Three origins of replication are active in vivo in the R plasmid RSF1040.

Replicating DNA molecules of RSF1040, a deletion derivative of the conjugative R plasmid R6K, are cleaved at a single site by the Eco RI restriction endonuclease. Microscopic analysis of Eco RI-cleaved RSF1040 replicative intermediates synthesized in vivo indicates that initiation of replication occurs at three unique sites, ori alpha, ori beta, and ori gamma. The relative frequencies of initiations at these three origins are different from those found in vitro.


Replicating
DNA molecules of RSF1040, a deletion derivative of the conjugative R plasmid R6K, are cleaved at a single site by the Eco RI restriction endonuclease.
Microscopic analysis of Eco RI-cleaved RSF1040 replicative intermediates synthesized in uiuo indicates that initiation of replication occurs at three unique sites, or&, oriJ3, and oriy. The relative frequencies of initiations at these three origins are different from those found in vitro.
The R plasmid R6K which confers resistance to ampicillin and streptomycin has several features that make it an attractive system for study. R6K is a 26-megadalton conjugative plasmid which replicates in Escherichia coli Kl2 cells as a multicopy pool as opposed to the rest of the conjugative plasmids, usually present in low copy number pools in E. coli (1). Nevertheless, R6K shares the ability of the other conjugative plasmids to replicate and be maintained in poZAmutants of E. co2i (2,3).
Generally, plasmid DNA replication is initiated at a unique site on the DNA molecule, although a few plasmids were shown to possess more than one replication origin (4)(5)(6). In uiuo, R6K and its deletion derivative RSF1040 showed initiation at two sites, aria and or&?, separated by a stretch of about 3900 nucleotides (2,7,8). Replication from either oria or or@ first proceeds unidirectionally to a specific terminus and then proceeds from the origin in the opposite direction to complete the replication process at the termination site T. These results suggested the existence of a specific termination site on the R6K genome (3,8).
Recently, Stalker et al. (9) and Kolter and Helinski (11) reported that they cloned and sequenced a 520-base pair region of R6K that included a functional origin of replication, but the position of this initiation site was not correlated with the location of the previously described oria and orip (7). I report in this communication that, by using very early replicative intermediates of RSF1040 DNA replicating in Co, it is possible to map a third replication initiation site on RSF1040 DNA which is located between aria and ori/% This third origin, which we named oriy, is used in uiuo at a much lower frequency than oria or ori/& (banding at a position very close to the covalently closed circular Form I peak) were selected in order to determine accurately the position of possible origin(s) of replication (Fig. 1).
Eco RI treatment of these molecules rendered linear molecules which, when examined in the electron microscope, showed an internal loop of replicated DNA and two unreplicated branches. In these molecules, approximate positions for origins of replication could be obtained by measuring the distance from one of the Eco RI-generated ends to the repli-  cation loop. Arbitrarily, we chose for measurement the Eco RI-generated end in which the distance of the unreplicated branch to the replication loop was shortest. All molecules of RSF1040 at early stages of replication could unequivocally be classified into two major subpopulations. Fig. 2 shows diagrammatically that in one subpopulation the replication loop was located at 21.4 f 1.3% of the total molecular length from the nearest Eco RI end, while the other group showed a replication loop at about 40.8 f 2.0%. Fig. 3, a, b,  There is also clear evidence of initiations occurring at a site 30.9 A 1.9% from the nearest Eco RI end. We have named this putative site oriy. Representative molecules in which initiation has occurred at oriy are shown in Fig. 3, d and (Table I).

DISCUSSION
An electron microscopic examination of RSF1040 replicating DNA revealed that replication may be initiated from either of two distinct origins, oria and or$ (7,8). Although the majority of RSF1040 molecules 'initiated replication from either oria or or@, some molecules were observed in which both origins were operating simultaneously (7).
Molecular cloning experiments (1 1) suggested the presence of an origin of replication near the junction of Hind111 fragments H4 and H9, but the position of this origin did not correspond to the positions of either ork or or$ as previously reported (7.8). When we became aware of Kolter and Helinski's (1 1) results, we reexamined the replicative properties of RSF1040 by using replicative intermediates which were at very early stages of replication. These molecules were cleaved with Eco RI and examined in the electron microscope. Although most of the initiations occurred at 21.4 f 1.3% and 40.8 f 2.0%, respectively, from one of the Eco RI ends, there is also very clear evidence of initiations occurring at a site 30.9 * 1.9% from the Eco RI end, suggesting the presence of an additional origin of replication which we named oriy. Presumably, in past in vivo experiments, oriy was overlooked because initiations at this origin are rare in vivo and molecules initiating from oria or or$ after replicating but for a short time would have masked this site. This is also a problem when one wants to determine directionality for molecules initiating replication at oriy. However, inspection of Fig. 2 suggests that, at least initially, replication from oriy must be counterclockwise, that is towards or$. The initial component of the sequential mode of replication exhibited by R6K and derivatives in vivo is counterclockwise from oria and clockwise from orip (3,7,8). It remains to be seen whether replication from oriy is bidirectional in vivo or unidirectional as is the case in vitro (12).
Our assignment of the y origin of replication would place it around the junction of fragments HindIII H4 and H9, while oria is located on HindIII fragment H4 and orip is located on HindIII fragment H2 near the junction with HindIII fragment H15 (see Fig. 4). The location of oriy at the junction of Bind111 fragments H4 and H9 suggests that oriy must be the origin cloned and sequenced in Helinski's laboratory (9,11).
It is interesting that initiation of DNA replication at oriy requires the T protein both in vivo (11) and in vitro,' and that initiation of DNA replication at oria or orip requires an intact sequence of Hind111 fragments H9 and H15 which encodes T protein (2). Despite these similarities, our findings indicate that the frequency of origin usage in vivo is quite different from that in vitro (12). In vivo, o r b is the origin used at highest frequency, while in uitro oriy and orip are the origins used at highest frequency. The reason for these differences in frequency of origin usage is not clear as yet. Lack of specific cellular components in the in vitro system could lead to a difference in selection of initiation origins. We are currently examining RSF1040 replication under a variety of physiological conditions to define the parameters that lead to selection of the origins of replication in RSF1040 DNA.