ABSTRACT
The process of strand exchange is considered to be the hallmark of DNA recombination. Proteins known to carry out such exchange are believed to operate via one or the other of two mechanisms. RecA-like proteins promote the formation of a three-stranded or triplex synaptic intermediate in which strand exchange occurs, whereas other proteins would allow the coordinated exonucleolytic degradation of one strand in the duplex DNA and its replacement by an invading strand of similar sequence and polarity. In view of properties ascribed to it, we have attempted to determine whether p53 belongs to one or the other of these groups of proteins. The in vitro assay used relies on a double-stranded (ds) oligonucleotide (oligo 1+2) and a single-stranded (ss) oligonucleotide (oligo 3), part of which is complementary to oligo 1. The data collected suggest that, under the conditions of the assay, oligo 1+2 undergoes partial denaturation; p53 then catalyzes renaturation of oligo 1 with oligo 3, rather than true strand exchange. Since p53 is not known for being able to `melt' DNA, it would seem unlikely that this protein would effect strand exchange in vivo without assistance from another, denaturing, protein.
In organisms ranging from bacteria to mammals, proteins have been found that are able to promote the kind of interaction between homologous DNA sequences which could be the initial step in DNA repair or recombination (1 ). The archetype of such proteins is the RecA protein of Escherichia coli (2 -9 ). The best characterized in vitro reaction between RecA and DNA is one in which, in the presence of ATP, RecA first polymerises onto ss DNA; the resulting nucleoproteic filament subsequently forms a joint structure with homologous ds DNA in which the three strands lie parallel; then, pairing occurs between the ssDNA (+)-strand initially present in the filament and the complementary or (-)-strand from the dsDNA; finally, unidirectional branch migration allows complete strand exchange, leading to the release of two DNA products, heteroduplex DNA and the (+)-strand originally included in the dsDNA (2 -8 ). The three-stranded intermediate, which was reported to be stable even after removal of the RecA protein (10 -12 ), is considered to be characteristic of strand exchange catalyzed by proteins of the RecA type (1 ,13 ), a family of proteins which includes the Rad51 (14 -16 ) and Dmc1 (17 ,18 ) proteins of Saccharomyces cerevisiae. Other proteins from several eukaryotic species have been shown to carry out strand transfer in vitro. However, these proteins have no sequence similarity to RecA, Rad51 or Dmc1, do not need ATP for carrying out strand exchange, and fail to give rise to a nucleoprotein filament (1 ,19 -23 ). Yet, they often have exonuclease-associated activity (1 ) which, in some instances at least, is not only essential for strand exchange but dictates the polarity of the reaction (1 ,23 ). Also, in the case of this second group of proteins, compelling genetic evidence for a role in recombination is so far lacking (1 ,23 ).
The p53 tumor suppressor (Fig. 1 A) is a eukaryotic protein known to be involved in DNA repair, at least indirectly (25 ,26 ). This transcriptional activator accumulates in cells in response to exposure to ionizing radiation or other DNA-damaging agents (25 ,26 ), and stimulates the activity of the GADD45 (DDT11) gene, which is believed to be involved in the response to DNA damage (27 -29 ). Aside from this indirect involvement in DNA repair, p53 is suspected to play another, more immediate, role in this process. Indeed, several groups have reported that p53 binds single-stranded DNA ends and catalyzes both DNA renaturation (30 -34 ) and strand exchange (31 ,34 ), in the absence of ATP. Recently, p53 has also been found to possess exonucleolytic activity (35 ). While DNA renaturation can be defined as the establishment of a duplex structure between two complementary ssDNAs, strand exchange consists in the displacement of a given strand [(+)-strand] in a dsDNA, by another strand of the same polarity. Hence, the (-)-strand exchanges a (+)-strand for another (+)-strand. Ideally, assessing whether a protein already known to facilitate DNA renaturation is also able to catalyze strand exchange requires the availability of (i) preparations of dsDNA totally free of ssDNA and (ii) conditions for strand exchange that preclude DNA denaturation. Otherwise, as will become clear from our results, there will be no other choice but to try and establish the distinction between true strand exchange and denaturation/renaturation. In this work, we have attempted to compare p53 with RecA, the archetype of strand exchange proteins, and E.coli single strand binding protein (SSB), a protein able to catalyze both DNA denaturation and DNA renaturation (36 ,37 ). While confirming the well-established DNA renaturing activity of p53 (30 -34 ), our results did not provide us with evidence of catalysis of strand exchange by p53.
Oligonucleotides were synthesized with a Gene Assembler Plus synthesizer (Pharmacia-LKB). Two 30mer oligonucleotides (oligo 1 and oligo 2) and a 60mer (oligo 3) were used in most of this work. Oligo 1 (5'-ATTCCTAAGCAGCCAAGTGTGACCAGTGTC-3') was labelled at its 5'-end (renaturation, strand exchange and exonuclease assays) or at its 3'-end (exonuclease assay; see ref. 35 ). Oligo 2 (5'-GACACTGGTCACACTTGGCTGCTTAGGAAT-3') is complementary to oligo 1 and was used to prepare a 30 bp duplex (oligo 1+2) used in strand exchange experiments. Oligo 3 (5'-TTTAGTTTAGTTTAG
RecA and SSB from E.coli were purchased from Pharmacia. All protein quantifications were carried out by the Bio-Rad protein assay, which relies on the method of Bradford (38 ).
Recombinant Autographa californica baculoviruses expressing either the wild-type (WT) or the 280-390 mutant murine p53 protein (39 ) were generously provided by Dr P.Tegtmeyer. These vectors code for proteins with hexa-histidine tags at their N-terminus, allowing purification through a metal affinity column (see below).
Spodoptera frugiperda cells (Sf9 strain) were grown as a suspension at 27oC, in Grace's medium (40 ) supplemented with 10% fetal calf serum. They were infected at a density of 2 × 106 cells per ml with recombinant baculovirus (~1 p.f.u. cell), diluted to 3 × 105 cells per ml, further incubated at 27oC and harvested 3 days later. After centrifugation at 2000 r.p.m. in a Sorvall RC-3 centrifuge at 4oC for 20 min, the pellet from one liter of cell suspension was resuspended in 20 ml of lysis buffer (50 mM Tris-HCl pH 8.0; 150 mM NaCl; 1% NP-40; 10% glycerol; 1 mM phenylmethylsulfonyl fluoride; aprotinin and leupeptin, 50 µg/ml of each; 1 mM benzamidin; 5 mM [beta]-mercaptoethanol) and left on ice for 30 min. After removal of cell debris (2000 r.p.m. for 15 min at 4oC in a lab top centrifuge), the nuclei were pelleted by centrifugation for 30 min at 20 000 r.p.m. and 4oC in a Sorvall SS34 rotor, and the supernatant (S100) was collected.
Proteins were purified on a nickel affinity column as already described (39 ), except that we used a Metal Chelating Sepharose Fast Flow matrix from Pharmacia instead of a Ni-NTA-Agarose column from Qiagen. Purified WT, 1-360 and 280-390 p53 proteins (39 ) were also kindly provided by Dr P.Tegtmeyer (Fig. 1 B) and used in strand exchange and renaturation assays.
Purified proteins were analyzed and quantified by SDS-PAGE (Fig. 1 B) according to Laemmli (42 ) and silver stained as described by Wray et al. (43 ).
Oligo 1+2 was made by annealing oligo 1 with oligo 2 (see `Templates') in a buffer containing 20 mM Tris-HCl pH 7.5, 10 mM MgCl2, 100 mM NaCl and 1 mM DTT (15 min at 94oC followed by 30 min at 55oC). The ds material was purified by electrophoresis through a 10% polyacrylamide gel and eluted from the gel slice in TE buffer (10 mM Tris-HCl pH 7.5; 1 mM EDTA). In a control experiment, the ds material was eluted in TC buffer (10 mM Tris-HCl pH 7.5; 10 mM CaCl2). In general, 0.1 ng of oligo 1+2 was mixed with 0.05 ng of oligo 3 (ratio of two duplexes 1+2 per ss oligo 3). Some experiments, however, were done at high DNA concentration and with a different ratio of ds to ss oligonucleotides, i.e., 6 ng of oligo 1+2 and 12 ng of oligo 3 (Fig. 3 ). Reactions were performed in 15 µl of p53 buffer (5 mM Tris-HCl pH 7.5; 10 mM KCl; 0.5 mM EDTA; 1.5 mM DTT; 1 mg/ml BSA; 3.7% glycerol; see ref. 31 ). Divalent cations were added to this buffer whenever appropriate (see Results). RecA was used as a control, and incubated with oligonucleotides in p53 buffer containing 4 mM ATP and an ATP-regenerating system [10 U/ml of creatine kinase (Boehringer) and 8 mM of phosphocreatine (Sigma)]. WT, mutant 1-360 and mutant 280-390 p53 proteins and SSB were tested in reaction mixtures including oligo 1+2 as well as oligo 3 [assays were carried out at protein concentrations, or protein-DNA mass ratios, that were comparable]. The protein was pre-incubated with oligo 3 for 10 min at 37oC, in p53 buffer with or without divalent cations, and the reaction started by adding oligo 1+2. The mixture was further incubated for 30 min at 37oC. SDS and EDTA were then added to a final concentration of 0.17% and 26 mM, respectively. The deproteinized material was subjected to electrophoresis through 10 or 15% polyacrylamide gels, which were dried and autoradiographed. WT and 280-390 mutant p53 proteins displayed essentially the same properties whether they had been purified in our laboratory or that of Dr P.Tegtmeyer; both preparations of 1-360 protein which we tested came directly from the latter laboratory.
Renaturation was performed as detailed for DNA strand exchange (see previous section), without preincubation however, and using only oligo 1 and oligo 3 (0.05 ng of each, thus with 2-fold molar excess of oligo 1) as substrates. Reactions were stopped and samples analyzed as stated in the previous section. Note that the amounts of oligos used in our assays were determined on the basis of results such as those displayed in Figure 4 .
The exonucleolytic activity of p53 was tested as already described (35 ) with either 3'- or 5'-labelled oligo 1 as substrate. Reaction products were analyzed by electrophoresis through 20% polyacrylamide gels containing 7 M urea.
Surprising among the results shown above was the mimicking of strand exchange by SSB. Indeed, strand exchange activity has never been claimed for this extensively characterized protein (9 ). Possibly, in our assay, SSB was actually denaturing oligo 1+2 (Fig. 3 , M band) in a reaction inhibited by Mg2+, and renaturation was occurring during or after deproteinization (see Materials and Methods). If such was actually the case, the concentration of oligonucleotides in our reaction mixtures was probably too high to allow an adequate characterization of a protein endowed with renaturing activity, such as p53.
We thus decided to look for conditions which would preclude spontaneous renaturation between the two oligos yielding band S in Figure 3 , i.e., oligo 1 (hot) and oligo 3 (cold). Constant amounts of oligo 1 (0.05 ng) mixed with variable amounts (from 0.025 to 2.4 ng) of oligo 3 were incubated at 37oC in either p53 buffer, or p53 buffer added with Mg2+, Ca2+ (not shown) or Na+, and the two expected radioactive species (F and S, Fig. 2 C) were separated by electrophoresis (Fig. 4 ). Two observations were made in this experiment. First, renaturation and/or duplex stability was rather poor in p53 buffer (see strong F band across the gel in Fig. 4 A), as compared with p53 buffer with added Mg2+ (Fig. 4 B); secondly, a mixture containing as little as 0.05 ng of oligo 1 and 0.05 ng of oligo 3 (vertical arrows, Fig. 4 ) was not subject to spontaneous renaturation unless the p53 buffer was added with 10 mM Mg2+ (or Ca2+, see legend of Fig. 4 ). Thus, in order to avoid spontaneous renaturation in subsequent experiments conducted in the presence of proteins, we decided to use 0.1 ng of oligo 1+2 and 0.05 ng of oligo 3 in three strand reactions, and 0.05 ng of each oligo 1 and oligo 3 in two strand reactions (see below). In so doing, we were actually adjusting our conditions to those already used by others working with p53 (31 ,34 ; see Materials and Methods).
With these lower concentrations of oligos, it now became clear that most of oligo 1+2 (species M) was denatured by SSB (Fig. 5 A, lane 3), oligo 1 thus migrating mostly as species F (Fig. 2 B, right), and not as a mixture of species S and M (Fig. 2 B, right) as before (Fig. 3 , lane 3). At the same reduced concentration of oligos however, RecA still behaved as if carrying out strand exchange, albeit at a somewhat lower efficiency (Fig. 5 A, lane 9). As to p53, it was now clear that it acted differently from either SSB or RecA. In amounts varying between 15 and 50 ng per reaction, p53 was quite effective in generating S material, but mostly, if not exclusively, at the expense of F material (Fig. 5 A, lanes 5 and 6). Such F material was also detectable in the absence of protein (Fig. 5 A, lane 1), unless Mg2+ cations were added to the buffer (Fig. 5 A, lane 2). We tentatively concluded from these observations that oligo 1+2 undergoes partial denaturation either during incubation in p53 buffer or before (see below), and that Mg2+ cations or low amounts of p53 allow renaturation yielding the S species (1+3, Fig. 2 B), p53 being more effective than Mg2+ in stimulating such renaturation. When p53 was used in amounts in excess of 50 ng however, the F band became undetectable, as if renaturation had been complete (Fig. 5 A, lanes 7 and 8); simultaneously however, the amount of S material appeared to level off or even decline (Fig. 5 A, lanes 7 and 8) as if, at high ratios of p53 to oligos, denaturation of oligo 1+2 was reduced or renaturation yielded a different ratio of M and S products (Fig. 2 B, right). See the Discussion for a detailed description of the mechanism which we believe accounts for the response to increasing amounts of p53.
In the next experiment, we decided to carry out three strand reactions (Fig. 6 A) and two strand reactions (Fig. 6 B) in the presence of a constant amount of p53 and of increasing concentrations of either of two divalent cations, Mg2+ and Ca2+. We already knew that, in the absence of protein, Mg2+ and Ca2+ exerted similar enhancing effects on renaturation and/or ds structure stability (Fig. 4 ). In the relatively straightforward two strand reaction (Fig. 6 B), 10 mM of each cation was found to have about the same effect on renaturation as 100 ng of protein in the absence of cation (Fig. 6 B, lanes 2, 3, 7 and 12), while no renaturation was detectable in the absence of p53 and either Mg2+ or Ca2+ (Fig. 6 B, lanes 1 and 11). In addition we noted that, regardless of Ca2+ concentration, the p53/Ca2+ combination was no more effective than either of its components alone (Fig. 6 B, lanes 7-10). Quite different results were obtained with the p53/Mg2+ combination: the higher the Mg2+ concentration, the weaker the S and F bands (Fig. 6 B, lanes 4-6), and the greater the occurrence of smaller molecular weight radioactive material (Fig. 6 B, lanes 4-6). The fact that the S band, while always weaker, seemed to disappear even more rapidly than the F band was consistent with oligo 1 and oligo 3 (Fig. 2 C) being separately subjected to degradation by a ss-specific exonuclease before they could renature (see below). The results of the three strand reaction (Fig. 6 A) were consistent with this conclusion. At high concentrations of Mg2+, the S band, but not the M band, virtually disappeared (Fig. 6 A, lanes 5 and 6). Thus, the Mg2+-dependent nucleolytic activity already detected in the two strand reaction appeared to have little effect on the ds template which already existed at the onset of the reaction (oligo 1+2, band M); in contrast, disappearance of the S band, and appearance of low molecular weight radioactive material (Fig. 6 A, lanes 4, 5 and 6), again suggested exonucleolytic degradation of oligo 1 and oligo 3, that of oligo 3 having presumably started during preincubation, before renaturation or strand exchange could even take place. Once more (Fig. 5 ), whether in the absence or the presence of divalent cation(s), p53 appeared somewhat more effective than either Mg2+ or Ca2+ alone in giving rise to an S band (Fig. 6 A, lanes 2, 3, 7, 8, 9, 10 and 12). The main conclusion from this experiment was that divalent cations added little to the effect of p53 (WT at least, see below) in either of the two reactions carried out, except for the uncovering by Mg2+ of a ss-specific nuclease activity which clearly was Mg2+-concentration dependent.
Deletion mutants of p53 were compared with the WT protein in three strand and two strand reactions, two different preparations of mutant proteins being tested simultaneously, preparations b and c for mutant protein 1-360, d and e for mutant protein 280-390 (Fig. 7 ). In the three strand assay, reactions run with Mg2+, Ca2+ or no divalent cation (and no protein) were compared with reactions in which p53 was acting in the absence of such cations (Fig. 7 A). It was clear from the data that the WT and the 280-390 mutant p53 proteins were at least as effective, if not more so, as divalent cations in prompting two effects: the disappearance of species F from, and the appearance of species S in, the reaction mixture (Fig. 7 A, compare lanes 2 and 3 with lanes 4, 7 and 8). In contrast, mutant 1-360 appeared clearly less active than the other two proteins in both respects (Fig. 7 A, lanes 5 and 6). Once again, the tight coupling of F species disappearance with S species appearance suggested that the two events were linked and probably reflected the renaturing activity of p53. In a second experiment, WT and mutant p53 proteins were compared for their ability to promote renaturation in the absence or presence of Mg2+ cations (Fig. 7 B). On this occasion, WT p53 proved clearly more active than Mg2+ ions in stimulating renaturation, actually achieving complete conversion of the F species into S species (Fig. 7 B, compare lanes 2 and 3). In the presence of WT p53 and Mg2+, both species disappeared, again as expected from extensive exonucleolytic degradation of both oligos 1 and 3 (Fig. 7 B, lane 8). Altogether different results were obtained with the mutant proteins. Renaturation was stronger with mutant 280-390 than with mutant 1-360 in the absence of Mg2+, and either mutant protein performed as well or better in the presence of cation as either Mg2+ or the protein alone (Fig. 7 B, lanes 4-7 and 9-12). These data indicate that while WT p53 is the only one of the three proteins tested with readily detectable exonucleolytic activity, mutant protein 280-390 has renaturing activity which is virtually as strong as that of WT protein, as already observed by others (34 ; see also Fig. 1 A).
As indicated in Figure 2 A, and also in Materials and Methods, the bulk of our strand exchange reactions were carried out with ds material consisting of two complementary oligos (1 and 2) of precisely equal lengths. We noticed that previous work on strand exchange by p53 had often resorted to complementary oligos differing in length by one, two or more residues (31 ,34 ). Conceivably, this could have markedly influenced the results. We thus decided to attempt strand exchange between oligo 3 and essentially double-stranded material, in which either oligo 1 or oligo 2 was shortened by two nucleotides, at either the 5'- or 3'-end (all four possible combinations were tried). The results obtained were no different from those obtained with fully double-stranded oligo 1+2 (not shown).
Mummenbrauer et al. (35 ) have described a 3' -> 5', Mg2+- dependent, single strand-specific exonuclease associated with purified p53, as well as with some deletion mutants of this protein. We, of course, wondered whether this activity could explain the degradation of single-stranded oligonucleotides which we had found to take place during reaction with WT p53 in the presence of Mg2+ but not Ca2+ (Figs 6 and 7 B). We thus tried to detect exonucleolytic activity in our preparations of proteins using precisely the conditions of Mummenbrauer et al. (35 ). Figure 8 shows the results obtained with WT p53 (designated Prep a in Fig. 7 ); our results could be superimposed on those of Mummenbrauer et al. (35 ) for the same protein. We also tested preparations of mutant proteins 1-360 and 280-390 for nuclease activity; while protein 1-360 possibly displayed a very weak activity, protein 280-390 appeared completely inactive (not shown). While the absence of a strong associated exonuclease activity with mutant protein 1-360 is at odds with the results of Mummenbrauer et al. (35 ), this data suggests that the exonucleolytic activity which we found associated with WT p53 is unlikely to be that of a contaminating nuclease. Actually, all three of the p53 proteins analyzed in Figure 7 had been purified by the same procedure, while the mutant proteins were used at somewhat higher molar concentrations (Fig. 7 legend) and yet only WT p53 appeared to possess exonucleolytic activity.
In this work, we resorted to an oligonucleotide assay (Fig. 2 A) which, according to reports from two laboratories (31 ,34 ), allowed the demonstration of strand exchange by p53. In this context, both SSB and RecA permitted us to make useful comparisons. SSB, a protein which like p53 exists as a tetramer in solution, is known to have multiple binding modes to DNA; its effect on the structure of oligonucleotides of the size used in this work was expected to be strongly influenced by Mg2+ ions (45 -47 ). Both SSB and RecA worked with oligonucleotides as we expected from the relevant literature, SSB being able to mediate denaturation (Fig. 5 A) but not renaturation (Fig. 5 B) in the absence of Mg2+ cations; in contrast, RecA proved able to catalyze strand exchange (Figs 3 and 5 A) but not denaturation (Fig. 5 A) in p53 buffer added with Mg2+ and ATP. It should be noted that our RecA-mediated reactions were all carried out in the absence of SSB. SSB is often added to the reaction mixture in which RecA-mediated strand exchange is carried out, even though the two proteins compete for binding to ssDNA, and strand exchange is known to occur in the absence of SSB (48 ,49 ). We omitted SSB when carrying out strand exchange with RecA because we found SSB to actually inhibit this reaction (D.Gendron, L.Delbecchi and P.Bourgaux, unpublished observations). Possibly, this effect was related to our use of oligonucleotides instead of DNA, which left SSB with little to achieve but impeding the binding of RecA (48 ,49 ). As to our use of a buffer containing ATP and Mg2+ but no SSB to demonstrate RecA-mediated renaturation, it was consistent with earlier work on DNA renaturation (50 ,51 ), some of which indicated that SSB is inhibitory to this reaction (44 ). Actually, renaturation mediated by RecA in the presence of Mg2+ (10 mM) and ATP was not obviously more extensive than that mediated by 10 mM Mg2+ alone (Fig. 5 B); this is perhaps not too surprising, since optimal DNA renaturation by RecA requires 30 mM Mg2+ (44 ), a concentration of cations which deviates significantly from that used in this work.
As alluded to in our Introduction, some proteins owe part of their strand exchange activity to being able to degrade rather than displace a given strand in a duplex DNA (1 ,23 ). Others (35 ) have described the association of exonucleolytic activity with p53 (WT or mutant) and our data (this work) are consistent with WT p53 being associated with such an activity. There is nothing in our data however that would support the notion that this activity would assist p53 in catalyzing strand exchange. In the three strand reaction (Fig. 2 A), enhanced strand exchange should translate into a greater occurrence of S component. Yet, at any of the Mg2+ concentrations tested, addition of this cation to the three strand reaction meant a decrease in the intensity of the S band (Fig. 6 A). Also, mutant p53 protein with little or no exonucleolytic activity (Fig. 7 B, lanes 11 and 12) proved almost as active as WT p53 in generating the S species during a three strand reaction (Fig. 7 A, lanes 4, 7 and 8). Finally, the Mg2+-dependent exonucleolytic activity associated with WT p53 appeared to preferentially degrade ss material (Fig. 6 , compare lanes 4, 5 and 6 between A and B), and thus would be expected to impede strand exchange rather than facilitate it; indeed, digesting oligo 3, rather than oligo 2 in oligo 1+2, would be equivalent to degrading the invading strand rather than the strand that should be displaced or somehow disposed (23 ).
Careful analysis of our data clearly indicates that p53 has renaturing activity which, in the two strand assay, results in the appearance of an S band. It would, however, be hard to argue that in the three strand assay, the same S band results from renaturation plus strand exchange, or strand exchange alone; actually, in this assay, the amount of S material generated in the presence of p53 appears to eventually match that of the F material detected in its absence (see, for instance, Fig. 5 A). Yet, we were initially puzzled by the observation that p53 was consistently better than Mg2+ in stimulating the generation of a S species, but only in the three strand reaction (Figs 5 A, 6 A and 7 A). We believe that this apparent discrepancy reflects a critical difference between the mechanism of action of Mg2+ and that of p53, which matches a critical difference between the two strand and the three strand assay. Presumably, Mg2+ acts by favoring ds over ss structure, while p53 acts primarily by binding to, and favoring renaturation of, ss material (31 ,32 ,52 ,53 ). Thus, in the three strand reaction where p53 is preincubated with oligo 3 (see Materials and Methods), p53-carrying oligo 3 would be expected to react readily with any complementary oligo added to the reaction mixture, such as oligo 1 arising from oligo 1+2 via spontaneous denaturation. Such denaturation does not come as a surprise, in view of the apparent instability of the ds structure (Fig. 4 A) in the p53 buffer employed here and elsewhere (31 ,34 ) to look for strand exchange between oligonucleotides. Thus, unlike Mg2+, p53 will favor the renaturation of oligo 3 with oligo 1 over that of oligo 1 with oligo 2 (Fig. 2 B). At high amounts of p53 however, oligo 3 might become saturated, such that excess protein would bind more and more to oligo 1 and/or oligo 2. Hence, regeneration of oligo 1+2 (species M) would compete with generation of oligo 1+3 (species S); this could be one reason for the amount of label in the S band leveling off beyond 50 ng of p53 per reaction (Fig. 5 A, lanes 7 and 8). In the two strand assay in contrast, bound p53 will necessarily be shared between oligo 1 and oligo 3, a requirement which may explain why sizeable generation of S material requires 50 ng of protein, rather than 15 ng of protein in the three strand assay (Fig. 5 , compare lanes 5 and 6 in A and B).
In conclusion, we find that in the three strand assay as carried out here, p53 does act as if promoting some displacement of oligo 2 by oligo 3. However, if our interpretation is correct, this displacement does not require more from p53 than the exercise of its renaturing activity.
This work was supported by a grant from the Medical Research Council of Canada. We thank Raymund Wellinger for critical review of an early version of this manuscript. We express our gratitude to Peter Tegtmeyer for discussion, constructive criticism and useful suggestions made during the writing up of this paper, and for the generous gift of baculoviruses encoding the p53 proteins, as well as samples of purified proteins. Text editing by Andrée Houle is gratefully acknowledged.
*To whom correspondence should be addressed. Tel: +1 819 564 5321; Fax: +1 819 564 5392; Email: ahoule@courrier.usherb.ca
REFERENCES