transfection of HeLa cells by a eukaryotic expression vector carrying HTopo I cDNA. Upon immunoprecipitation with anti-ARF serum, revelation of the blot with anti-Topo I serum (Topogen) showed reproducibly the presence of a faint band that was not detected by non immune serum (data not shown). This result confirms the data obtained from in vitro experiments. The faintless of the HeLa cell ARF band may be due to the particular localisation of the protein in HeLa cells (see Discussion).

In contrast, Saos-2 cells have been shown to contain nearly undetectable levels of endogenous ARF (Zang and Xiong, 1999 and our data). Co-transfection with both ARF and Topo I eukaryotic expression vectors (Figure 4b) also allowed to detect the interaction between ARF and Topo I. In that case, cell lysates were immunoprecipitated by anti-ARF serum and immunoblotted with anti-Topo I serum (Topogen). A band running at the position of Topo I was visible only in those samples from cells co-transfected with both genes (lanes 2 and 4). It should be noticed that in this experiment, transfection of ARF-HA or ARF-GFP cDNA alone were unable to immunoprecipitate endogenous Topo I (lanes 1 and 3). These results indicated that the status for ARF expression may be an important factor for ARF/Topo I interaction and will be discussed in light of the nuclear localisation of both proteins.

Simultaneous expression of ARF and Topo I proteins results in their localisation in nucleolus

A characteristic and biologically significant property of ARF is its propensity to relocalize its mdm2 partner into the nucleolus (Weber et al., 1999; Pomerantz et al. 1998). Because of the link between Topo I and ARF, we decided to examine the localization of Topo I expressed with and without ARF in Saos-2 cells. Saos-2 cells stably transfected with pcDNA3.1-Topo I plasmid encoding the human Topo I cDNA (referred herein as Topos cells) were transiently transfected with a GFP-ARF construct (see Materials and methods). Control consisting of Saos-2 cells transfected with the pEGFP-C1 plasmid generated as expected a green fluorescent staining pattern encompassing the whole cells (data not shown). In contrast, the fusion protein GFP-ARF appeared to be concentrated in round intranuclear structures that were demonstrated to be nucleoli (Figure 5a,b). More accurate inspection, revealed two distinct patterns of staining: first, rare small cell pairs or cells at close proximity from each other displaying numerous discrete stained foci (data not shown), second, cells with an intense staining of large nucleoli (Figure 5b). By referring to the distribution of nucleolar proteins, the former pattern is likely to be linked to early G1 phase cells (Kill, 1996). In Topos cells, the Topo I staining was distributed throughout the nucleoplasm and in the nucleoli. In the latter, the staining was heterogeneous with denser areas at the periphery. A very similar but fainter pattern was also found in original Saos2 cells, indicating that endogenous Topo I was present in these cells (data not shown). In Topos cells transfected with GFP-ARF, cells clearly showed a colocalization of both ARF and Topo I proteins as revealed by discrete yellow areas (Figure 5d). At higher magnification (Figure 5e-h), Topo I and ARF appeared to be colocalized at the periphery of the nucleolus, whereas no staining was detected in the central part of the organite. 3D projection of GFP fluorescence Z serie (Figure 5i and i') confirmed the colocalization of ARF and Topo I in the granular component as reported elsewhere (Lindstrom et al., 2000) and their absence in the fibrillar center.

These data are consistent with the existence of an ARF-Topo I interaction in vivo. Nevertheless we cannot exclude that they also reflect some attachment of both proteins to a common nucleolar structure.

Extracts from SF9 cells infected by ARF or/and Topo I recombinant viruses display a stimulation of the relaxation activity of Topo I

We tried to reproduce the in-vitro stimulatory effect of GST-p14^ARF by examining the Topo I relaxation activity in cell extracts from SF9 cells monoinfected by AcNPV, BacHARF, BacHTopo I or coinfected by BacHARF and BacHTopo I viruses (Figure 6). As expected, a DNA relaxation activity likely to reflect insect Topo I in SF9 cells was detected in AcNPV-infected cells (Figure 6a). In the presence of BacHARF, some slight increase (1.5-fold) in DNA relaxation was observed (Figure 6b). In BacHTopo I-infected cells, a 2.3-fold increase in DNA relaxation activity was found, that corresponded to the exogenous enzyme activity from the recombinant Topo I baculovirus (Figure 6c). Finally, when SF9 cells were co-infected by BacHARF and Topo I viruses at equal multiplicities of infection, DNA relaxation activity dramatically increased (8.4-fold) compared to that of extracts separately expressing either p14^ARF or Topo I (Figure 6d). The activities measured in the different cell extracts preparations were expressed as specific DNA relaxation activities (see Materials and methods and Table 1).

Similar experiment was performed in the presence of 0.1% SDS in the extract, reflecting the conditions used for immunoprecipitations (Table 1). The presence of SDS in the extracts had some impact as the Topo I DNA relaxation activity is reduced by 1.5-2-fold (AcNPV, BacHTopo I). In addition the relative stimulation of Topo I by ARF was still detected but reduced by nearly two-fold. The presence of SDS has a dramatic decrease effect on the Topo I activity measured for BacHARF alone and in these conditions no stimulation was observed for the endogenous insect Topo I. Therefore, in spite of the presence of the detergent, its influence on protein conformation and ARF-Topo I association was sufficiently conserved as to maintain the stimulation of the catalytic activity of human Topo I.

We next examined whether the stimulation capacity of ARF could be detected by mixing cell extracts from SF9 cells infected separately. Consistent with the previous data, the DNA relaxation activities found in extracts from cells infected by AcNPV, BacHARF and BacHTopo I alone were comparable to those of previous experiments (Figure 7a and see Table 2). Upon mixing of AcNPV and BacHARF in ice prior to determination of activity, a slight 1.8-fold stimulation of endogenous insect Topo I activity was detected, by comparison to the theoretical specific activity (this activity was deduced from measurements on monoinfected extracts and is referred to as expected specific activity). In contrast, mixing AcNPV and BacHTopo I extracts did generate a specific activity value close to the expected one. Finally, the combination of BacHARF and BacHTopo I extracts resulted in a 1.7-fold stimulation of the specific activity, compared to the expected calculations (Figure 7b and Table 2). Altogether, these results suggested that expression of p14^ARF in insect cells through baculovirus infection was able to stimulate the catalytic activity of recombinant Topo I, whatever they were co-expressed in the cells or put together by mixing cell extracts.

Discussion

In view of recent results, the ARF tumour suppressor protein appears more and more to regulate the cell cycle through p53-dependent and independent mechanisms, namely the so-called ARF-mdm2-p53 pathway and the second most likely corresponding to the Rb pathway. This involvement of ARF in intricated mechanisms that play a role in tumour acquisition makes the research of new partners of ARF particularly important. We report here that human ARF exhibits a capacity to interact in vitro and in vivo with Topo I, first by directly binding to the enzymatic protein and second by stimulating its DNA relaxation activity. In addition, we have found that ARF and Topo I proteins colocalize in a particular structure of the nucleolus, the granular component.

Several facts argue in favour of a specific ARF-Topo I association. Indeed, IP-Western blot experiments have clearly shown that ARF and Topo I coprecipitate from cell extracts of SF9 cells infected with baculoviruses containing cDNA inserts of each protein. Importantly, the complex was found whatever the sequence of antibodies used for performing the IP step. This complex was also present in mixtures of cell extracts from cultures separately infected with ARF and Topo I recombinant baculoviruses. Furthermore, our studies on mammalian cells achieved the same conclusion. The IP/Western Blot experiments on 293 cell extracts indicated that the ARF-Topo I complex was formed from endogenous proteins. In addition, the transient expression of Topo I in HeLa cells, that abundantly express endogenous ARF, resulted in coprecipitation of Topo I in the immune complexes. Also important is the fact that these results were obtained by use of serums of different origins, for example, polyclonal rabbit serum to human Topo I and autoimmune serum from a patient with scleroderma. We are now preparing ARF mutants that will allow to define the region(s) of the protein participating to the interaction.

The stimulatory effect of ARF on Topo I activity exhibited several features. First, it was dependent on ARF as assays with GST alone or wild type baculovirus did not stimulate Topo I activity, assays with purified Topo II preparations instead of Topo I revealed no activity of ARF on this related enzyme thus supporting the notion of a specific effect of ARF. Moreover, the fact that it was observed in reactions containing purified ARF and Topo I proteins strongly suggested that there was no need for another protein than ARF for achieving stimulation of the enzyme. This leads us to conclude that the stimulatory activity results from a physical interaction between the two proteins. The ARF protein mutants we are currently preparing should contribute to solve this point.

A preliminary study of some parameters of the relaxation reaction catalysed by Topo I showed that stimulation by ARF was detected for equimolar quantities of each protein. This stoechiometry can be advantageously compared to that of the p53-Topo I complex for which a p53/Topo I ratio of 20 is required for inducing stimulation of Topo I activity (Gobert et al., 1996). Incidently, it may be not fortuitous that p53 and ARF which belong to a common pathway also stimulate Topo I.

In addition, we were able to reproduce in vitro the enzymatic stimulation of Topo I by ARF in a more complex situation, i.e., corresponding to a cell extract containing both Topo I and ARF recombinant proteins. This result strengthens the specificity of such protein-protein interaction. Furthermore, it should be noticed that the enzymatic effect of the interaction could be detected by using the same conditions of detergent that were used for co-immunoprecipitations of both proteins. The interference of SDS in the enzymatic reaction process (see Tables 1 and 2) may reflect an underestimation of the protein/protein stoechiometry for the IP/Western experiments, thus explaining the absence of detection of the complex when Topo I is not overexpressed in Saos-2 cells (Figure 4b).

In Saos-2 cells transfected with Topo I alone, the protein appeared to be located inside the nucleoli and in the nucleoplasm, in agreement with other data in the literature (Guldner et al., 1986; Kill, 1996). For this reason, our results do not demonstrate that ARF is involved in a relocalization process of Topo I in the nucleolus. Such an ARF-mediated process has been shown to exist for mdm2 and a new partner of ARF, E2F-1 (Eymin et al., submitted). Nevertheless,confocal analysis of cotransfected Saos-2 cells clearly showed a colocalization of part of nucleolar ARF and Topo I in a peripheral structure of the organite that has been identified as the granular component (Lindstrom et al., 2000). It is tempting to speculate that this colocalization corresponds to ARF-Topo I complexes which have been detected by IP/Western blot experiments. The tight area of Topo I/ARF colocalization, conjugated with a destabilizing effect of SDS may also explain the need for an overexpression of both Topo I/ARF to be easily detected by IP/Western in our conditions of experiments. Other data in the literature have reported that nucleolar Topo I is mainly located in the fibrillar center of the nucleolus, where synthesis of rRNA takes place (Guldner et al., 1986; Kill, 1996). On the other hand, a more recent study indicated that Topo I showed a very prominent granular pattern inside the nucleoli (Meyer et al., 1997), in agreement with our results on this cell line.

Because the granular component is not involved in RNA transcription, it is possible that it is a site of accumulation of some protein components such as ARF and Topo I.

The HeLa cell clone used in our laboratory is unusual because ARF is not concentrated in the nucleolus but is dispersed throughout the nucleoplasm (the same observation has been noted elsewhere, Della Valle et al., 1997; Gazzeri, personal communication). This abnormal behaviour may result from a mutation in the ARF protein itself or from an intrinsic cell defect in this cell clone (for example the absence of a structural protein that prevents ARF from binding in the nucleolus). Assuming that the ARF-Topo I association occurs in nucleoli, this would explain the faint intensity of the Topo I band found in HeLa cell complexes compared to the relative abundancy of ARF in these cells.

Topo I was reported to bind, mainly through its N-terminal region, a large number of cellular or viral proteins involved in RNA splicing such as ASF/SF2, PSF /p54 (Rossi et al., 1996; Straub et al., 1998); in transcriptional regulation such as TBP, p53, SR/RING, nucleolin (Merino et al., 1993; Gobert et al., 1996; Haluska et al. 1999; Bharti et al., 1996); in replication such as LT-SV40 (Simmons et al.; 1996) and in heat shock response such as hsc70 (Ciavarra et al., 1994). Two of these proteins, hsc70 and nucleolin accumulate into the nucleolus. Among this list of Topo I interacting proteins only p53 and PSF/p54 were demonstrated to activate the enzymatic activity of Topo I, as does ARF.

Our current results provide no evident clue for understanding the biological significance of the ARF-Topo I interaction. Because of the localization of the two proteins in the nucleolus, it is tempting to speculate on a function in rRNA biosynthesis. Consistent with this notion, it is of interest that cell treatment by low doses of actinomycin not only blocks the RNA biosynthetic process but also induces translocation of ARF to nucleoplasm (Lindstrom et al., 2000). But a more general implication of ARF in reactions driven by Topo I should also be considered.

Materials and methods

Cell culture

Mammalian cells were cultured (37°C, 5% CO₂) in Dulbecco's modified Eagle Medium (Saos-2 and 293 cell lines) or in RPMI1640 (Hela cells) supplemented with 10% fetal bovine serum (Gibco-BRL) and antibiotics. Split ratio was 1 : 2 to 1 : 4, and medium was renewed twice a week.

The Sf9 subclone of Spodoptera fugiperda cells IPLB-Sf21-AE (Vaughn et al., 1977) was maintained in Grace's insect medium (Gibco-BRL) supplemented with Yeastolate and Lactalbumine hydrolysate at 3.3 g/l each, 10% fetal bovine serum-insect qualifed (Gibco-BRL) and Penicilline 5000 Ul+Streptomycine 5000 g (Gibco-BRL).

Construction of recombinant plasmids and baculoviruses

ARFwt was first generated by PCR from DeltaEB vector (Della Valle et al., 1997) containing the full-length ARF, the BamHI and XbaI sites were incorporated into the sens and antisens primers, ARFwt was then sequenced and cloned into BamHI and XbaI sites of the transfert vector pFastBac. The recombinant baculovirus expressing wild-type p14^ARF were constructed using the Bac to Bac baculovirus expression system (Gibco-BRL). The recombinant baculovirus expressing the Topo I gene has been previously described (Rossi et al., 1996).

In parallel, the same cDNA was cloned into the BamHI-XbaI sites of the pEGFP-C1 vector (Clontech), in frame with the C-terminus of green fluorescent protein. The cDNA expressing Topo I was cloned into the BamHI-EcoRI sites of the pcDNA3.1 vector (Invitrogen).

Transfections and virus infection

Twenty four hours before transfection, Saos-2 cells were subcultured and seeded in a 6-well plate (3´10⁵ cells/well). Transfection experiments were carried out using 0.5-1 g of DNA, whatever the plasmid used, and 15 l of Effectene reagent (Qiagen), in presence of serum. Selection of stable transfected clones was performed by maintaining the culture at a 300 ng/ml G418 concentration during 8 weeks. After amplification of surviving cells (referred herein as Topos clones), proteins were extracted and analysed by Western blot. Transient transfection of Topos or Hela cells were performed following the same protocol or alternatively 1´10⁵ cells were cultured in chamber slides (Lab-Tek system), and transfected with 0.4 g of DNA and 7 l of Effectene reagent.

Sf9 cells were infected for 72 h at a multiplicity of infection (MOI) of 20 pfu/cell. In coinfection experiments, cells were simultaneously infected by two recombinant baculoviruses at a total MOI ratios of 20 pfu/cell and with the respective MOI ratios of 20 : 0 (control single infections).

Sf9 and mammalian cell lysis and immunoprecipitation

After 72 h of simple infection (p14^ARF or Topo I) or coinfection (p14^ARF and Topo I) the Sf9 cells (10⁷) were washed 3´with PBS 1´lysed with 1 ml of cold lysis buffer (tris 10 mm pH 7.5, NaCl 120 mM, EDTA 1 mM, DTT 1 mM, 0.5% NP40, 0.05% or 0.1% SDS) supplemented with insect cells protease inhibitor cocktail (Pharmingen). The cells were incubated on ice for 30 mn and sonicated 3´7 s. They were then centrifugated at 10 000 r.p.m. at 4°C for 15 min. The supernatant was collected in a new microtube and incubated with a laboratory made antibody against p14^ARF (Della valle et al., 1997) or alternatively with a human sclerodermia serum with a specificity against Topo I overnight. Complexes precipitated with 20 l of protein-A agarose were washed four times with cold wash buffer (Tris 50 mM pH 8, NaCl 150 mM, Tween-20 0.1% and 1 mM of PMSF). Precipitates were separated on denaturing polyacrylamide gels and transferred to nitrocellulose. p14^ARF was detected by immunoblotting with the same antibody, Topo I was detected by immunoblotting with a rabbit polyclonal antibody against human Topo I (Topogen), and visualized by enhanced chemiluminescence as per the manufacturer's instructions (Amersham). Hela, 293 and Saos-2 cells lysis and immunoprecipitations were done as previously described.

Indirect immunofluorescence

Cells were harvested 48 h after transfection, and fixed 5 mn on ice with cold acetone. Mouse monoclonal primary antibody against human Topo I (NCL-TOPOIA from Novocastra) was diluted to 1/100 before incubation. Topo I was then detected with Alexa fluor 568 goat anti-mouse antibody diluted to 1/500 (Molecular probes).

Confocal imaging

The cells were examined by confocal laser scanning microscopy using a Biorad MRC 1024 equipped with a 15 mW argon-krypton gas laser. The confocal unit was attached to an inverted microscope (Olympus I´70). Maximal resolution was obtained with Olympus plan apo´60 water, 1.3 numerical aperture objective. Fluorescence images were generated sequentially to avoid spectral overlap of fluorescent emissions of the fluorochromes. GFP used to stain ARF was excited with the 488 nm blue line and green fluorescence emission of the dye was collected via a photomultiplier through a 522 nm pass band filter. The Alexa 568 fluorochrome used to reveal Topo I was excited with the 568 nm yellow line and red emission of the dye was collected via a photomultiplier through a 605 nm pass band filter. Double fluorescence images were further obtained by merging single fluorescence images. Non confocal transmission images were generated by transmitted light detector. Fluorescence signal collection, the image construction, and scaling were performed through the control software (Lasersharp 3.2; Bio-Rad). Optical sectionning of the nucleolar structure (Z serie) was driven by a Z-axis stepping motor and 3D reconstruction was further generated (Bitplane software).

Interaction assay

In simple infections experiments 10⁷ Sf9 cells were infected for 72 h at a multiplicity of infection (MOI) of 20PFU/cell by baculoviruses encoding wild type p14^ARF and Topo I. The cells were collected and lysed like discribed previously. The same amount of cell lysate (in term of proteins) were mixed together under agitation during 1 h at 4°C. Then each mixture was incubated with antibodies against p14^ARF or Topo I. Complexes were precipitated with 20 l of Protein-A agarose and analysed by immunoblotting. The mixture immunoprecipitated by the antibody against p14^ARF was immunoblotted with the antibody against Topo I and the mixture immunoprecipitated by the antibody against Topo I was immunoblotted with the antibody against p14^ARF The detection was made by ECL Western blotting detection system (Amersham) according to the manufacturer's instructions.

In coinfection experiments, 10⁷ Sf9 cells were simultaneously infected for 72 h at a total (MOI) ratios of 20 PFU/cell by two recombinant baculoviruses encoding wild type p14^ARF and Topo I. After infection the cells were divided into two parts, and after lyse they were incubated with the antibody against p14^ARF or Topo I. Complexes precipitated with 20 l of Protein-A agarose were, after separation on denaturing polyacrylamide gels, immunoblotted with the antibody against Topo I or p14^ARF respectivly.

Topoisomerase I stimulation

DNA relaxation assay: DNA relaxation was assembled in ice, in a final volume of 20 l containing 20 mM Tris-HCl pH 7.5, 120 mM KCl, 0.5 mM EDTA, 0.5 mM, DTT, 30 g/ml bovine serum albumine and 0.5 g supercoiled pBR322 DNA. Purified calf thymus Topo I and GST-p14^ARF or GST at different concentrations (as specified in the text) were added on ice to the reaction mixture and incubated at 37°C for 30 mn. Reaction was stopped on ice with 6 l of loading buffer (50 mM EDTA, 0.5% SDS, 0.1% bromophenol blue and 50% (v/w) sucrose). DNA samples were then electrophoresed in 1% agarose gels at 6 V/cm for 4 h at room temperature in TBE buffer. Gels were stained with ethidium bromide and photographed under UV light.

For the determination of Topoisomerase I relaxation activity in SF9 infected cell extracts, protein concentration was determined by the BioRad assay and serial dilutions of each extract were extemporary prepared in ice cold Topoisomerase buffer (20 mM Tris-HCl pH 7.5, 120 mM KCl, 0.5 mM EDTA, 0.5 mM DTT, 30 g/ml bovine serum albumine), then added to the DNA relaxation reaction mixture as source of enzyme.

Negative of the gel pictures were scanned with a Pharmacia ultroscan densitometer in order to quantify the relaxation reaction. In this way, we determined the minimal amount of protein extract (mg) necessary to relax 50% of the supercoiled DNA in the assay conditions, i.e one relaxation unit and calculated the specific activity of the extract.

For reconstitution experiments, SF9 infected cell extracts from ACNPV, BacHARF, or BacHTopo I were mixed on ice with an equal volume of Topoisomerase buffer or an equal volume of another extract giving the following combinations: ACNPV/buffer, BacHARF/buffer, BacHTopocl/buffer, ACNPV/BacHARF, ACNPV/BacHTopo I and BacHARF/BacHTopo I. Specific DNA relaxation activity was determined as described.

For the Topo II decatenation activity, reaction mixture (20 mM Tris-HCl pH 7.5, 120 mM KCl, 10 mM MgCl₂, 1 mM ATP, 0.5 mM EDTA, 0.5 mM DTT and 0.25 mg of T. cruzi kinetoplast DNA) was assembled on ice. Purified calf thymus Topo II and GST-p14^ARF were added to the reaction mixture to give a final volume of 20 ml and incubated at 37°C for 30 mn. The reaction was stopped and electrophoresed as described for DNA relaxation but in 2% agarose gels.

Acknowledgements

This work was supported by grants from the Association pour la Recherche contre le Cancer (ARC, contrat 5524), and from the Ligue Nationale contre le Cancer (Comités de la Vienne et de la Charente Maritime). L Karayan held a postdoctoral fellowship from la Ligue nationale contre le Cancer.

Bates S, Phillips AC, Clarke P, Stott F, Peters G, Ludwig RL and Vousden KH. (1998). Nature 395, 124-125. Article MEDLINE

Beck WT. (1997). In Principle of Antineoplastic Drug Development and Pharmacology, Schilsky Rl, Milano Ga and Ratain MJ (eds.). Dekker: New York.,

Bharti AK, Olson MO, Kufe DW and Rubin EH. (1996). J. Biol. Chem. 271, 1993-1997. MEDLINE

Burden DA and Osheroff N. (1998). Biochim. Biophys. Acta. 1400, 139-154. MEDLINE

Carnero A, Hudson JD, Price CM and Beach DH. (2000). Nature Cell Biol. 2, 148-154. Article MEDLINE

Castano IB, Brzoska PM, Sadoff BU, Chen H and Christman MF. (1996). Genes Dev. 10, 2564-2576. MEDLINE

Ciavarra RP, Goldman C, Wen KK, Tedeschi B and Castora FJ. (1994). Proc. Natl. Acad. Sci. USA 91, 1751-1755. MEDLINE

Chin L, Pomerantz J and DePinho RA. (1998). TIBS 23, 291-296. MEDLINE

Christiansen K and Westergaard O. (1994). J. Biol. Chem. 269, 721-729. MEDLINE

Cowell IG, Okorokov AL, Cutts SA, Padget K, Bell M, Milner J and Austin CA. (2000). Exp. Cell Res. 255, 86-94. Article MEDLINE

Della Valle V, Duro D, Bernard O and Larsen C-J. (1997). Oncogene 15, 2475-2481. MEDLINE

De Stanchina E, McCurrach ME, Zindy F, Shieh SY, Ferbeyre G, Samuelson AV Prives C, Roussel MF, Sherr CJ and Lowe SW. (1998). Genes Dev. 12, 2434-2442. MEDLINE

Dimri GP, Itahana K, Acosta M and Campisi J. (2000). Mol. Cell. Biol. 20, 273-285. MEDLINE

Downes CS, Clarke DJ, Mullinger AM, Gimenez-Abian JF, Creighton AM and Johnson RT. (1994). Nature 372, 467-470. MEDLINE

Duro D, Bernard O, Della Valle V, Berger R and Larsen C-J. (1995). Oncogene 11, 21-29. MEDLINE

Eymin B, Karayan L, Séité P, Brambilla C, Brambilla E, Larsen C-J and Gazzéri S. (2000). Oncogene (in press).

Gazzeri S, Della Valle V, Chaussade L, Brambilla C, Larsen C-J and Brambilla E. (1998). Cancer Res. 58, 3926-3931. MEDLINE

Gimenez-Abian JF, Clarke DJ, Devlin J, Gimenez-Abian MI, De la Torre C, Johnson RT, Mullinger AM and Downes CS. (2000). Chromosoma 109, 235-244. MEDLINE

Gobert C, Bracco L, Rossi F, Olivier M, Tazi J, Lavelle F, Larsen KA and Riou JF. (1996). Biochemistry 35, 5778-5786. Article MEDLINE

Gobert C, Skladanowski A and Larsen AK. (1999). Proc. Natl. Acad. Sci. USA 96, 10355-10360. MEDLINE

Goswami PC, Roti JL and Hunt CR. (1996). Mol. Cell. Biol. 16, 1500-1508. MEDLINE

Guldner HH, Szostecki C, Vosberg HP, Lakomek HJ, Penner E and Bautz FA. (1986). Chromosoma 94, 132-138. MEDLINE

Haluska P, Saleem A, Rasheed Z, Ahmed F, Su EW, Liu LF and Rubin EH. (1999). Nucleic Acids Res. 27, 2538-2544. MEDLINE

Honda R and Yasuda H. (1999). EMBO J 18, 22-27. Article MEDLINE

Kamijo T, Zindy F, Roussel MF, Quelle DE, Downing JR, Ashmun RA, Grosveld G and Sherr CJ. (1997). Cell. 91, 649-659. MEDLINE

Kill IR. (1996). J. Cell Sci. 109, 1253-1263. MEDLINE

Kimura K, Saijo M, Ui M and Enomoto T. (1994). J. Biol. Chem. 269, 1173-1176. MEDLINE

Kurokawa K, Tanaka T and Kato J. (1999). Oncogene 18, 2718-2727. MEDLINE

Kwon Y, Shin BS and Chung IK. (2000). J. Biol. Chem. 275, 18503-18510. MEDLINE

Larsen AK and Gobert C. (1999). Pathol. Oncol. Res. 5, 171-178. Article MEDLINE

Larsen C-J. (1997). Progress in Cell Cycle Research. Meijer L, Guidet S and Philippe M. (eds). Plenum Press: NY, USA., pp. 109-124.

Lee HH, Chiang WH, Chiang SH, Liu YC, Hwang J and Ng SY. (1995). Gene Exp. 4, 95-109.

Lindstrom MS, Klangby U, Inoue R, Pisa P, Wiman KG and Asker CE. (2000). Exp. Cell Res. 256, 400-410. Article MEDLINE

Lohrum MAE, Ashcroft M, Kubbutat MHG and Vousden KH. (2000). Nature Cell Biol. 2, 179-181. Article MEDLINE

Mao L, Merlo A, Bedi G, Shapiro GI, Edwards CD, Rollins BJ and Sidransky D. (1995). Cancer Res. 55, 2995-2997. MEDLINE

Megonigal MD, Fertala J and Bjornsti MA. (1997). J. Biol. Chem. 272, 12801-12808. MEDLINE

Merino A, Madden KR, Lane WS, Champoux JJ and Reinberg D. (1993). Nature 365, 227-232. MEDLINE

Meyer KN, Kjeldsen E, Straub T, Knudsen BR, Hickson ID, Kikuchi A, Kreipe H and Boege F. (1997). J. Cell Biol. 136, 775-788. MEDLINE

Nitiss JL. (1996). DNA Topoisomerase in DNA and DNA Damage Tolerance, Nickloff Ja and Hokstra MF (eds.). Humana, Clifton: NJ.,

Palmero I, Pantoja I and Serrano M. (1998). Nature 395, 125-126. Article MEDLINE

Pomerantz J, Schreiber-Agus N, Liégeois NJ, Silverman A, Alland L, Chin L, Potes J, Chen K, Orlow I, Lee HW, Cordon-Cardo C and DePinho R. (1998). Cell 92, 713-723. MEDLINE

Pommier Y, Pourquier P, Fan Y and Strumberg D. (1998). Biochim. Biophys. Acta 1400, 83-105. MEDLINE

Quelle DE, Zindy F, Ashmun RA and Sherr CJ. (1995). Cell 83, 993-1000. MEDLINE

Radfar A, Unnikrishnan I, Lee HW, DePinho RA and Rosenberg N. (1998). Proc. Natl. Acad. Sci. USA 95, 13194-13199. Article MEDLINE

Rizos H, Darmanian AP, Mann GJ and Kefford RF. (2000). Oncogene 19, 2978-2985. MEDLINE

Rossi F, Labourier E, Forné T, Divita G, Derancourt J, Riou JF, Antoine E, Cathala G, Brunel C and Tazi J. (1996). Nature 381, 80-82. MEDLINE

Ruas and Peters G. (1998). Biochim Biophys Acta. 1378, 115-177.

Simmons DT, Melendy T, Usher D and Stillman B. (1996). J. Virol. 222, 365-374.

Sherr CJ. (1998). Genes Dev. 12, 2984-2991. MEDLINE

Stone S, Jiang P, Dayananth P, Tavtigian SV, Katcher H, Parry D, Peters G and Kamb A. (1995). Cancer Res. 55, 2988-2994. MEDLINE

Stott FJ, Bates S, James MC, McConnell BB, Starbor M, Brookes S, Palmero I, Ryan K, Hara E, Vousden K and Peters G. (1998). EMBO J. 17, 5001-5014. Article MEDLINE

Straub T, Grue P, Uhse A, Lisby M, Knudsen BR, Tange TO, Westergaard O and Boege F. (1998). J. Biol. Chem. 273, 26261-26264. MEDLINE

Tan KB, Dorman TE, Falls KM, Chung TD, Mirabelli CK, Crooke ST and Mao J. (1992). Cancer Res. 52, 231-234. MEDLINE

Tao W and Levine AJ. (1999). Proc. Natl. Acad. Sci. USA 96, 3077-3080. Article MEDLINE

Tazi J, Rossi F, Labourier E, Gallouzi I, Brunel C and Antoine E. (1997). J. Mol. Med. 75, 786-800. MEDLINE

Vaughn JL, Goodwin RH, Tompkins GJ and McCawley P. (1977). In Vitro 13, 213-217. MEDLINE

Wang JC. (1996). Annu. Rev. Biochem. 65, 635-692. MEDLINE

Weber JD, Taylor LJ, Roussel MF, Sher CJ and Bar-Sagi D. (1999). Nature Cell Biol. 1, 20-27. Article MEDLINE

Weber JD, Kuo ML, Botner B, DiGiammarino EL, Kriwacki RW, Roussel MF and Sherr CJ. (2000). Mol. Cell. Biol. 20, 2517-2528. MEDLINE

Yuwen H, Hsia CC, Nakashima Y, Evangelista A and Tabor E. (1997). Biochem. Biophys. Res. 234, 194-197.

Zhang Y, Xiong Y and Yarbrough WG. (1998). Cell 92, 725-734. MEDLINE

Zhang Y and Xiong Y. (1999). Mol Cell. 3, 579-591. MEDLINE

Zindy F, Eischen CM, Randle D, Kamijo T, Cleveland JL, Sherr CJ and Roussel MF. (1998). Genes Dev. 12, 2424-2433. MEDLINE

Figure 1 GST-ARF recombinant protein stimulates DNA relaxation activity of purified calf thymus Topo I and does not affect kDNA decatenation activity of purified calf thymus Topo II. (a) Effect of GST-ARF (250 ng). Topo I at various concentrations (60, 20, 6.6 and 2.2 ng) was incubated with (lanes 7, 8, 9 and 10) or without (lanes 2, 3, 4 and 5) GST-ARF as indicated on the figure. Lane 1, control plasmid DNA, and lane 6, control GST-ARF without Topo I. (b) Effect of GST-ARF compared to GST. Topo I at various concentrations (180, 20, 2.2 and 0.7 ng) was incubated alone (lanes 2-5), or with GST-ARF (lanes 6-9), or with GST (lanes 10-13) as indicated on the figure. Lane 1, control plasmid DNA, lanes 14 and 15, control GST-ARF and GST, without Topo I. (c) Topo I relaxation in the presence of different concentration of GST-ARF. Twenty or 6.6 ng of Topo I were incubated in the presence of 250, 83, 27, 9.2, 3.1 or 0 ng of GST-ARF as indicated in the figure. Minimal stimulatory effect were observed in lanes 10 and 13, corresponding to equimolar ratio of the two proteins. (d) Effect of GST-ARF (250 ng). Topo II at various concentrations (60, 30 and 10 ng) was incubated with (lanes 5, 6 and 7) or without (lanes 2, 3 and 4) GST-ARF. Lane 1 control kDNA. GST-ARF does not activate Topo II decatenation

Figure 2 Determination of conditions for solubilizing ARF in cell extracts. All tests were performed on BacHARF infected SF9 cells. (1) supernatant of cell extract lysed in buffer containing 0,5% NP40. (2) pellet from the same extract dissolved in buffer containing 0.1% SDS. (3) supernatant of cells lysed in buffer containing 0.5% Triton X100. (4) supernatant of cells lysed in buffer containing 0.5% Na Deoxycholate. (5) supernatant of cells lysed in buffer containing 0.05% SDS. The front is located beyond the position of the 7 kDa marker

Figure 3 ARF and Topo I can be recovered in immune complexes with antibodies to either proteins. (a) 10⁷ Sf9 cells were coinfected with BacHARF and BacHTopoI for 72 h, cell lysates in 0.1% SDS were immunoprecipitated with Anti-ARF serum (73SA). Immune complexes were runned on polyacrylamide gels and transferred onto nitrocellulose. Revelation of the blot was done with Anti-Topo I (Topogen) serum. (b) Same conditions as in (a) but cell extracts were immunoprecipitated with Anti-Topo I from sclerodermia patients serum (Anti-hSc-Topo I) and immunoblotted with Anti-ARF serum (73SA). (c) 10⁷ SF9 cells were infected separately with either BacHARF or BacHTopo I. Equal amounts of proteins of Sf9 cell lysates in 0.05% SDS infected by BacHARF or BacHTopo I were mixed, immunoprecipitated with Anti-ARF serum (73SA) and immunoblotted with Anti-Topo I serum (Topogen). (d) Equal amounts of proteins of Sf9 cells lysates in 0.05% SDS infected by BacHARF or BacHTopoI was mixed, immunoprecipitated with Anti-hSc-Topo I serum and immunoblotted with Anti-ARF serum (73SA). U: uninfected cells; TopoI: BacHTopo I virus infected Sf9 cell extract; ARF: BacHARF virus infected Sf9 cell extract; ARF+TopoI: Sf9 cells coinfected with BacHARF virus and BacHTopoI virus (a and b) or mixtures of cell lysates sepatately infected (c and d). All lanes presented in c are from the same gel. Impaired migration of TopoI in lane 3 is due to abnormal conditions of electrophoresis in this lane (air bubble)

Figure 4 Interaction of ARF with Topo I is detected in human cell extracts. (a) Extracts of 293 cells prepared in the presence of 0.1% SDS were immunoprecipitated first with preimmune serum then with anti-ARF (73SA) serum. Resulting complexes were immunoblotted with anti-Topo I serum (Topogen). (b) Saos2 cells were transfected with a cDNA expressing the HA-ARF fusion protein (lanes 1, 5), or a cDNA expressing a GFP-ARF fusion proteins (lanes 3, 7), or cotransfected with one of the previous constructe plus a cDNA expressing Topo I (lanes 2, 4, 6, 8). Extracts in 0.1% SDS were immunoprecipitated with anti-ARF (73SA) serum (lanes 1-4) and rabbit preimmune serum (lanes 5-8). Resulting complexes were immunoblotted with anti-Topo I serum (Topogen)

Figure 5 Cellular localization of ARF and TopoI by confocal microscopy. Topos cells transiently transfected by GFP-ARF (see text) were analysed for localization of TopoI and ARF. (a) Phase contrast. Nucleoli are indicated by arrows. (b) Green-staining of GFP-ARF is detected only in nucleoli of one of the cells presented in a. (c) Same cells as in a, displaying TopoI staining in nucleoplasm and nucleoli (arrows point to nucleoli). (d) Same cells with superimposition of ARF and Topo I staining:interestingly, the colocalization is concentrated at the periphery of the nucleoli (granular component). (e-h) Topos cells transiently transfected by GFP-ARF were examined at higher magnification. (f) GFP-ARF staining. (g) TopoI staining. (h) Superimposition of ARF and TopoI staining. Notice the absence of signal in the central part of the nucleolus. (I and I'): reconstitution of the previous structure by Bitplane software

Figure 6 Topo I DNA relaxation activity in SF9 cell extracts infected by ARF and Topo I recombinant baculoviruses. Extracts from cells either monoinfected or co-infected by recombinant baculoviruses were prepared in the absence of SDS, then measured for DNA relaxation activity by serial dilution. Lanes 1, 2, 3, 4, and 5: corresponding to 1/30, 1/90, 1/270, 1/810 and 1/2430th extract dilution, respectively. Extract from SF9 cells infected by (a) AcNPV, (b) BacHARF, (c) BacHTopo I, and (d) BacHARF and BacHTopo I. Negative of the gel picture were scanned in order to quantify the relaxation reaction and specific Topo I DNA relaxation was calculated. Values were reported in Table 1

Figure 7 Topo I DNA relaxation activity in SF9 cell extract infected separately by AcNPV, ARF and Topo I baculoviruses and reconstituted by mixing in vitro. Each extract was prepared in the absence of SDS, then mixed with an equal volume of either Topoisomerase buffer (a) or another extract (b) as indicated in the figure, then measured for DNA relaxation activity by serial dilution. Lanes 1, 2, 3 and 4: corresponding to 1/90, 1/270, 1/810 and 1/2430th dilution of extract, respectively. Negative of the gel picture were scanned in order to quantify the relaxation reaction and specific Topo I DNA relaxation was calculated. Measured values and theoritical expected values were reported in Table 2

Table 1 Topo I DNA relaxation activity in SF9 cell extracts infected by ARF and Topo I recombinant baculoviruses prepared in the presence or absence of SDS

Table 2 Topo I DNA relaxation activity in SF9 cell extract infected separately by AcNPV, ARF and Topo I baculoviruses and reconstituted by mixing in vitro