Role of the Double-Strand Origin Cruciform in pT181 Replication

Plasmid 46, 95–105 (2001) doi:10.1006/plas.2001.1535, available online at http://www.academicpress.com on

Role of the Double-Strand Origin Cruciform in pT181 Replication Ruzhong Jin and Richard P. Novick1 Molecular Pathogenesis Program, Skirball Institute, and Department of Microbiology and Department of Medicine, New York University School of Medicine, New York, New York 10016 Received March 15, 2001; revised June 2, 2001 pT181 is a small rolling-circle plasmid from Staphylococcus aureus whose initiator protein, RepC, melts the plasmid’s double-strand origin (DSO) and extrudes a cruciform involving IR II, a palindrome flanking the initiation nick site. We have hypothesized that the cruciform is required for initiation, providing a single-stranded region for the assembly of the replisome (R. Jin et al., 1997, EMBO J. 16, 4456–4566). In this study, we have tested the requirement for cruciform extrusion by disrupting the symmetry of the IR II palindrome or by increasing its length. The modified DSOs were tested for replication with RepC in trans. Rather surprisingly, disruption of the IR II symmetry had no detectable effect on replication or on competitivity of the modified DSO, though plasmids with IR II disrupted were less efficiently relaxed than the wild type by RepC. However, in conjunction with IR II disruption, modification of the tight RepC binding site IR III blocked replication. These results define two key elements of the pT181 initiation mechanism—the IR II conformation and the RepC binding site (IR III)—and they indicate that pT181 replication initiation is sufficiently robust to be able to compensate for significant modifications in the configuration of the DSO. © 2001 Academic Press

Palindromic elements are widespread in DNA and are often associated with regulatory functions. In double-stranded DNA, the key feature of these elements is their sequence symmetry, which enables them to serve as symmetric binding sites for dimeric regulatory proteins. At the same time, palindromic sequences are capable of cruciform extrusion driven by the free energy of supercoiling, and this has been readily demonstrated in vivo as well as in vitro by means of various secondary structure-specific reagents (see review by Lilley, 1989). Most of these reagents, however, target the singlestranded loop that is usually present at the cruciform tip and do not attack the double-stranded stem. For superhelical DNA, it is generally assumed that if the loop region becomes melted, cruciform extrusion will inevitably follow, as it relieves superhelical tension. In some cases, it has been possible to confirm the structure by means of cruciform-specific antibodies (biaFrappier et al., 1989), binding proteins

(Bianchi et al., 1989), or enzymes (Panayotatos and Fontaine, 1987) or by chemical or photocrosslinking (Cech and Pardue, 1976) within the stem. Though their existence in superhelical DNA is unquestioned, it has been rather difficult to identify specific biological functions for cruciforms as distinct from functions attributable to sequence symmetry. One area where cruciforms may be important is regulation of transcription. In several cases, cruciforms have been shown to modulate promoter function by altering the polymerase binding affinity (Horwitz and Loeb 1988). Another area is replication, in which cruciform structures may be important for assembly of the initiation complex (Zannis-Hadjopoulos et al., 1988; Noirot et al., 1990; Jin et al., 1997). For all known rolling-circle replicons, the nicking site in the leading or double-strand origin (DSO)2 must be melted prior to nicking because the initiator proteins have a strong preference for a single-stranded substrate. As the nick site is often at the center of a palindrome and as the DNA must be supercoiled, it is rea-

1

To whom correspondence and reprint requests should be addressed. Fax: (212) 263-5711. E-mail: [email protected]. nyu.edu.

2

Abbreviation used: DSO, double-strand origin.

95

0147-619X/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.

96

JIN AND NOVICK

sonable to assume that melting would be followed by extrusion of the corresponding cruciform (Noirot et al., 1990; Higashitani et al., 1994). In some cases, the DSO must be melted not only because the initiator requires a singlestranded substrate, but also to allow assembly of the replication complex—for example, pT181 DNA that has been nicked by the initiator and allowed to relax is not a substrate for in vitro replication (Jin et al., 1997), and we have proposed that the DSO cruciform in pT181 provides the required melted region for assembly of the replisome (Jin et al., 1997). In this report, we present the results of several studies designed to test the hypothesis that the pT181 DSO cruciform is required for the initiation of replication. We have found that the cruciform is not absolutely required but may be regarded as one key organizational feature of the DSO that is involved in initiation in addition to the adjacent initiator binding site, also a palindrome (IR III) but almost certainly not a cruciform. Any one of the two, but not both, may be modified without significantly compromising replication. These results suggest that the initiation mechanism possesses internal safeguards that enable it to function in the face of significant genotypic or environmental interference with its key organizational features. EXPERIMENTAL PROCEDURES Bacterial strains, plasmids, and growth conditions. Staphylococcus aureus wild-type strain RN450 (NCTC 8325) and its derivative SA1350, carrying the plaC1 mutation (lordanescu, 1983), have been used as hosts. PlaC1 is an RNA polymerase sigma factor (s70) mutant. In this mutant, transcription of the plasmid negative regulatory gene (cop) is greatly reduced, greatly increasing repC expression. Plasmids used are listed in Table 1. All strains were grown in CY broth (Novick and Brodsky, 1972) with vigorous aeration at different temperatures. Growth was monitored turbidimetrically using a Klett-Sumersen colorimeter with a green (540 nm) filter. RepC protein. N-terminal histidine-tagged RepC/C (homodimeric) protein was purified from S. aureus strain RN8601 containing

pRN6921 as previously described (Jin et al., 1996). DNA mutagenesis. Site-specific mutagenesis of the pT181 DSO IR II region by the polymerase chain reaction was carried out according to the method of Ito et al. (1991). Three common primers and the primers specific for the indicated mutations were used in the PCR. Mutagenized DNA fragments containing pT181 DSO were cloned into either pRN5101 (temperaturesensitive mutant of pE194) or directly back into pT181. Mutations were confirmed by DNA sequencing. Transformation and transduction. Protoplast transformation of S. aureus was carried out according to the method of Chang and Cohen (1979). Transduction was performed as described (Novick, 1967). Specific plasmid DNA relaxation assay. Reaction mixtures (20 ml) contained 10 mM Tris–HCl (pH 8.0), 100 mM KCl, 10 mM Mg(OAc)2, 5% ethylene glycol, 1 mM EDTA, equimolar amounts of Rep protein (RepC/C), and supercoiled plasmid DNA (pT181 and its DSO mutant derivatives). The mixtures were incubated at 32°C for different lengths of time and reactions stopped by the addition of 2ml 500 mM EDTA. Samples were resolved on 0.8% agarose gels containing 1 mg/ml ethidium bromide in TBE buffer. Gels were analyzed by scanning densitometry using an Alpha-innotech videoimager. Competition assays. Strains containing pSA5000 and pT181 DSO mutant derivatives were grown initially on doubly selective GL agar plates. These were used as starting cultures to inoculate broth cultures without antibiotic selection and grown for 3 h. Whole-cell lysates were prepared from these cultures and separated by agarose gel electrophoresis. Complementation assays. These assays made use of the temperature-sensitive vector pRN5101 to which the various mutant origins had been cloned. Replication at the restriction temperature was dependent on a functional pT181 origin and on Rep protein provided in trans by a co-resident plasmid. For electrophoretic analysis, cells were grown on GL plates, with plasmid-selective antibiotics, at

TABLE 1 Plasmids Used in This Study Plasmid

Phenotype Tcr Cmr Tcr Tcr Emr Emr Tsr Emr Tsr Cmr

pRN6949 pRN6972 pRN6973 pRN6974

Emr Tcr Tcr Tcr

PRN6411 pRN7027

Emr

pRN7028

Emr

pRN7029

Emr

pRN6921

Cmr

Inc3 oriC RepC Inc4 oriD RepD Cointegrate, pT181(cop620)::pE5, with oriC inactivated (RepC donor) Similar to pSA7542 except repD sequences substituted for repC (RepD donor) Temperature-sensitive (Tsr) mutant of pE194 pRN5101::pT181-Taql-C (pT181 origin cloned to pRN5101) pRN5101::pC221-Taql-E (pC221 origin cloned to pRN5101) pT181 with tetracycline resistance (tetK) gene replaced by chloramphenicol resistance gene; it retains wild-type pT181 replicon pRN5101::pT181-oriC with the distal arm of IR II in DSO substituted with GC-rich sequences (see Fig. 1) pT181, the distal arm of IR II in DSO substituted with GC-rich sequences (see Fig. 1) pT181, the distal arm of IR II in DSO substituted with AT-rich sequences (see Fig. 1) pT181, 4 nt adjacent to the distal arm of IR II in DSO substituted with sequences complementary to those adjacent to the proximal arm (see Fig. 1) pRN5101::pT181-oriC with the distal arm of IR III deleted (see Fig. 1) pRN5101::pT181-oriC with the left arm of IR II in DSO substituted with GC-rich sequences plus the distal arm of IR III deleted (see Fig. 1) pRN5101::pT181-oriC with the left arm of IR II in DSO substituted with AT-rich sequences plus the distal arm of IR III deleted (see Fig. 1) pRN5101::pT18-oriC with 4 nt adjacent to the distal arm of IR II in DSO substituted with sequences complementary to those adjacent to the proximal arm plus the distal arm of IR III deleted (see Fig. 1) pRN5548::P-bla-repC (repC gene under the control of b-lactamase promoter)

Reference Iordanescu (1976) Khan and Novick (1983) Iordanescu (1989) Iordanescu (1989) Novick et al. (1984) Carleton et al. (1984) Projan et al. (1985) Iordanescu and Surdeanu (1980) This work This work This work This work Gennaro et al. (1989) This work

THE pT181 DSO CRUCIFORM

pT181 pC221 pSA7542 pSA7461 pRN5101 pRN6397 pRN6385 pSA5000

Description

This work This work Jin et al. (1996)

97

98

JIN AND NOVICK

32°C overnight. These cultures were restreaked onto GL plates without antibiotics and duplicate plates were grown at 32 and 43°C, respectively, overnight. Whole-cell lysates were prepared from these cultures and separated by agarose gel electrophoresis. The ability of RepD to initiate the replication in trans of plasmids containing cloned target origins was tested in SA1350 (plaC1) in the presence of pSA7461 (RepD donor). The ability of RepC to initiate the replication in trans of plasmids containing cloned target origins was tested in RN450 in the presence of pSA7542 (low-level RepC donor) or pRN6921 (high-level RepC donor). KMnO4 analysis. Two micrograms of pT181 plasmid DNA or its DSO mutant derivatives was incubated with 10 pmol RepC/C protein in binding buffer (10 mM Tris–HCl, pH 8.0, 100 mM KCl, 0.1 mM EDTA) without Mg2⫹ in a total volume of 50 ml at 37°C. After 30 min, 2.5 ml 80 mM KMnO4 was added to each reaction for 1 min at 37°C. The reactions were stopped by the addition of 2.5 ml b-mercaptoethanol and plasmids recovered by using Qiaprep miniprep spin columns. NaOH (200 mM) was used to break the backbone of DNA at the sites of KMnO4 at-

tack and to denature the double-stranded plasmid DNA. Primer extension reactions were carried out using 32P 5⬘ end-labeled primers hybridized to either of the plasmid strands and 3 units of Sequenase for 15 min at 43°C. The reactions were stopped by the addition of a formamide–dye mixture and heated at 90°C for 2 min prior to denaturing gel electrophoresis. RESULTS Effects of modifying IR II. Since previous studies have shown that most modifications between the nick site and IR III, the RepC recognition site in the pT181 DSO, interfere with replication whether they affect the IR II palindrome or not (Gennaro et al., 1989; Wang et al., 1993) and because the proximal but not the distal arm is covered by the RepC footprint (Koepsel et al., 1986; Jin et al., 1996), we tested for an explicit requirement for the IR II cruciform by replacing five nucleotides in the distal arm, as shown in Fig. 1, with five noncomplementary nucleotides, either preserving the G⫹C content of the palindrome (pT181 G⫹C) or using only A or T residues (pT181 A⫹T). We also prepared a construct in which four nucleotides adjacent to the

FIG. 1. The pT181 DSO and its modifications. DNA sequence of the pT181 double-strand replication origin (DSO). The nucleotides in boxes are those that have been replaced in various combinations as indicated. ⌬44 represents the deletion of the distal arm of the IR III palindrome.

THE pT181 DSO CRUCIFORM

distal arm were replaced by four nucleotides complementary to those adjacent to the proximal arm, thus lengthening the potential hairpin by 4 bp (pT181 long stem). In the case of “G⫹C” substitution, amino acids Thr(24)–Gly(25) of RepC were mutagenized to Arg(24)–Gln(25); in the “A⫹T” substitution, Thr(24)–Gly(25) to Ile(24)–Leu(25); and in the “long” mutant, Ser(22)–Lys(23) to Cys(22)–Pro(23). These amino acid changes in the N-terminal region are not expected to cause any phenotypic changes in the initiator protein, since this region of the protein is nonessential. These constructs were tested for their DSO functions in two different sequence contexts. Initially, the modified DSO sequences were inserted into pRN5101, a thermosensitive derivative of staphylococcal plasmid pE194 (Novick et al., 1984), at this plasmid’s unique Clal site. Once it was found that the G⫹C substitution in the DSO (see Fig. 1) could support replication (see below), all three were transferred to the intact pT181 plasmid. The latter constructs were tested for their sensitivity to nicking/relaxation by purified RepC in vitro, and both sets were tested for their ability to replicate in vivo. The pRN5101 clones were tested at the restrictive temperature (43°C) in the presence of a co-resident plasmid expressing RepC; the pT181 constructs were tested in the homoplasmid state. As can be seen in Fig. 2, in an experiment with the pT181 constructs, disruption of the IR

99

II palindrome caused a significant reduction in in vitro relaxability. In the presence of RepC, the G⫹C- and A⫹T-substituted DSOs showed less relaxation after 10 min than the wild-type DSO after 2 min. In contrast, the long-stem construct showed a visibly greater relaxation response than the wild type. These results are consistent with the notion that ability to form the cruciform facilitates the RepC-induced nicking/relaxation of the DSO. However, disruption of the palindrome had no detectable effect on RepC-driven replication in vivo (Fig. 3) or in vitro (not shown) with either set of constructs. The stability and copy number of these constructs are essentially the same as those of the wild-type pT181 (not shown). These results mean that the IR II cruciform is not absolutely essential for pT181 nicking/relaxing or replication. They also mean that even a considerable reduction in nicking/relaxation activity in vitro did not cause any detectable reduction in replication frequency. We have previously observed that competition with the wild-type plasmid provides a sensitive test for the replication proficiency of a mutant plasmid. Thus, deletions affecting the pT181 replication enhancer as well as many mutations involving IR II, which have little or no detectable effect on replication proficiency in the homoplasmid state, cause a significant decrease in the ability of the mutant plasmid to compete with the wild type for a limited supply

FIG. 2. Effects of IR II modifications on RepC/C-dependent relaxation of pT181 and its derivatives. Reactions were performed with equimolar amounts of protein (20 ng) and supercoiled plasmid DNA (0.8 mg) and stopped at different time points (shown at top) with EDTA. Samples were resolved on 1% agarose gels containing 1 mg/ml ethidium bromide.

100

JIN AND NOVICK

FIG. 3. Effects of IR II modifications on in vivo competition of pT181 derivatives with pSA5000. Agarose gel electrophoresis of whole-cell lysate. Cell cultures containing pSA5000 and pT181 or one of its derivatives were grown on selective medium (a) and nonselective medium (b).

of RepC in vitro as well as in vivo (Gennaro and Novick, 1988; Gennaro et al., 1989). In order to test for competitivity, we replaced the DSO in the intact pT181 wild-type plasmid with each of the three IR II modifications in turn and tested the resulting constructs for competitivity in vivo with pSA5000, which contains the wild-type pT181 replicon but expresses resistance to chloramphenicol instead of to tetracycline (Iordanescu and Surdeanu, 1980). As shown in Fig. 3, and to our surprise, the G⫹C and A⫹T palindrome mutations had essentially no effect on competitivity with the wild-type DSO. This indicates that any role for the cruciform conformation of the DSO is dispensable when the other elements on the plasmid are intact. However, we did observe some decreased competitivity of the long-stem construct, presumably due to the disruption of the IR I sequence, which has previously been shown to cause impaired competition (Gennaro et al., 1989). Effects of IR II modification on RepC-Induced melting. As a corollary of the cruciform hypothesis, we have suggested that the extent of the palindrome would correspond to the melted region required for assembly of the initiation complex (Jin et al., 1997). Since the nucleotides within the cruciform stem are base-paired, they would not be sensitive to single-strand-specific reagents such as KMnO4, and so modification of

the loop nucleotides by KMnO4 (Jin et al., 1996) would not differentiate between melting of the entire palindrome and melting of only the loop region. The availability of the DSO modified by the insertion of a set of A⫹T residues enabled a test of the extent of RepC-driven melting in the DSO in the absence of cruciform formation. As shown in Fig. 4, the melting patterns of the three modified DSOs and the wild type were very similar. Since the A⫹T substitution had no significant effect on replication, it can be concluded that only the loop region need be melted for initiation of pT181 replication. Nevertheless, RepC interacts in a critical way with nucleotides in the proximal arm of IR II, such as G66 (see Fig. 1), since the mutation G66A profoundly decreases replication (Gennaro et al., 1989). It is noted that with the long-stem derivative there is significant KMnO4 sensitivity in the loop region in the absence of RepC, which we take to indicate spontaneous extrusion of the cruciform. In this configuration, the two flanking T residues (68 and 76) are not KMnO4-sensitive and are therefore base-paired, whereas they are KMnO4-sensitive in the RepC-melted DSO, suggesting that the actual RepC-melted region is 10 nt in length—the 6 loop nucleotides plus the 4 A and T residues at the top of the stem (see Fig. 1) .

THE pT181 DSO CRUCIFORM

101

FIG. 4. KMnO4 analysis of melting of pT181 and its derivatized DSOs. 2 mg supercoiled pT181 and its derivatives was incubated with an equimolar amount of RepC/C for 30 min and then treated with KMnO4 as described under Experimental Procedures. Reaction mixtures were separated on 6% polyacrylamide sequencing gels and autoradiographed. Primer-extension results showing the probing pattern of the top (leading) strand of pT181 DSO are presented.

Effects of modifying the Rep–DSO binding interaction. Origin recognition within the pT181 family is based on sequence recognition between a six-amino-acid segment of the protein and a 15-nucleotide palindromic region of the DSO, IR III (Novick, 1989; Wang et al., 1992; Dempsey et al., 1992), and is ordinarily absolutely plasmid-specific. Nevertheless, crossreactivity between two members of the family, pT181 and pC221, and their respective initiator proteins, RepC and RepD, is seen when the heterologous protein is provided in excess, either in vivo or in vitro (Iordanescu, 1989; Thomas et al., 1990; Zock et al., 1990). This phenomenon is illustrated in Fig. 5, in an experiment in which the pT181 (lanes 2 and 3) and pC221 (lanes 6 and 7) DSOs, cloned to pRN5101, were tested for replication in the presence of RepD in host SA1350, in which RepD is overexpressed. As can be seen, the high concentration of RepD supported replication of both of the intact DSOs (that of pC221 predictably better than that of

pT181). Lanes 3 and 4 show the results of this test using the pT181 (G⫹C) substitution, which shows no detectable replication at the restrictive temperature. Note that the lower band in lane 4 represents replication of the pRN5101 vector at the permissive temperature and does not represent amplification by RepD. This, then, is the first hint of any role of the IR II cruciform in replication; although these data do not directly distinguish between secondary structure and sequence, as shown above and in previous studies (Gennaro et al., 1989; Zhao et al., 1996), the sequence, per se, of the distal arm of IR II has no detectable role in replication. Since the plasmidspecific difference between the pT181 and pC221 DSOs is in IR III, the Rep protein binding site (Novick, 1989; Wang et al., 1992; Dempsey et al., 1992), it appears that poor binding by the heterologous initiators can be overcome by excess protein, permitting replication, when IR II is intact but not when its distal arm has been disrupted by mutation.

102

JIN AND NOVICK

FIG. 5. Complementation by RepD of pT181 and its modified DSOs. Agarose gel electrophoresis of whole-cell lysates from host SA1350. Due to the plaC1 mutation of the host, a high level of RepD is supplied in trans from pSA7461. Cell cultures were grown at 32 and 43°C as shown at top.

Similarly, deletion of the distal arm of the IR III palindrome, which has been shown previously to be dispensable with IR II intact (Wang et al., 1993), eliminated RepC-driven replication of the two mutant IR II elements, as shown in Fig. 6A, but, as expected, had no effect on replication of the intact DSO. Again, it is assumed that deletion of the distal arm of IR III weakens the binding of RepC to the DSO and that this can be compensated for by an intact IR II. To test whether excess homologous RepC can overcome the limitations of the IR III distal arm deletion (⌬44) plus IR II (G⫹C or A⫹T) double mutation, we have performed an in vivo replication experiment in host RN450 similar to the experiment shown in Fig. 6A, except with a RepC donor plasmid carrying the repC gene under control of the b-lactamase promoter. This promoter has a high transcriptional level and thus the intracellular RepC concentration is much higher than that in the experiment shown in Fig. 6A. The results shown in Fig. 6B indicate that higher RepC concentration in the cell can support replication of the (⌬44)-(G⫹C) but not the (⌬44)-(A⫹T) double mutant. This result is surprising because one would expect just the opposite since the AT-rich sequence is assumed to facilitate melting of the DSO. However, this may suggest that the GC-rich sequence of IR II has a specific role in the initiation of pT181 replication.

DISCUSSION In this study, we have continued a long-term analysis of the structural elements of the pT181 DSO that are necessary for RepC-driven initiation. In earlier studies (Gennaro et al., 1989; Wang et al., 1993; Zhao et al., 1996), it has been found that an inconstant palindrome, IR I, distal to IR II, is not required for DSO function in the pT181 family, that the proximal arm and central part of IR III are required, and that deletions up to position 86, distal to IR II, are compatible with normal replication. Additionally, only one of several point mutations within IR II significantly interfered with replication, though several of the mutations in this region caused a severe reduction in competitivity. As noted, we have assumed that melting of the IR II loop, which is induced by RepC on supercoiled DNA, would be followed by extrusion of the IR II cruciform, driven by the free energy of superhelix formation. Although we have never demonstrated the cruciform directly, we have no reason to doubt this. The absolute conservation of the IR II sequence and potential structure coupled with the lack of symmetrical binding of RepC to the palindrome and with RepC-induced melting of the DSO, presumably followed by cruciform extrusion, suggested that the cruciform would be an essential feature of the initiation complex and might represent the means of providing a melted region large enough for assembly of the replication complex (Jin et al., 1997). Results

THE pT181 DSO CRUCIFORM

103

FIG. 6. Complementation by RepC of pT181 and its modified DSOs. Agarose gel electrophoresis of whole-cell lysate from host RN450. (A) Low level of RepC is supplied in trans from pSA7542. (B) High level of RepC is supplied in trans from pRN6921. Cell cultures were grown at 32 and 43°C as shown at top.

described here suggest that the situation is rather more complex than initially envisioned. Clearly, the actual cruciform is not absolutely required for replication—which suggests that the single-stranded region, 10 nt, generated by RepC-induced melting per se is sufficient for assembly of the replication complex. The intact palindrome is, however, required when the primary binding of the initiator to its recognition site, IR III, is suboptimal. We propose that under such conditions, the protein does not bind with sufficient strength to melt the palindrome loop fully, but that it destabilizes the loop region sufficiently to drive cruciform extrusion, permitting nicking to occur. We also note that the elimination of negative supercoiling of plasmid DNA

eliminated the ability of RepD to catalyze initiation from the pT181 DSO but not from the pC221 DSO (Thomas et al., 1990). Here, the substrate had to be covalently closed so that there would have been a gaussian distribution of superhelical density centering on zero. Since RepC does not melt linear DNA (Jin et al., 1996), it can be assumed that it does not melt fully relaxed circular DNA either. It is thus our view that only those molecules with a significant number of negative superhelical turns could have served as substrates here. These observations together define three key features of the pT181 initiation mechanism—the IR II cruciform, the RepC binding site (IR III), and the superhelical density of the DNA. Any one of

104

JIN AND NOVICK

these, but not two, can be substantially modified without seriously compromising the initiation process. In other words, initiation can take place if at least two of the three critical elements of the DSO are intact, and we suggest but have not proven directly that one of these elements is the actual IR II cruciform. It has been shown that complementarity between the two arms of IR II enhanced the ligation of nicked circles by RepC in vitro (Jin et al., 1997; Thomas, 1997), which is expected to happen during termination of pT181 replication. Our in vitro relaxation experiment also indicated a slower transition from the open circular to the relaxed form for the mutant plasmids with IR II disrupted (Fig. 2), presumably owing to the loss of IR II complementarity. Therefore, it is possible that interference with termination is responsible for the observed replication defects. However, it has been demonstrated that the distal arm of IR II is not essential for termination (Zhao et al., 1997). As noted, the copy number and the competitivity of the IR II mutants are the same as those of the wild-type pT181. Moreover, we have never observed plasmid multimer formation either in vivo or in vitro, which would be expected if termination were inefficient, nor have we observed the accumulation of a complex relaxed form of the plasmid, which would be expected if termination were blocked. Thus, we consider it very unlikely that the incapability of (⌬44)-IR II double mutants to replicate is due to defective termination. The conclusion that the initiation of replication is compromised only when two of the three parameters discussed above are modified led us to analyze in more depth the nature of the replication block under such circumstances, testing the hypothesis that there would be a demonstrable defect in the ability of RepC to configure the DSO for initiation. An additional surprise, therefore, was that the DSO derivatives lacking both the distal cruciform stem and the distal arm of IR III (⌬44) are melted in vitro approximately as well as the intact DSO, and they have about the same decrease in the efficiency of RepC-induced nicking and relaxation in vitro as the derivatives lacking only the distal stem (data not shown). However, they are profoundly defective in RepC-initiated replication, especially

when the intracellular RepC level is not in excess (Fig. 6A). One simple interpretation for this is that Rep-induced melting of the DSO coupled with nicking and relaxation is not sufficient for replication, and some subtle factors that have yet to be identified are required for normal replication. It is also possible that strong binding of RepC to the DSO may be required for the recruitment of one or more replication factors. This idea is consistent with the observation that a high concentration of RepC can enable replication from the (⌬44)-(G⫹C) but not the (⌬44)-(A⫹T) double mutant (Fig. 6B), which implies that high GC content in IR II may play a role in recruiting such factors. Another possibility is that some in in vivo specific and/or nonspecific DNA binding proteins, which are not present in our in vitro assays, might bind to the plasmid replication origin and prevent RepC from interacting efficiently with the DSO and initiating replication under the above-discussed suboptimal conditions. ACKNOWLEDGMENT This work was supported by NIH Grant RO1-GM14372 to R.P.N.

REFERENCES biaFrappier, L. Price, G., Martin, R., and Zannis-Hadjopoulos, M. (1989). Monoclonal antibodies to cruciform DNA structures. J. Mol. Biol., 193, 751–758. Bianchi, M., Beltrame, M., and Paonessa, G. (1989). Specific recognition of cruciform DNA by nuclear protein HMGI. Science 243, 1056–1059. Carleton, S., Projan, S. J., Highlander, S. K., Moghazeh, S. and Novick, R. P. (1984). Control of pT181 replication II. Mutational analysis. EMBO J. 3, 2407–2414. Cech, T. R., and Pardue, M. L. (1976) Electron microscopy of DNA crosslinked with trimethylpsoralen: Test of the secondary structure of eukaryotic inverted repeat sequences. Proc. Natl. Acad. Sci. USA 73, 2644–2648. Chang, S. and Cohen, S. N. (1979). High frequency of transformation of Bacillus subtilis protoplasts by plasmid DNA. Mol. Gen. Genet. 168, 111–115. Dempsey, L. A., Birch, P., and Khan, S. A. (1992). Six amino acids determine the sequence-specific DNA binding and replication specificity of the initiator proteins of the pT181 family. J. Biol. Chem. 267, 24538–24543. Gennaro, M. L., Iordanescu, S., Novick, R. P., Murray, R. W., Steck, T. R., and Khan, S. A. (1989). Functional organization of the plasmid pT181 replication origin. J. Mol. Biol. 205, 355–362. Gennaro, M. L., and Novick, R. P. (1988). An enhancer of DNA replication. J. Bacteriol. 170, 5709–5717.

THE pT181 DSO CRUCIFORM Higashitani, A., Greenstein, D., Hirokawa, H., Asano, S. and Horiuchi, K. (1994). Multiple DNA conformational changes induced by an initiator protein precede the nicking reaction in a rolling circle replication origin. J. Mol. Biol. 237, 388–400. Horwitz, M. and Loeb, L. (1988). An E. coli promoter that regulates transcription by DNA superhelix-induced cruciform extrusion. Science 241, 703–705. Iordanescu, S. (1976). Three distinct plasmids originating in the same Staphylococcus aureus strain. Arch. Roum. Pathol. Exp. Microbiol. 35, 257–264. Iordanescu, S. (1983). Staphylococcus aureus chromosomal mutation specifically affecting the copy number of Inc3 plasmids. Plasmid 10, 130–137. Iordanescu, S. (1989). Specificity of the interactions between the Rep proteins and the origins of replication of Staphylococcus aureus plasmids pT181 and pC221. Mol. Gen. Genet. 217, 481–487. Iordanescu, S. and Surdeanu, M. (1980). Complementation of a plasmid replication defect by autonomous incompatible plasmids in Staphylococcus aureus. Plasmid 4, 1–7. Ito, W., Ishiguro, H. and Kurosawa, Y. (1991). A general method for introducing a series of mutations into cloned DNA using the polymerase chain reaction. Gene 102, 67–70. Jin, R., Fernandez-Beros, M., and Novick, R. P. (1997). Why is the initiation nick site of an AT-rich rolling circle plasmid at the tip of a GC-rich cruciform? EMBO J. 16, 4456–4466. Jin, R., Zhou, X. and Novick, R. (1996). The inactive pT181 initiator heterodimer, RepC/C*, binds but fails to induce melting of the plasmid replication origin. J. Biol. Chem. 271, 31086–31091. Koepsel, R. R., Murray, R. W. and Khan, S. A. (1986). Sequence-specific interaction between the replication initiator protein of plasmid pT181 and its origin of replication. Proc. Natl. Acad. Sci. USA 83, 5484– 5488. Lilley, D. (1989). Structural isomerization in DNA: The formation of cruciform structures in supercoiled DNA molecules. Chem. Soc. Rev. 18, 53–83. Noirot, P., Bargonetti, J., and Novick, R. P. (1990). Initiation of rolling circle replication of pT181 plasmid: Initiator protein enhances cruciform extrusion at the origin. Proc. Natl. Acad. Sci. USA 87, 8560–8564. Novick, R. (1967). Properties of a cryptic high-frequency transducing phage in Staphylococcus aureus. Virology 33, 155–166.

105

Novick, R., Adler, G., Projan, S., Carleton, S., Highlander, S., Gruss, A., Khan, S. and Iordanescu, S. (1984). Control of pT181 replication I. The pT181 copy control function acts by inhibiting the synthesis of a replication protein. EMBO J. 3, 2399–2405. Novick, R. P. (1989). Staphylococcal plasmids and their replication. Annu. Rev. Microbiol. 43, 537–565. Novick, R. P. and Brodsky, R. J. (1972). Studies on plasmid replication: Plasmid incompatibility and establishment in Staphylococcus aureus. J. Mol. Biol. 68, 285–302. Panayotatos, A. and Fontaine, A. (1987). A native cruciform DNA structure probed in bacteria by recombinant T7 endonuclease. J. Biol. Chem. 262, 11364–11368. Projan, S. J., Kornblum, J., Moghazeh, S. L., Edelman, I., Gennaro, M. L. and Novick, R. P. (1985). Comparative sequence and functional analysis of pT181 and pC221, cognate plasmid replicons from S. aureus. Mol. Gen. Genet. 199, 452–464. Thomas, C. D. (1997). A novel role for the inverted complementary repeat. Biochem. Soc. Trans. 25, S648. Thomas, C. D., Balson, D. F. and Shaw, W. V. (1990). In vitro studies of the initiation of staphylococcal plasmid replication. J. Biol. Chem. 265, 5519–5530. Wang, P.-Z., Projan, S. J., Henriquez, V. and Novick, R. P. (1992). Specificity of origin recognition by replication initiator protein in plasmids of the pT181 family is determined by a six amino acid residue element. J. Mol. Biol. 223, 145–158. Wang, P.-Z., Projan, S. J., Henriquez, V. and Novick, R. P. (1993). Origin recognition specificity in pT181 plasmids is determined by a functionally asymmetric palindromic element. EMBO J. 12, 45–52. Zannis-Hadjopoulos, M., Frappier, L., Khoury, M. and Price, G. (1988). Effect of anti-cruciform DNA monoclonal antibodies on DNA replication. EMBO J. 7, 1837–1844. Zhao, A. C. and Khan, S. A. (1996). An 18-base-pair sequence is sufficient for termination of rolling-circle replication of plasmid pT181. J. Bacteriol. 178, 5222–5228. Zhao, A. C. and Khan, S. A. (1997). Sequence requirements for the termination of rolling-circle replication of plasmid pT181. Mol. Microbiol. 24, 535–544. Zock, J. M., Birch, P. and Khan, S. A. (1990). Specificity of RepC protein in plasmid pT181 DNA replication. J. Biol. Chem. 265, 3484–3488. Communicated by D. Chattoraj