Articulo Micro

Plasmid 55 (2006) 126
www.elsevier.com/locate/yplas
Review
Plasmid R1Replication and its control

Kurt Nordstrom
Department of Cell and Molecular Biology, Biomedical Center, Uppsala University, P.O. Box 596, S-751 24 Uppsala, Sweden
Received 22 April 2005, revised 4 July 2005
Available online 30 September 2005
Communicated by Dhruba K. Chattoraj
Abstract
Plasmid R1 is a low-copy-number plasmid belonging to the IncFII group. The genetics, biochemistry, molecular biology,
and physiology of R1 replication and its control are summarised and discussed in the present communication. Replication of
R1 starts at a unique origin, oriR1, and proceeds unidirectionally according to the Theta mode. Plasmid R1 replicates during
the entire cell cycle and the R1 copies in the cell are members of a pool from which a plasmid copy at random is selected for
replication. However, there is an eclipse period during which a newly replicated copy does not belong to this pool.
Replication of R1 is controlled by an antisense RNA, CopA, that is unstable and formed consti-tutively; hence, its
concentration is a measure of the concentration of the plasmid. CopA-RNA interacts with its complementary target, CopTRNA, that is located upstream of the RepA message on the repA-mRNA. CopA-RNA post-transcriptionally inhibits
translation of the repA-mRNA. CopA- and CopT-RNA interact in a bimolecular reaction which results in an inverse
proportionality between the relative rate of replication (replications per plasmid copy and cell cycle) and the copy number;
the number of replications per cell and cell cycle, n, is independent of the actual copy number in the individual cells, the socalled +n mode of control. Single base-pair substitutions in the copA/copT region of the plasmid genome may result in
mutants that are compatible with the wild type. Loss of CopA activity results in (uncontrolled) so-called runaway replication,
which is lethal to the host but useful for the production of proteins from cloned genes. Plasmid R1 also has an ancillary
control system, CopB, that derepresses the synthesis of repA-mRNA in cells that happen to contain lower than normal
number of copies. Plasmid R1, as other plasmids, form clusters in the cell and plasmid replication is assumed to take place in
the centre of the cells; this requires tra c from the cluster to the replication factories and back to the clusters. The clusters
are plasmid-specific and presumably based on sequence homology.
2005 Elsevier Inc. All rights reserved.
Keywords: Antisense RNA; Basic replicon; CopA RNA; Eclipse period; Incompatibility; Plasmid copy number; Plasmid R1; Random
replication; Replication control
Fax: +46 18 53 03 96.

E-mail address: Kurt.Nordstrom@icm.uu.se.
0147-619X/$ - see front matter 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.plasmid.2005.07.002
K. Nordstrom / Plasmid 55 (2006) 126
1. Introduction
Plasmid R1 was isolated in 1965 from a pa-tient
suering from a Salmonella infection ( Datta and
Kontomichalou, 1965); the plasmid was named 1818
and renamed R1 by Meynell and Datta (1966).
Plasmid R1 and the closely related plasmids NR1
(also called R100; mainly studied by R.H. Rownd
and his co-workers), and R6-5 (mainly studied by
S.N. Cohen and his co-work-ers and K.N. Timmis
and his co-workers) belong to the incompatibility
group IncFII. They are large conjugative plasmids
with a size of 90 100 kb.
Plasmid R1 was used to define the so-called ba-sic
replicon (the smallest piece of the plasmid that is able
to replicate with the same copy number as the parent
plasmid) (Kollek et al., 1978). Regions (cassettes) of
plasmids R1 and NR1 were found to stabilise the
inheritance of the plasmids (Miki et al., 1980;
Nordstrom et al., 1980a). This led to the discovery of
two stabilisation cassettes (Gerdes et al., 1985), one
(parA) securing that each daugh-ter cell receives at
least one copy of the plasmid during partition of the
plasmids to the daughter cells at cell division (Gerdes
and Molin, 1985), and one responsible for killing
from within of the daughters that do not receive the
plasmid at cell division (parB or hok/sok) (Gerdes et
al., 1986). Plasmid R1 has also a second killer
system, parD (Bravo et al., 1987). Plasmid R1 further
encodes a tra system that mediates conjugal transfer
of the plasmid (Achtmann et al., 1978). All these cassettes are located in the resistance transfer factor
(RTF) part of the plasmids, whereas the antibiot-icresistance genes are present in the r-determi-nant; the
RTF moieties of R1, NR1, and R6-5 are very similar
(Sharp et al., 1973). The borders are between the RTF
and the r-determinant are IS1 sequences (Clerget et
al., 1981). The r-determi-nant is spontaneously
deleted in a fairly high fre-quency by homologous
recombination between these insertion elements. A
map of plasmid R1 is shown in Fig. 1.
In the present publication, I will concentrate on
plasmid replication and its control. Most of the
studies of plasmid R1 have been performed with
Escherichia coli as host.
Fig. 1. Genetic map of plasmid R1, showing the basic replicon

(BR; see Fig. 2), the partition (parA), killer (hoksok and parD),
and conjugation (tra) cassettes, the genes encoding resistance to
ampicillin (bla), chloramphenicol (cat), kanamycin (aph), spectinomycin/streptomycin (aad), and sulphonamide (sul), and the
insertion sequence IS1 that separate the resistance transfer factor
(RTF) from the resistance determinant (r-det) and splits the r-det in
two parts. The total size of the plasmid is about 95 kb (Blohm and
Goebel, 1978).
2. The basic R1 replicon

Replication of plasmids uses the enzymes and
proteins that replicate the host chromosome. A ba-sic
replicon is small, typically about 2 kb, and con-tains
an origin (ori) of replication (some plasmids have
two or three replication origins), (in most cases) the
structural gene (rep) for an initiation protein (Rep),
and genetic information for control of replication
(cop) (Nordstrom, 1985). The basic replicon of
plasmid R1 (Nordstrom et al., 1984) is shown in Fig.
2.
The origin of replication of plasmid R1, oriR1, is
about 188 bp (Masai et al., 1983) and has a structure
that is typical of Theta origins (Bramhill and Kornberg, 1988; Del Solar et al., 1998). There are two
structural genes for proteins in the basic
replicon, copB and repA, and one gene
encoding an untrans-lated RNA, copA. The
RepA protein is a typical ini-tiator protein

with a molecular weight of about
1998; Masai and Arai, 1987; Masai et al., 1983) in a

fashion similar to that of the binding of
Fig. 2. The basic replicon of plasmid R1 consists of an origin of

replication, oriR1, the structural gene, repA, for the initiator
protein RepA that binds to oriR1; tap that encodes a leader peptide
the synthesis of which by translational coupling is required for the
synthesis of the RepA protein, the gene copA for the antisense
RNA, CopA, that negatively regulates the translation of the repAmRNA, the structural gene, copB, for an inhibitor of expression of
the prepA promoter, and the constitutive pcopA and pcopB
promoters. Under normal conditions, the prepA promoter is almost
totally switched o by the CopB protein (Womble et al., 1985).
33,000 (Del Solar et al., 1998; Rosen et al., 1980;

Yoshikawa, 1974). Two promoters, pcopB and prepA, transcribe the repA gene. The CopB protein is a
repressor of transcription from the prepA promoter.
Just upstream of the repA gene there is a region
encoding a small leader peptide, tap (Blomberg et al.,
1992). Downstream of oriR1, there are termi-nators
of replication (see Fig. 3; Hill et al., 1988). The copA
gene is transcribed in the opposite direc-tion
compared to the two other transcripts. The CopARNA is an antisense RNA with its target, CopT, in
the complementary region of the repA-mRNA,
located upstream of the RepA information, and is the
main control element of plasmid R1 rep-lication (see
below).
3. Initiation of replication of plasmid R1
Plasmid R1 replicates according to the Theta
mode and replication starts at oriR1 (Del Solar et al.,
1998). Replication requires the binding of a fairly
large number (may be 20) of the plasmid-en-coded
initiator protein RepA to oriR1 (Del Solar et al.,
Fig. 3. The start region for leading strand synthesis of plasmid R1

(Bernander et al., 1992). The following features are shown: oriR1,
the R1 minimal origin (filled box) (Masai et al., 1983); DnaA and
RepA footprints (open and shaded ellipses, respec-tively) within
the oriR1 (Masai and Arai, 1987); the Ter sites Lter and Rter (open
rectangles with arrows indicating the polarity of replication fork
blockage) (Hill et al., 1988); the ssi region (striped box) (Bahk et
al., 1988; Masai and Arai, 1989; Nomura et al., 1991), the start site
for synthesis of the leading strand (vertical arrow). The nucleotide
sequence of the start sites found in vitro (Masai and Arai, 1989)
and in vivo (Bernander et al., 1992) are also shown. All features
are drawn to scale (shown as a horizontal bar).
the DnaA protein to the chromosomal origin, oriC

(Bramhill and Kornberg, 1988), i.e., the DNA is
wound two turns around a sphere of RepA molecules. This causes relaxation of the AT-rich region of
the origin; the substrate of initiation of replica-tion is
negatively supercoiled origin DNA (Diaz and
Staudenbauer, 1982; Funnell et al., 1986; Mar-ians et
al., 1986). The partially single-stranded re-gion then
recruits the proteins that form the replicon. The RepA
protein is cis-acting and a se-quence denoted CIS
was identified between the end of the repA gene and
oriR1 (Masai and Arai, 1988). Replication is
(essentially) unidirectional from oriR1 (Diaz and
Staudenbauer, 1982).
The start site of the leading strand (the 5 0-end)
during initiation of replication of plasmid R1 was
determined in vitro (Masai and Arai, 1989; Miyazaki
et al., 1988) as well as in vivo (Bernander et al.,
1992) and found to be located about 380 bp
downstream of the minimal origin of replication; this
position is located within a ssiA (single-strand
initiation) site (Bahk et al., 1988; Nomura et al.,
1991) (Fig. 3). Replication of plasmid R1 does not
require the DnaA protein. However, there is one
dnaA box in the origin region and the
eciency of R1 replication is increased in the presence of DnaA (Bernander et al., 1992). Similarly, the
30-end of the lagging strand was found to be located
within oriR1 (Krabbe, 1995).
Maas and Wang (1997) and Maas et al. (1997)
proposed an alternative model for initiation of replication of R1-like plasmids. A main feature of the
model is that transcription is an integral part of the
initiation process; in this model, initiation of
replication resembles that of the ColE1 family of
plasmids in which an RNA is used as a preprimer and
converted to a primer by the action of Rnase H (Itoh
and Tomizawa, 1980); however, plasmids of the
ColE1 family do not require any plasmid-encod-ed
initiation protein. The RepA protein is assumed to
function in the transition from RNA to DNA
synthesis. The model also explains the cis-action of
the RepA protein. However, I regard the scheme of
Maas and co-workers to be less likely, since plas-mid
R1 can replicate in vitro in the absence of pro-tein
synthesis and RNA synthesis by RNA polymerase
(replication is rifamicin resistant) if the RepA protein
is added (Diaz and Ortega, 1984).
Downstream of oriR1, there is a complete termination region with two Ter sites in opposite directions (Fig. 3). The Ter sites are targets for the Tus
protein; the Tus/Ter complex acts as a road block for
the replisome (Hill et al., 1988). The function of these
termination sites was studied by Krabbe et al. (1997).
Inactivation of the tus gene led to a sig-nificant
amount of rolling-circle replication of the plasmid;
hence, the function of the TerR R site seems to be to
secure that replication of R1 stops after one round.
The 30-end of the lagging strand did not cor-respond
to the TerRL site (Krabbe, 1995).
Replication of plasmid R1 requires RNA and
protein synthesis as shown by Diaz et al. (1981);
presumably, new RepA molecules have to be synthesised for each initiation event; plasmid R1 replication in vitro is independent of transcription by
host RNA polymerase (Diaz and Ortega, 1984).
4. Replication of plasmid R1 is random in time and

not coupled to the cell cycle
Rownd (1969) introduced density-shift experiments (Meselson and Stahl, 1958) in plasmid
studies and showed that selection of plasmid NR1 for

replication is random, i.e., a newly formed daughter
copy has the same probability of being selected for
replication as the other plasmid copies in the cell. In
these studies, Pro-teus mirabilis, in which NR1 is a
high-copy-num-ber plasmid, was used as host.
Bazaral and Helinski (1970) studied another highcopy-num-ber plasmid, ColE1, and found that also
this plasmid replicates randomly. At that time, the
idea was that some plasmids are relaxed (high copy
number) and other stringent (low copy number), and
that the latter replicate cell-cycle specifically as
experimental data obtained with the Baby Machine
(the membrane-dilution tech-nique; Cooper and
Helmstetter, 1968; Helmstet-ter and Cooper, 1968)
indicated (Cooper, 1972; David and Helmstetter,
1973; Zeuthen and Pato, 1971). The division between
relaxed and strin-gent plasmids turned out to be false,
since even the low-copy-number plasmids R1 and F
lac were found to replicate randomly (Gustafsson and
Nordstrom, 1975; Gustafsson et al., 1978). Later
experiments with the Baby Machine have confirmed
that all plasmids, with the possible exception of P1,
replicate randomly and indepen-dent of the cell cycle
(see discussion by Bogan et al., 2001). Hence, there
is no fundamental dif-ference between plasmids that
justifies a division into stringent and relaxed ones,
and the terms have disappeared from the plasmid
literature. Cell-cycle-independent replication of
plasmid R1 was also demonstrated by analysis of
cells of dif-ferent size (age) in the population; R1
replication was spread evenly over the entire cell
cycle (Gustafsson et al., 1978).
5. Eclipse period
By using the classical MeselsonStahl densityshift experiments (Meselson and Stahl, 1958),
Rownd (1969) found that selection of copies of
plasmid NR1 for replication was random. In such an
experiment, the bacteria are grown exponential-ly in
13
15
+
heavy medium (e.g., C6-glucose and NH4 ) and
12
14
+
shifted to light medium ( C6-glucose and NH4 ).
Before the density shift, all DNA
has two heavy strands (HH), after replication HL

DNA appears. A second replication gives LL DNA.
The reactions involved are the following:
HH-DNA + L-deoxyribonucleotides = 2 HL-DNA
HL-DNA + L-deoxyribonucleotides = HL-DNA
+ LL-DNA
LL-DNA + L-deoxyribonucleotides = 2 LL-DNA
On the assumption that the rate constant, k, is the
same for these three reactions, the kinetics of the
process are as follows (t is the time from the density
shift):
Relative amount of HH-DNA x1 e 2kt ;
Relative amount of HL-DNA x2

2e kt
e 2kt ;
HLmax 0.5 Ds=2.
5
4
It should also be stressed that the eclipse is an

operational or experimental parameter and refers to
large populations of cells and plasmids. The causes
of the eclipse will be discussed below.
The main molecular species of R1 in the cells is a
negatively
supercoiled
circle
(Diaz
and
Staudenbauer, 1982). The correct degree of
superhelicity is required for initiation of replication
(Marians et al., 1986). Initiation of replication relaxes this structure; the topological cycle during replication was recently discussed by Schwartzman and
Stasiak (2004). It takes about 4 min from initiation of
replication back to the superhelical state (Gustafsson
et al., 1978). This is part of the reason for the eclipse
period mentioned above and dis-cussed below.
Relative amount of LL-DNA x3
6. Control of replication
1 2e kt e 2kt .
3
This set of equations shows that x 1 starts to decline and x2 starts to increase directly after the
density shift, whereas x3 only slowly and gradually
increases. After one doubling time, the relative
amount of HL-DNA reaches its maximum, 50%, and
HH- and LL-DNA are equally abundant, 25% each.
However, Gustafsson et al. (1978) found that the
experimental results deviated from the results
predicted by Eqs. (1)(3) and that this deviation
could be explained by the presence of an eclipse
period (Ds) during which a newly repli-cated plasmid
copy could not participate in a new round of
replication; it did not belong to the pool from which
plasmids were recruited for replication [as discussed
by Nordstrom (1983)]. The eclipse was found to be
0.22 host generation times (Gustafsson et al., 1978)
in glycerol-minimal medi-um at 37 LC. (The heavy
medium in their experi-ment contained D2O and
15
NH4Cl.)
The relative amount of HL-DNA reaches its
highest value (HLmax) exactly one generation after
the density shift. This peak value is a function of the
length of the eclipse, Ds, and is expressed as a
fraction of the generation time, s. It should be noted
that Ds is a number and not a physical time (Olsson
et al., 2002, 2003):
Plasmids are present at defined copy numbers that

are functions of the plasmid, the host, and the growth
conditions and can vary from a few to hundreds of
plasmid copies per cell. In many cases, the
biologically most interesting copy num-ber is the
number of copies per average cell or per newborn
cell (the n value) (Nordstrom et al., 1980a).
However, for technical reasons, the copy number is
often determined as the number of plas-mid copies
per genome equivalent of the host DNA; this is also
appropriate in studies of control of replication, since
the size of E. coli cells increas-es with increasing
growth rate (Bremer and Den-nis, 1996) and since
control studies in most cases involve large
populations of cells and plasmids. Because the
relative amount of DNA (DNA/mass) in E. coli is
almost the same at moderate and high growth rates
(Bremer and Dennis, 1996) the ratio plasmidDNA/chromosomal-DNA gives a reason-able value
for the concentration of a plasmid, which is the most
appropriate expression of copy number in kinetic
studies.
Plasmid R1 is a low-copy-number (or a unit-copynumber plasmid since its concentration is about the
same as that of oriC) that is present in 45 copies per
fast-growing cell (n value about 3) (Nordstrom et
al., 1980a), which means that the
number of R1 copies per oriC (origin of the chromosome) is about 1.

Plasmids control their replication (see the car-toon
in Fig. 4). During exponential growth, a plas-mid
population attains a constant value of the copy
number, i.e., the rate constant for the growth of the
plasmid population is the same as that of the host
population. Hence, the number of plasmid
replications is exactly one per plasmid copy during
one doubling in cell massit should be stressed that
this is true for the population, not for the indi-vidual
plasmid copy (Nordstrom et al., 1984). Deviations
from the steady-state copy number are corrected such
that the steady state is main-tained. Hence, at reduced
copy numbers (e.g., immediately after the
introduction of a plasmid copy into a plasmid-free
host), there are more than one replication per plasmid
copy per doubling in mass, and that at increased copy
numbers, there is less than one replication per
plasmid copy per doubling in mass. Consequently,
the control curve for a plasmid falls within the shaded
areas in Fig. 4. This figure also shows that a plasmid
mea-
Fig. 4. Kinetics of control of plasmid replication (Nordstrom et

al., 1984). The control curves lie within the shaded areas (see
discussion in the text). Control of R1 replication follows an inverse
proportionality between the frequency of replication (replications
per plasmid copy and cell generation) and the relative copy number
(Nordstrom and Wagner, 1994; Nord- strom et al., 1984).
Experimental values were obtained in metabolic shifts (points 2
and 3; Gustafsson and Nordstrom, 1980), in temperature shifts
with an R1 mutant with temper-ature-dependent copy number
(points 1 and 4; Gustafsson and Nordstrom, 1980), and in
temperature shifts with R1 cloned onto a vector that is
temperature-sensitive for replication (point 5; Nielsen and Molin,
1984).
sures its concentration and uses this measurement to

adjust its frequency of replication accordingly.
Negative control would be the simplest and most
plausible way to get a control curve that falls within
the shaded areas of Fig. 4 (see dis-cussion by
Nordstrom, 1985); the first to suggest negative
control of replication was Pritchard et al. (1969) with
their enormously influential inhibitordilution model.
The isolation of mu-tants that have a 2- to 8-fold
increased copy number showed that IncFII plasmids
carry genetic information for their copy number
(Bra-dy et al., 1983; Morris et al., 1974; Nordstrom,
1969, 1971; Nordstrom
et al., 1972; Rosen
et al., 1980). The finding
that copy number is
a quantitative genetic trait opened replication control
for genetic and functional studies.
The isolation of copy mutants indicated that R1
encodes a negative control element and it was later
shown that also other plasmids encode such elements. These are of three types, repressor proteins,
antisense RNA, and DNA iterons (Del Solar et al.,
1998; Nordstrom, 1985, 1990). Plasmid R1 en-codes
an antisense RNA, CopA-RNA (Stougaard et al.,
1981a; see below).
The control curve of plasmid R1 shows that the
frequency of R1 replication (expressed as replications per plasmid copy and doubling in mass) is
inversely proportional to the copy number. This was
shown by shifts in copy number during shifts
between dierent media (growth rates) ( Gustafs-son
and Nordstrom, 1980) and in shifts from arti-ficially
increased to normal copy numbers (Nielsen and
Molin, 1984; Olsson, 2003). An inverse proportionality between the frequency of replication and
the copy number gives a constant number of
replications per cell and cell cycle irrespective of the
number of plasmid copies at the birth of the cells
the so-called +n mode of replication (Nord-strom et
al., 1984). This mode of control gives a fairly rapid
adjustment of copy number deviations: half a
deviation is corrected after one generation, 75% after
two generations, etc. (If the copy num-ber in a
newborn cell before the shift is m, it will be m 3(m
n)/4 after two generations.)
Careful analyses of the control curve
with other plasmids are scarce, but the
control curve of plas-mid F seems to be
much steeper than that of R1,
almost a step function (Tsutsui and Matsubara,

1981).
7. Antisense-RNA controls replication of plasmid
R1
Mutants of plasmids R1, NR1, and R6-5 that have
an increased copy number were isolated; the fact that
an increased gene dosage results in an in-creased
level of antibiotic resistance (Uhlin and Nordstrom,
1977) was used in the isolation of such cop mutants
(Brady et al., 1983; Morris et al., 1974; Nordstrom,
1969, 1971; Nordstrom et al., 1972). The mutations
were found to be located in a very small region of the
basic replicon (Fig. 5). Some of the mutations did not
aect the amino acid sequence of any putative
protein (Brady et al., 1983). This led to the idea that
the copy number was determined by interaction
between nucleic acids (RNA/DNA or RNA/RNA)
(Brady et al., 1983) and Stougaard et al. (1981a)
could show that a short (about 90 nucleotides) RNA
(CopA-RNA) was the controlling element.
Fig. 5. Secondary structure of CopA-RNA and the locations of cop

mutations. (Left) The secondary structure of CopA (Wagner and
Nordstrom, 1986). (Right) The location of single-base-pair
substitutions in loop II of CopA-RNA that increase the copy
number. The figure within brackets after each changed base refers
to the increase in copy number normalised to the copy number of
the wild type (=1). The dierent mutants (17) were described by
(1) Rosen et al. (1980) and Danbara et al. (1981); (2) Brady et al.
(1983); (3) Stougaard et al. (1981b);
(4) Brady et al. (1983); (5) Danbara et al. (1981); (6) Rosen et al.
(1980); (7) Givskov and Molin (1984).
The sequence of the CopA-RNA (identical in R1,

NR1, and R6-5) predicted a high degree of secondary
structure with a long stem and a 6-nu-cleotide loop
(Fig. 5); this structure was later con-firmed
experimentally (Wagner and Nordstrom, 1986). The
copy-number mutations in the copA gene were found
not only to be narrowly clustered but also to aect
the sequence of the loop in the CopA-RNA (Fig. 5).
Light and Molin (1981) dem-onstrated that CopARNA inhibits the expression of the repA gene and
that inhibition is exerted post-transcriptionally by a
small RNA molecule (Light and Molin, 1983).
Hence, CopA-RNA interacts with the complementary
region of the mRNA for the RepA protein; the target
is called CopT (Danbara et al., 1981); the concept
CopT was introduced by Pritchard (1978).
The fact that cop mutations aect the loop
sequences of CopA and CopT strongly suggests that
the initial interaction is by base-pairing be-tween
these loops. This was studied in vitro and the rate
constant for the formation of a complex between
CopA and CopT was determined and found to be
6
1
1
about 3 10 mole s . The eects of the cop
mutations are that the apparent rate constant for the
formation of a CopA/CopT com-plex is reduced
(Persson et al., 1990a,b); there was an inverse
correlation such that a fourfold reduc-tion in the rate
constant led to a fourfold increase in the plasmid
copy number (Persson et al., 1988).
The locations of the cop mutations indicate that
the initial interaction is between the loops in CopARNA and CopT-RNA; this has been called the
kissing reaction. The initial complex is rapidly
converted into a more stable complex (deep kiss-ing).
The subsequent reaction is slow and hardly ever
results in a complete duplex (Tomizawa, 1984;
Wagner and Simons, 1994; Wagner et al., 2002).
However, the formation of the deep-kissing complex
is enough for inhibition of translation of the repAmRNA. The o rate (i.e., resolution of the kissing
complex into CopA-RNA and CopT-RNA) is very
5
1
slow [about 10 s (Persson et al., 1990a), later
3 1
corrected to about 10 s (Hjalt and Wagner, 1995)].
With a slow o-rate and a rapid forward-rate,
equilibria are not obtained and, hence, the reaction
follows Briggs-Haldane kinetics rather than
MichaelisMenten
kinetics (Nordstrom and Wagner, 1994) (see discussion by Persson et al., 1990a).
The cop mutations are of two kinds. Some aect
the sequence of the loops without aecting the loop
size, whereas others also aect the loop size. The
eects on the plasmid copy number were up to
eightfold. Hjalt and Wagner (1992a) performed an
experimental study in which the loop size was varied;
they found the most rapid on-rate for the formation of
the kissing complex was obtained with loops of 67
unpaired nucleotides.
The stem in CopA-RNA is not a perfect duplex,
but contains bulges. This is crucial, because a per-fect
stem of the size of CopA-RNA is a substrate of
RNaseIII; a CopA-RNA with a perfect stem was
rapidly degraded in vivo (Hjalt and Wagner, 1992b).
8. CopA-RNA inhibits translation of the repA

message
The antisense RNA, CopA-RNA, interacts with
the complementary region of the two RepA-messages
to inhibit translation of the repA-mRNA (Fig. 2).
However, the site of interaction is up-stream of the
repA gene and, therefore, the interac-tion has to be
indirect.
The ShineDalgarno of the repA information is
hidden in a stem structure (Fig. 6B) and expression of
the repA message is therefore very ineective. It
turned out that RepA synthesis is translationally

coupled to expression of a leader sequence, tap. This
is short (24 codons) (Fig. 6A) and the puta-tive Tap
protein is of no significance, only transla-tion of the
short sequence is required for RepA formation by
translational coupling (Blomberg et al., 1992). This
was verified by a careful muta-genesis study that
showed that opening of the stem by mismatches leads
to tap-independent RepA synthesis (Blomberg et al.,
1994).
The antisense RNA, CopA-RNA, interacts with
the complementary region of the two RepA-messages
to inhibit translation of the repA-mRNA (Fig. 2). The
site of interaction is upstream of the repA message
and the eect has therefore to be indirect. The
interaction between CopA and its target CopT (on the
repA-mRNA) was found to inhibit translation of tap
and thereby of repA (Blomberg et al., 1992), which
explains how the indirect inhibition of RepA
formation is exerted.
The duplex between CopA and CopT-RNA is
subjected to cleavage by RNaseIII which results in a
reduced RepA expression (Blomberg et al., 1990).
Although this is not the main copy-number control, it
seems to be the first example of RNa-seIII-mediated
eect on gene expression, which is interesting now
when RNA interference and the control activity of
small RNAs (micro RNA and siRNA) is an
extremely hot topic (Couzin, 2002) and RNaseIIItype enzymes are key players (Carmell and Hannon,
2004).
Fig. 6. (A) The repA-leader sequence. (B) The secondary structure of the region around the repA ShineDalgarno and initiation site
(Blomberg et al., 1994).
9. Plasmid incompatibility
Closely related plasmids are incompatible, i.e.,
they cannot be maintained stably in a growing
population of the host bacteria (Novick and
Hoppensteadt, 1978). Incompatibility can be caused
by replication, by partition, and by the killing
systems (Nordstrom and Austin, 1989). Fig. 7 shows
the result of a hypothetical experi-ment in which two
dierently labelled variants of a plasmid (e.g.,
having dierent antibiotic-re-sistance markers; in the
figure, the derivatives are marked as open circles and
x, respectively) behave due to the randomness in
replication. If a cell starts with equal numbers of two
plasmid derivatives, random selection for replication
inev-itably causes unbalances in the ratio between
them; the number of classes with dierent ratios will
be 2n 1, and all classes will be equally abun-dant.
These unbalances cannot be corrected and cells with
only one copy of the one or the other derivative will
appear; such cells cannot give two
daughters containing both plasmid derivatives. If the

plasmids contain a par cassette it is most likely that
partition involves pairing of plasmid copies and that
the plasmids are equipartitioned (half the number of
copies go into each cell half) (Nordstrom and
Austin, 1989). Data suggest that plasmid pairs are
formed randomly, which ex-plains how the partition
process causes incompat-ibility (Austin and
Nordstrom, 1990; Nordstrom and Austin, 1989);
plasmid pairing was reviewed by Funnell (2005).
Ebersbach et al. (2005) recent-ly reported
experiments that they claim do not support plasmid
pairing as a cause of partition incompatibility; this
will be discussed in Section 16.
The frequency of formation of pure lines in each
cell generation (L) is shown by the following
equation (Nordstrom et al., 1980b):
L 2n=2n 1n 1&.
The equation is approximate, because all cells at

cell division are assumed to contain 2n plas-mid
copies; no copy-number variation is considered in
Eq. (5). Experimental evidence supports Eq. (5); a
cop mutant with a 50% in-creased copy number (n
value 4.5 compared to 3 for the wild type) was lost
with about a ten-fold decreased L value (Nordstrom
et al., 1980a); according to Eq. (5), the decrease
should have been eightfold.
To summarise, replication causes incompatibili-ty
between two plasmids because (i) the plasmids have
the same copy-number-control system and, therefore,
are accommodated within the same copy number (n
value), and (ii) selection for replication is random.
Fig. 7. Incompatibility due to random selection for replication and

partition. Two dierently-labelled (e.g., one with a bla and one
with a aph gene) derivatives of the same plasmid are marked o and
x, respectively. The n value of the plasmid is assumed to be 3 and
there are exactly 3 replications in every cell during one cell
generation. The starting population consists of equal proportions
of oox and oxx. Due to random replication, skewed distributions
between the two derivatives occur and five dierent equally
abundant classes of distributions are formed. If the plasmid
+
derivatives are Par , 30% of the progeny in each cell cycle will be
pure lines, while 35% will be oox and oxx, respectively
(Nordstrom et al., 1980b).
Mutations in the copA/copT region of plasmid R1

may change the incompatibility properties such that a
cop mutant may exclusively cause loss of the wild
type (Brady et al., 1983) or the cop mutation may
even make the mutant and wild type compat-ible
(Easton and Rownd, 1982; Miki et al., 1980) (Table
1). Table 1 shows the interactions between the loops
of CopA-RNA and CopT-RNA from wild-type and
mutant plasmids. In some combina-tions (mutants 4
and 5) there are mismatches (C/A pairs) that result in
the exclusive loss of the wild type in an
incompatibility experiment, whereas
10
Table 1
Incompatibility eects of mutations aecting the loop sequence of CopA-RNA (see Fig. 5)
Mutation (see Fig. 5)
Loop sequence
Wild type
4
CGCC
CGUC
CGCU
CGCA
Wild-type CopA/mutant CopT
Mutant CopA/wild-type CopT
Incompatibility
CGCC
GCAG
CGCC
GCGA
CGCC
GCGU
CGUC
GCGG
CGCU
GCGG
CGCA
GCGG
Wild type lost

Wild type lost
Compatible
Miss-matched base pairs are shown in bold letters.
there are unpaired bases in both combinations with

mutant 6, resulting in compatibility (Easton and
Rownd, 1982; Miki et al., 1980). Data as in Table 1
prompted Brady et al. (1983) to suggest that both the
eector and its target were encoded by the same
small region of the R1 genome.
It should also be stressed that CopA-RNA and
CopT-RNA are totally complementary also in the cop
mutants; hence, incompatibility is exerted be-tween
two dierently marked derivatives of a cop mutant,
but the rate of formation of pure clones decreases
according to Eq. (5) (Nordstrom et al., 1980b). The
L value approaches 1/n at high copy numbers.
The cause for the incompatibility results shown in
Table 1 is that the RepA protein is cis-acting (Masai
and Arai, 1988), i.e., binds essentially to the oriR1
that is present on the same plasmid copy as that from
which the corresponding repA-mRNA was formed.
Hence, the dierent plasmid copies do not share a
joint RepA-pool. Similarly, the CopA-RNA from the
mutant does not interact with the wild-type CopT and
vice versa. Other plasmids have a Rep protein that
acts both in cis and trans. In these cases, compatible
cop mutants cannot exist since mutant and wild type
share a pool of Rep protein (see Fig. 8).
In addition to partition systems, plasmids often
carry killer systems that kill from within daughter
cells that do not receive any plasmid copy (Gerdes et
al., 1997). These killer systems express incompatibility as shown by the following example: A basic
replicon of a plasmid P will be lost is an fre-quency
2n
of (1/2) (see Eq. (7) below). If the plas-mid carries
a killer cassette (kil), the daughters that do not
receive the plasmid will be killed
Fig. 8. Compatible R1 cop mutants due to cis-acting Rep proteins.

Replicon 2 diers from Replicon 1 by a single base-pair
substitution in the copA gene that makes the mutant CopA-RNA
incapable of interaction with the wild-type CopT-RNA and vice
versa (see mutant 6 in Table 1). The upper part shows the
properties of plasmid R1 with cis-acting RepA protein and the
lower the properties of a hypothetical plasmid with a Rep protein
acting both in cis and in trans; plasmid pT181 has such a Rep
protein. The two CopA-RNAs only interact with the mRNA from
the same replicon, whereas they produce identical RepA
molecules. Compatible plasmid derivatives are only possible in the
cis case, since the RepA molecules formed do not belong to a

common pool available to both replicons.
11
Table 2
Plasmid incompatibility caused by a killer system
Plasmid copies in
newborn cells
Frequency
Fate of
daughter cells
Fate of daughter cells

in presence of Vkil
0
1
2
3
4
1/16
4/16
6/16
4/16
1/16
Killed
Survives
Survives
Survives
Survives
Survives
Survives
Survives
Survives
Survives
The basic replicon of plasmid P carrying a killer cassette (kil) is assumed to be present in exactly four copies in every mother cell; each
plasmid copy has 50% probability of entering each of the daughter cells at cell division. The vector V is also carrying kil.
(Table 2). However, if the cells harbours another

plasmid also carrying the kil cassette, all daughter
cells will survive and cells without plasmid 0 will
appear; hence, kil is an incompatibility function but,
in contrast to replication and partition incompatibility, the eect is indirectit does not act on the
behaviour of the plasmids but directly on the survival
of the host. Furthermore, the incompati-bility eect
will be drastically reduced if the plas-mid carries a
par cassette, since the frequency of formation of
plasmid-free daughters is much re-duced in this case.
10. Parameters that determine the copy number of

plasmid R1
The copy number, Cop, of plasmid R1 is a function of a number of rate constants as shown in Eq.
(6) and Fig. 9; for deduction of the formula, see
Nordstrom and Wagner (1994).
Cop k kR kDA kT=kA kapp kH;
k is a proportionality constant; kR is the rate constant for initiation of replication by the RepA protein; kA and kT are the rate constants for the synthesis
of CopA-RNA and repA-mRNA, respec-tively; k DA
is the rate constant for the decay of CopA-RNA; k app
is the rate constant for the inter-action between
CopA- and CopT-RNA; and kH is the rate constant
for the growth of the host population.
The values of k and kR are not known, but there
are experimental data for the remaining constants. A
reduction of the rate of decay of CopA-RNA has
been shown to cause a proportional decrease
Fig. 9. Schematic diagram showing the reactions that deter-mine

the copy number of plasmid R1 (Nordstrom and Wagner, 1994).
The following constants are shown in the figure: kH, the rate
constant for the growth of the host population; k A and kT, the rate
constants for the synthesis of CopA-RNA and repA-mRNA,
respectively; kDA and kDT, the rate constants for the decay of
CopA-RNA and CopT-RNA, respectively; k app, the rate constant
for the interaction between CopA- and CopT-RNA; kR, the rate
constant for initiation of replication by the RepA protein; and k, a
proportionality constant.
in the R1 copy number (Soderbom and Wagner,

1998; Soderbom et al., 1997). A decrease in the rate
constant kapp causes a proportional increase in the R1
copy number (Persson et al., 1988). The ratio
between R1 DNA and chromosomal DNA increases
with decreasing growth rate, kH, of the host (Engberg
and Nordstrom, 1975).
The situation for kA and kT is more complicat-ed,
since transcription of the repA gene negatively
aects the rate of synthesis of CopA-RNA by socalled convergent transcription (Blomberg et al.,
12
1990; Stougaard et al., 1982). Therefore, an increased kT does not give a proportional increase in
the R1 copy number, but a much more drastic eect
(Fig. 10; Nordstrom and Uhlin, 1992). This is the
basis for runaway replication (see next section).
11. Runaway replication

The promoter pcopA is much stronger than pcopB
(Givskov et al., 1987), but even a moderate increase
in the eciency of the pcopB promoter causes the
copy number to increase drastically (Fig. 10)
(Nordstrom and Uhlin, 1992) as was shown by the
so-called runaway-replication mu-tants that were
isolated by Uhlin and Nordstrom (1978). These
mutants contain two mutations, one making
transcription from pcopB temperature dependent (the
rate increases with temperature) and one that reduces
the eciency of pcopA about twofold (Givskov et
al., 1987). These plasmids have a controlled but fairly
low (but higher than the wild-type) copy number at
30 LC, an increased copy number at temperatures
above 34 LC, and have lost replication control at 42
LC (Nordstrom and Uhlin, 1992; Uhlin and
Nordstrom, 1978).
Later Larsen et al. (1984) constructed another
class of runaway-replication mutants (Fig. 11). In
these, the repressor gene cI857 and the pL promoter
from phage k were inserted upstream of pcopB. The
repressor protein totally represses p L at tem-peratures
634 LC; the k promoter is gradually derepressed at
increasing temperature due to the temperature
sensitivity of the cI 857 repressor. At temperatures
>39LC, replication is totally uncon-trolled. Hence,
the copy number increases with increasing
temperature (Larsen et al., 1984).
In a shift from 30 to 42 LC with each type of
runaway-replication mutant, the plasmid popula-tion
immediately starts to grow much faster than the host
population (Larsen et al., 1984; Olsson et al., 2003;
Uhlin and Nordstrom, 1978). The rate of growth of
the host population after a while starts to decline and
eventually comes to a halt. At that time, the amount
of plasmid DNA exceeds that of chromosomal DNA
(Uhlin and Nord-strom, 1978).
Fig. 10. Eect of convergent transcription on the copy number of

plasmid R1. (A) The transcription pattern of the wild-type R1 with
its normalised copy number of 1, of a mutant in which pcopB and
the main part of the copB gene had been deleted (Nordstrom and
Nordstrom, 1985), and of a mutant in which the copB gene had
been inactivated by a frame-shift mutation (Nordstrom,
unpublished). (B) The copy number of the strains as a function of
the relative rate of repA transcription. This figure also includes
data from a mutant with a mutation that increases the e ciency of
the pcopB promoter (Givskov et al., 1987).
The kpL promoter is much stronger than the pcopB

promoter. Therefore, the rate constant for growth
of the plasmid population of this runaway
mutant is much higher than that of the

host popu-
13
12. The eclipse is determined by structural factors

and by the copy-number-control system
The runaway-replication plasmids oered an
opportunity to analyse the mechanism(s) behind the
eclipse period because these plasmids have a normal
copy-number-control system at 30 LC and lack
replication control at 42 LC. MeselsonStahl densityshift experiments were performed with a runawayreplication mutant and the corresponding wild-type at
42 LC. The eclipse was about 10 min for the wild
type and 34 min for the mutant, indicating that both
structural events and the copy-number-control
system contribute to the length of the eclipse period
(Olsson et al., 2003). The generation time of the
mutant-plasmid population was about the same as the
eclipse, indicating that the rate limitation was the
time it takes to replicate the plasmid and then to
regain the correct superhelicity of the plasmid; this
time has been determined to be about 34 min
(Gustafsson et al., 1978). I will come back to this below when microanalysis of the copy-number-con-trol
system will be discussed (Section 15).
Fig. 11. Transcription pattern of a runaway-replication mutant of

R1. (A) Map of the plasmid derivative: The k promoter p Rk was
introduced upstream of the R1 basic replicon and the plasmid also
contained the repressor gene cI857, encoding a temperaturesensitive k repressor. (B) Transcription pattern at low temperature,
at which the repressor is fully active, and at high temperature,
when the repressor is denaturated.
lation. After shift from 30 to 42 LC, the plasmid

population grows with a generation time of about 4
min (Olsson et al., 2003). This high rate of growth
continues for several plasmid generations, but after
that the growth of the plasmid popula-tion becomes
linear with a rate of about 400 R1 replications per
cell and cell generation (Olsson et al., 2003). The
slowing down of the amplifica-tion process is
presumably due to shortage of DNA polymerase III.
The runaway-replication mutants are useful
cloning vehicles (Uhlin et al., 1979) because the host
population continues to grow during runaway
replication. Hence, not only the plasmid but also
gene products of cloned genes can be amplified
(Nordstrom and Uhlin, 1992; Uhlin et al., 1979).
13. Copy-number distributions

The experimental analyses of the R1 system have
used large populations of R1-carrying host bacteria
because there is so far no way of perform-ing studies
of individual bacteria. Therefore, the data obtained
reflect average values of copy num-ber, etc. The data
indicate that plasmid R1 repli-cates according to the
+n mode, i.e., the (average) number of replications
per cell and cell cycle is n which equals the (average)
number of R1 copies per newborn cell. This number
is about 3. However, there has to be variations in the
num-ber of plasmid copies per cell. Theoretical
distribu-tions have been determined assuming that
the number of plasmid replications per cell has a
Pois-sonian spread around n. This gives the predicted
distributions shown in Fig. 12 (Nordstrom and
Aagaard-Hansen, 1984; Olsson et al., 2004).
Lbner-Olesen (1999) made the so far best attempt to experimentally determine the copy-num-ber
distributions for several plasmids, including R1. He
cloned the gene (gfp) for the Green Fluores-
14
Fig. 12. Copy-number distributions of plasmid R1. The open

symbols show the calculated distributions obtained if all cells in
the population had 4 R1 copies at birth and 8 at division, and if
there were a Poissonian spread around 4 in the number of
replications per cell and cell cycle. The closed symbols show
experimental data calculated from (Lbner-Olesen, 1999).
cent Protein behind an inducible promoter on dierent plasmids, grew the bacterial populations exponentially, induced the promoter, and analysed the
fluorescence by flow cytometry. A limitation was that
fairly long periods of induction before harvest-ing the
bacteria were required to obtain signals that were
strong enough. Hence, the data do not show the
instantaneous copy number but copy numbers
integrated over a period of time. Furthermore, the
data do not directly determine the amount of plasmid, but of a product formed from a gene present on
the plasmid. Nevertheless, the data are interest-ing
and I have used them to calculate the copy-num-ber
distribution assuming that the n value is 4 (Fig. 12).
The data are in reasonable accord with the
theoretically deduced data, thereby supporting the +n
mode of replication. The copy-number dis-tributions
are fairly broad, which has physiological
consequences (see next section).
14. The CopB system
In addition to the CopA/CopT system, plasmid R1
has the CopB system. The copB gene is tran-scribed
from the pcopB promoter as part of the long repAmRNA (Fig. 2). The CopB protein is a repres-sor of
the prepA promoter from which the short repAtranscript is transcribed (Fig. 2). The second
promoter is about twice as strong as the former (Light
et al., 1985), but due to the presence of the
CopB protein very little transcription takes place

from the prepA promoter under normal conditions
(Light et al., 1985; Womble et al., 1985). The pcopB
promoter is constitutive and the active repressor is a
tetramer (Riise and Molin, 1986). Hence, if the plasmid copy number decreases, the CopB concentra-tion
will decrease due to growth of the cells; the eect on
repressor activity will be drastic because of the
tetrameric nature of the repressor.
Riise et al. (1982) constructed a derivative of the
basic R1 replicon in which the copB gene with its
promoter were deleted and inserted somewhere else
on the plasmid. Hence, replication was totally
governed by the prepA promoter and the CopB
concentration. Such a plasmid had a copy number of
about 1/4 of that of R1. Hence, a reduction of the R1
copy number due to the variations in copy-number
distribution would cause a drastic derepression of the
prepA promoter in the cells that have low copy
numbers.
The function of the CopB system would there-fore
be to speed up replication in cells with low copy
numbers to decrease plasmid loss. This was tested by
the introduction of a vector with the copB gene under
its own promoter into a cell car-rying plasmid R1. In
such a situation, derepression of the prepA promoter
would not occur resulting in slight decrease in the
average R1 copy number and a more pronounced
increase in the frequency of cells with low copy
numbers. Plasmid R1 would be lost in all cell classes
m
with a frequency of (1/2) , where m is the plasmid
copy number in each class, if the plasmid is Par , but
would ideally only be lost from the cells with one R1
copy at the time of cell division if the plasmid carried
+
a functional par cassette (Par ). The results were as
predicted (Table 3); the eect on the loss rate was
about 70% in the Par case (Nordstrom and AagaardHansen, 1984), whereas there was a sevenfold
+
increase in the rate of plasmid loss of the Par
plasmid in the presence of high concentrations of the
CopB protein (Olsson et al., 2004).
15. Computer simulations of replication control
Macroscopic analyses of replication of

IncFII plasmids (the analyses of large
populations of plas-
15
Table 3
Eect on the CopB system on the stability of inheritance of plasmid R1
Plasmid
Par
+
Par
Loss frequency per cell generation

Extra CopB
2
1.510
4
1 10
References
+Extra CopB
2.5 10
4
7 10
Nordstrom and Aagaard-Hansen (1984)

Olsson et al. (2004)
Par , the R1 basic R1 replicon; Par , the R1 basic replicon with the par cassette.
The bacterial populations were grown exponentially in LB medium. Samples were taken at intervals and the relative amount of R1b
containing cells was determined.
A compatible high-copy-number vector (pBR322) carrying the copB gene under its own promoter was transduced into cells that
c
contained the R1 derivative.
a
mids and cells or of the average cell) have been

performed by e.g., Nordstrom et al. (1984), Womble and Rownd (1986) and Nordstrom and Wag-ner
(1994). In these, all plasmids form one large
population. The macroscopic analysis by Nordstrom and Wagner (1994) was discussed in Section
10 and Fig. 9; it was an algebraic deduction of how a
number of rate constants determine the copy number
of R1 and did not involve any computa-tional
analysis. Womble and Rownd (1986) deter-mined the
rates of transcription of the various promoters and
used these data to model NR1 rep-lication (Womble
and Rownd, 1986). They simu-lated what would
happen to a large number of parameters (i) during the
establishment of a plas-mid population after the
introduction of one plas-mid copy into a cell, (ii)
during one cell cycle in the steady state, (iii) during
the first generations after an unequal distribution of
the plasmid copies to the daughter cells. They also
compared the wild-type with cop mutants of plasmid
NR1. They con-cluded that the simulations predict a
very stable pattern of inheritance of NR1 despite its
low copy number, which is in agreement with
experimental observation. These analyses fit well
with experi-mental data (see e.g., Fig. 9). However,
microanal-ysis, analysis of individual cells, gives
important additional information and takes into
account that there is variation between dierent cells
in the population.
Microscopic analysis of the R1 replication-control system was first performed by Rosenfeld and
Grover (1993). The knowledge of the R1 system
available at that time was used in their analysis. The
following assumptions were made:
(i) The promoters pcopA and pcopB are transcribed constitutively.

(ii) The CopB repressor is a tetramer.
(iii) The pcopA promoter is subjected to convergent transcription.
(iv) The host cells grow exponentially and divide
equally at a fixed age.
(v) Plasmid copies are selected randomly from a
plasmid pool; the newly formed plasmid copies do not belong to the replication pool dur-ing
an eclipse period.
(vi) The plasmid copies are distributed randomly to
the daughter cells at cell division since the
plasmids do not contain a par cassette.
The computations of Rosenfeld and Grover (1993)
resulted in stable steady states with respect to
average copy number and to copy-number distributions irrespective of the initial conditions. The loss
rates (formation of plasmid-free cells) were also
determined and were found to fit experimental data.
Addition of copB mRNA in trans gave a min-or (6%)
reduction in the average copy number of the plasmid,
whereas no eect was found in real experiments
(Riise et al., 1982); a reduction by 6% would not
have been possible to detect experimen-tally. As
shown in Section 14, the experimentally determined
loss rate increased by 70% upon the addition of extra
copA mRNA (Table 3) compared to 82% in the
computer simulations.
The analyses by Rosenfeld and Grover (1993)
were consistent with the +n mode of replication, i.e.,
the average number of replications per cell and cell
cycle is independent of the copy number at birth of
the cells (Gustafsson and Nordstrom, 1980).
16
Light and Molin (1982) suggested that the CopB

system is a rescue system preventing cells with low
numbers of R1 copies from forming plas-mid-free
daughters. Rosenfeld and Grover (1993) computed
what would happen to a newborn cell that received
one R1 copy by infection. They found an overshoot
in replication of the wild-type plasmid, but also in the
absence of the copB system there was a rapid
increase in the copy number of the progeny of the
infected cells. Hence, the eect of the CopB on
establishment does not seem to be crucial. The only
experimental data about what happens during transfer
of a wild-type R1 to reci-pient cells were obtained by
Gustafsson and Nord-strom (Nordstrom, 2002);
they found that, in accordance with the computer
simulations, there was a substantial overshoot in
plasmid replication such that the average copy
number one generation after the transfer was about
normal. No transfer experiment have been performed
in which the CopB system in the recipient cells is
inactivated.
Ehrenberg and Sverredal (1995) developed a
model for control of replication of plasmid R1; they
performed both macroscopic and microscopic
simulations. In contrast to previous attempts they
took into consideration that the RepA protein is cisacting. Ehrenberg and Sverredal also wanted to
account for the eclipse period during which a newly
formed plasmid molecule was not available for a new
round of replication (Gustafsson et al., 1978; Koppes
and Nordstrom, 1986; Nordstrom, 1983). They
introduced the idea that the copy-number control is a
multistep process (Fig. 13). They assumed that the
active repA-mRNA is at-tached to the plasmid and
never leaves the plasmid from which it is transcribed.
They also assume that there is never more than one
repA-mRNA on an R1 copy. Furthermore, the CopA
system inacti-vates the mRNA, which then leaves the
plasmid and, to get more RepA, a new mRNA has to
be synthesised. The authors modelled the behaviour
of the system and compared their results to published data. They were able to faithfully calculate the
increase in the population during shifts be-tween
dierent
steady
states
as
demonstrated
experimentally by Gustafsson and Nordstrom
(1980); the simulations were in accord with the +n
mode of replication. They used the data of
Fig. 13. A model for multistep control of initiation of replication of

plasmid R1 (from Ehrenberg and Sverredal, 1995). It is assumed
that n cis-acting RepA molecules are required for initiation of
replication. Plasmids Xi and Zi have i RepA molecules bound to
oriR1 and Zi in addition has an active repA-mRNA bound to the
plasmid. Rate constants: kIT, the rate of transcription of the repA
gene; kT, the rate of translation of the repA message; q, the rate of
inactivation of mRNA by CopA RNA.
Koppes and Nordstrom (1986) to determine the

length of the eclipse period and the kinetics of rereplication of a replicated R1 copy and find that the
best fit is that the RepA molecules required for the
firing on an R1 origin are formed in about 3 bursts;
hence, the control is a multistep process.
Ehrenberg and Sverredal (1995) further addressed runaway replication of R1. In this, there is no
CopA synthesis at nonpermissive temperature, but
still the rate of plasmid replication is only about
threefold that of the host. They used their model to
explain the relatively slow replication of the plasmid
population as a consequence of the replication-control system. Their analysis was based upon the mutant that was isolated by Uhlin and Nordstrom
(1978). They were surprised that the growth rate of
the plasmid population was slow, but concluded that
that might be a consequence of the properties of the
multistep character of initiation of replication.
However, the growth rate of the runaway-replica-tion
plasmid population was found to be very short if a
strong promoter was introduced upstream of the
basic replicon (Fig. 11). As discussed in Section 12,
the copy-number-control system is responsible for
part of the eclipse period, again supporting a multistep mechanism.
Still it might be surprising that the generation time
of the runaway-replication mutants of Uhlin and
Nordstrom (1978) is so long. The average number
of RepA molecules formed per repA17
or plasmid NR1 (that is

n the pcopB region) by
); at a generation time of
essages were formed per
rate of transcrip-tion from
naway-replica-tion mutant.
uld be formed about every
of protein from a mRNA,
enough RepA for an
anslation initiation region
odon is GUG (Fig. 6B)
ence, more than one repAfor the formation of the
needed for initiation of
why the replication rate is
rate of repA transcription
Sverredal (1995) discuss

lasmid such as R1 can be
owing population of host
smid replication has been

y the Ehrenberg group
al, 1995; Paulsson and
aulsson et al., 1998). Two
d compared, ColE1 and R1
2000, 2001). They show
cation of ColE1 requires
rimer, RNA II. This is
nse RNA, RNA I. The
act during a short time
ring which about 250
he growing chain of RNA
RNAII is insensitive to
represents the about 250
NA I is inactivated by the
ss has to start from the
esis of a new RNAII.
the synthesis of about 20
antisense RNA, CopA,
epA-mRNA. If inhibition
hat the process has to go
e already formed RepA
.
he formation of plasmiddetermined by the aver-
age copy number and the spread around this value

(Nordstrom and Austin, 1989). The presence of a
plasmid in a cell involves a metabolic burden and the
ideal situation would be to have a low copy number
and a small variation around this copy number;
however, the latter would also impose a cost
(Paulsson and Ehrenberg, 2000). Plasmids ColE1 and
R1 use distinctly dierent types of mul-tistep
replication control (see previous paragraph) and the
consequences are dierent as analysed by Paulsson
and Ehrenberg (2000): The multistep control system
of ColE1 very eciently reduces the variation in
copy number, whereas that of R1 does not aect the
variation in copy number apart from reducing the
incidence of cells with only one R1 copy at the end
of the cell cycle. This would be most important for a
plasmid that has a functioning partition system, as R1
does, because this will eectively prohibit the
formation of plas-mid-free cells from dividing cells
with two or more plasmid copies. ColE1, on the other
hand, is Par .
A well-working multistep control predicts that the
copy number of a plasmid dimer should be half that
of the monomer (Paulsson et al., 1998). This has
been shown to be the case for plasmid ColE1
(Chiang and Bremer, 1988; Summers and Sherratt,
1984; Summers et al., 1993). In our laboratory, we
have performed the same analysis for plasmid R1 and
found that the copy number of a head-to-head dimer
is 50% of that of the monomer (E. Bjork-man,
masters thesis 2000: Stability of maintenance of
plasmid R1). This indicates that copy-number control
of R1 is multistep.
Paulsson and Ehrenberg (2000) also carefully and
in detail discuss the published models for rep-lication
control.
16. Clustering of R1 and replication factories

The basic replicon of plasmid R1 is lost in a frequency (L) of
L 1=22n;
where 2n is the copy number in the dividing cells.

This indicates that the unit of segregation is one
plasmid copy (Nordstrom et al., 1980a). However,
18
experiments aimed at finding where plasmids are

located in the cells gave a big surprise; dierent
plasmids were found to form clusters in the cells
(Eliasson et al., 1992; Ho et al., 2002; Pogliano et al.,
2001; Weitao et al., 2000a). Plasmid R1 is present in
45 copies per average cell (n value 3) but on the
average forms 2.3 clusters per cell (Weitao et al.,
2000a). A cop mutant of R1 (n value 11) also formed
2.3 clusters (Weitao et al., 2000a), but its loss rate
4
was 10
compared to 0.015 per cell and cell
generation for the wild-type plasmid (Nordstrom et
al., 1980a). This might seem to be a paradox between
the unit of segrega-tion and the clustering.
Nordstrom and Gerdes (2003) suggested a solution to the apparent paradox. The main feature of the
presented model was that segregation of the plasmid
occurs throughout the cell cycle (Fig. 14). Plasmid
replication is random and is as-sumed to take place in
replication factories in the middle of the cells. The
plasmid copy to be repli-cated is recruited from the
plasmid clusters that are located towards the cell
poles. The fate of the daughter copies is dependent
upon whether the plasmid has a functional par system
+
or not. For Par plasmids, the partition machinery
directs the two daughter copies in opposite directions
according to the active system described by Ml-lerJensen et al. (2002). Par plasmids do not have this
means of directioning but are randomly mov-ing in
either direction. Hence, both daughter cop-ies go
together with a probability of 1/2, thus leaving one
out of four daughter cells without plasmid. This
process is repeated for every repli-cation event
leading to the random segregation found
experimentally Eq. (7). The model fits the current
idea that chromosome replication and partition are
parallel processes that take place simultaneously
(Nordstrom and Dasgupta, 2001; Wu, 2004).
The chromosome replicates in replication factories in the cell centres (Gordon and Wright, 2000;
Koppes et al., 1999; Lemon and Grossman, 1999,
2000, 2001; Sawitzke and Austin, 2001). Onogi et al.
(2002) showed that also plasmid F replicates in the
cell centres. Hence, it is most likely that the same is
the case for other plasmids, but experimen-tal
evidence for plasmid R1 is still lacking. Such
Fig. 14. The plasmid replication cycle: a model that explains how
random segregation of Par plasmids and plasmid clustering do not
contradict each other (From Nordstrom and Gerdes, 2003).
Plasmids are recruited at random from the clusters to the
replication factories located in the middle of the cells. The
+
daughters of a Par plasmid are moved in opposite directions by
the partition system, whereas the daughters of a Par plasmid
randomly move in the cells. Hence, partition occurs during the
whole cell cycle since replication is random in time.
studies are under way in our laboratory. However, it

should also be noted that the Grossman group (Wang
et al., 2004) reported data for Bacillus subtilis that indicate that the replisomes
assemble where the origin of replication is

rather than the
o the factories. They also

so for the chromosome.
ection 11) can lead to very
per cell. We per-formed an
naway replica-tion was
the plasmid copy number
ditions to normal (Olsson,
smids presumably formed
et demonstrated, though).
asmid replication followed
shift experiments showed
did not have the same
ted; in the end, though, all
have been replicated. This
the +n mode implies that
process and that the CopA
onal to the total plasmid
s studies of where in the
the clusters or at midcell
kers studied the nature of

idman et al., 1996; Reich
erstand how large copy
usands of copies per cell)
he cells. They found that
d in vitro formed liquid
cture of the plasmids was
the for-mation of liquid
f the same process would
id concentrations, but the
group seem to oer a
mid clusters are formed.
y stable, and experiments
been without success
n our laboratory).
ters at defined preformed

hey determine their own
is that dierent plasmids
their clusters (Eliasson et
). This suggests that each
clusters and that the basis
ology; it should also be
clusters are dis-tinct from
19
Par plasmids have distinct locations at the cell

quarters or even closer to the cell poles
eitao et al., 2000b). The

ect of the parti-tion system
ers is at present unknown,
se-quence of the way the
e two daughter copies in
r-Jensen et al., 2002); the
eloaded from the partition
nce from the replication
ed by data reported by
o showed that the average
s increases with increasing
rowing the
populations
ch medium,
replication
3/4 positions, and the Par
lasmid copies a defined
n factories. This might be
(Mein-hardt and de Boer,
pole oscillation of Min
re-placed by oscillation
dle of the cells. Hence, the
not invoke any special
es for the localisation of
interesting to determine if
ible replicons having the
e.
1 have a more random

2000b). Presumably this is
copies diuse away from
nd may appear any-where
mid copies find each other,
urthermore, the formation
pendent of the par system,
+
s is the same for Par and
et al., 2000b).
ve suggest that the ba-sis

mid clusters is se-quence
the molecular nature of
the forma-tion of liquid
t al., 1996; Reich et al.,
rences between dierent
which cluster a plasmid
interesting to study the
dent, compatible replicons
gment.
20
The model described in Fig. 14 also explains how

new clusters might be formed. Many cells contain
only two clusters at the time of division (Weitao et
al., 2000b); hence, the daughters will contain only
one cluster located close to the cell poles. Replication
in the middle will then directly lead to the formation
of a new cluster in the plas-mid-free cell half.
Experiments to analyse the building up of new
clusters during growth of the cells are urgently
needed.
In the model described in Fig. 14 there is trac
from the plasmid clusters to the putative replication
factories and back again to the clusters. This trac
has so far not been experimentally demonstrated. The
model assumes that clusters are fairly stable, but that
one copy at a time leaves the cluster and moves to the
replication factories. What causes this release of a
plasmid copy from the cluster? It is likely that the
RepA protein binds to the plasmids already when
they are located in the cluster since the RepA protein
is cis-acting. This might change the topolo-gy of the
plasmid (Levin-Zaidman et al., 1996) such that it is
less firmly held in the cluster.
A putative problem with the model shown in Fig.
14 is the existence of partition incompatibility. Two
independent replicons having the same par cassette
are incompatible and the data support the notion that
homo- and heteroplasmid pairs of plasmids are
formed at random (Nordstrom and Austin, 1989;
Nordstrom et al., 1980b). Hence, partition of the
daughter plasmids cannot take place immediately
after replication, but some time has to be allowed for
the formation of heter-ologous plasmid pairs. Perhaps
this is not a prob-lem, because it has to take some
time for the binding of the Par protein to the par site
on the plasmid. The formation of heterologous
plasmid pairs has so far not been directly visualised in
vivo.
It should also be noted that Ebersbach et al. (2005)
presented data that they claim contradict plasmid
pairing as a cause of partition incompati-bility. They
found that, in cells with two dierent replicons (F
and R1) carrying the same par cas-sette, plasmids
never form clusters with both types of plasmids but
that each cluster always contains only one type of
plasmid. However, the location of the clusters was
changed in the cells. Hence, they proposed that
partition incompatibility is
not caused by pairing of heterologous plasmids but

instead by random positioning of pure plasmid
clusters along the long axis of the cell. It is unclear
how Ebersbach et al. (2005) envisage how the partition system works in a heteroplasmid cell and how
plasmid incompatibility is exerted. If the par-tition
system works by sending the two daughter copies in
opposite directions without interference from the
other plasmid, why would that give insta-bility?
There result is very important and provoca-tive, but, I
am not totally convinced that their interpretation
excludes the involvement of plasmid pairing: in the
scheme by Austin and Nordstrom (1990) and that of
Fig. 14, plasmid pairing is an intermittent process
occurring right after replica-tion has been completed.
Therefore, in the scheme of Fig. 14, it is still possible
that plasmids may pair before they are partitioned. It
would be interesting if the experiments by Ebersbach
et al. (2005) were repeated with a multicopy plasmid
+
because in this case, the low-copy-number Par
plasmid would behave as if it were Par ; how would
that show up in the distribution of the plasmid
clusters in the cell?
17. Other systems for control of plasmid replication

by antisense RNA
Antisense RNA as a substance that controls
molecular events was first observed in bacteria. In
the ColE1 system, antisense RNA (RNAI) con-trols
initiation of replication of the plasmid by inhibiting
the ability of the preprimer (RNAII) to change into a
primer of replication (Eguchi et al., 1991). In other
plasmids, the antisense RNA negatively controls the
ability of a mRNA for a rate-limiting initiator protein
to be translated (Wagner and Simons, 1994; Wagner
et al., 2002). However, the mechanisms are very
dierent: in R1, the structure of the translation start
region of the mRNA is aected, whereas in pT181,
the transcriptional terminator system is aected.
Many plasmids use the R1 system in principle, but
there are many variations of the same theme (Wagner
and Simons, 1994; Wagner et al., 2002).
Plasmids replicate in several dierent
ways, the Theta mode, the rolling-circle
mode, and the
R-loop mode (Del Solar et al., 1998). Plasmids that

use antisense RNA to control their copy num-ber
exist in all three classes. Hence, antisense con-trol
and replication mode have not co-evolved.
18. What is missing?
In spite of a lot of work in several laboratories
over a long period of time, fundamental aspects of
the R1 system are still missing, e.g.,
(i) Is there a time window for the CopA-RNA to
interact with CopT-RNA as there is in the ColE1
system in which the antisense RNA, RNAI, has to
interact with its target within 5 s, otherwise the
preprimer cannot become a primer irrespective of the
binding of RNAI (Tomizawa, 1986). How-ever, the
ColE1 system is dierent from the R1 sys-tem
because, to be eective, the reaction has to take
place before transcription has proceeded past the
transition point, whereas in the R1 system, it might
be possible for the antisense-RNA to inter-act as long
as the repA-mRNA still is intact. How-ever, we do
not really know.
(ii) There is a switch in the fate of a repA-mRNA
transcript: either it is inactivated by the antisense
RNA or translated into RepA protein. A so far
unanswered question is if initiation of translation
prohibits the antisense RNA to bind or, if CopT-RNA
during translation is still sensitive to CopA-RNA.
Perhaps the sterical changes in the structure of the
repA-mRNA will not be possible once trans-lation
has started, but the translation start se-quence is
fairly week, which may cause delays in subsequent
initiations of translation.
(iii) Can more than one RepA molecule be
formed from a repA-mRNA? This is linked to the
previous question and is important for the
understanding of the multistep process discussed
above. The average number of RepA molecules
formed per repA-mRNA was determined by Womble and Rownd (1986); at a generation time of 60
min, about 0.2 repA-mRNAs were formed per min
per plasmid copy. This gives about 12 per cell
generation per plasmid copy and, hence, on the
average about 2 RepA molecules per repA-mRNA.
(iv) The paradox between R1 clustering and that
the unit of segregation is one R1 copy raises
21
important questions about where in the cell R1

replicates and how new clusters are formed.
(v) In the idealised models, the cell is regarded as
one compartment. However, the plasmids are present
in clusters. Even at very high copy num-bers, the +n
mode of replication is upheld even when, in large
clusters, all copies do not have the same probability
of replication but eventually they all seam to be
replicated (Olsson et al., 2003). This suggests that
CopA is spread in the entire cell vol-ume, otherwise
its concentration would be dier-ent in clusters of
dierent size, which might explain why all plasmid
copies do not have the same probability of being
replicated. Therefore, replication in large clusters has
to be analysed.
(vi) The molecular mechanism of partition
incompatibility is still not explained. Does partition
work essentially as proposed in Fig. 14 and does
partition incompatibility involve pairing. If pairing of
heteroplasmid copies is not involved, why does
the presence of a compatible replicon with the same
+
par cassette destabilise a Par plasmid.
19. Concluding remarks
R1 was (fortuitously) a lucky choice of system,
because its antisense-RNA-mediated copy-number
control is fairly simple and straightforward. Plas-mid
R1 uses the concentration of the CopA-RNA as a
measure of its concentration, and the CopAconcentration is used to settle the level of initiation
of replication via a bimolecular reaction between
itself and its target, CopT-RNA (cf. Fig. 4). Due to
the relative instability of CopA-RNA, deviations in
copy number rapidly aect the replication frequency
this is dierent from protein-repressor-based
systems (as e.g., in kdv), because most proteins are
stable; the kinetics of the replication control system
of kdv was modelled mathematically by Lee and
Bailey (1984a,b). The iteron-based systems seem to
be much more com-plicated (Morrison and Chattoraj,
2004).
R1 was (fortuitously) a lucky choice of system,
because it turned out to be a system in which a regulatory RNA post-transcriptionally regulates
translation of a mRNA; another early identified
similar system was the system that controls the
22
frequency of transposition of Tn10 (Simons and

Kleckner, 1983). This at least to some extent inspired Izant and Weintraub (1984) in their at-tempts
to develop artificial control of gene expression by the
introduction of RNA antisense to the gene to be
silenced. Small regulatory RNAs have turned out to
be crucial control elements both in prokaryotes
(Vogel et al., 2004; Wagner et al., 2002) and
eukaryotic systems (Break-through of the Year
2002, Couzin, 2002).
R1 (and the other IncFII plasmids NR1 and R6-5)
is presumably the best known plasmid sys-tem, since
it has been studied molecularly, geneti-cally,
biochemically, and, in my mind, most importantly,
physiologically. A fair amount of computer modelling
has also been performed with the R1 system (see
Section 15 above). Early key discoveries were the
finding by Rownd (1969) that plasmid replication is
random and the discovery of cop mutants
(Nordstrom, 1969, 1971; Nordstrom et al., 1972).
These discoveries opened plasmid rep-lication for
genetic and physiological studies, and lead to the
realisation that plasmid replication is independent
and not coupled to the cell cycle.
Acknowledgments
I want to express my sincere gratitude to a large
number of people who have participated in our
studies of plasmid replication and its control:
graduate students (Helle Aagaard-Hansen, Rolf
Bernander, Pontus Blomberg, Asa Eliasson, Bir-gitta
Engberg, Petter Gustafsson, Margareta Kra-bbe,
Annika Lundback, Marie Nordstrom, Jan A.
Olsson, Christine Persson, Erik Riise, Peter Stougaard, Bernt Eric Uhlin, Tao Weitao, Marie Oh-man),
postdoctoral fellows (Ramon Diaz, Luud Koppes,
Janice Light, Sophie Maisnier-Patin, An-drew
Merryweather, Helen Withers), and senior co-workers
(Santanu Dasgupta, Sren Molin, E. Gerhart H.
Wagner). I have enjoyed co-operation with several
scientists: Otto G. Berg, Mans Ehren-berg, Fleming,
G. Hansen, Lewis C. Ingram, Jo-han Paulsson, John
W. Perram, Jan Zabielski. I also thank many
colleagues for interesting and fruitful discussions
over the years: Stuart Austin, Stanley N. Cohen,
Bruce R. Levin, Richard P.
Novick, Robert R. Rownd, Robert Simons, David

Summers, Michael Yarmolinsky.
The work was supported by generous grants from
The Danish Medical Research Council, The Knut and
Alice Wallenberg Foundation, The Swedish Cancer
Society, The Swedish Natural Sci-ence Research
Council (now The Swedish Re-search Council), and
The Wennergren Foundations.
References
Achtmann, M., Kusecek, B., Timmis, K.N., 1978. tra Cistrons and
proteins encoded by the Escherichia coli antibiotic resistance
plasmid R6-5. Mol. Gen. Genet. 163, 160179.
Austin, S., Nordstrom, K., 1990. Partition-mediated incompatibility of bacterial plasmids. Cell 60, 351354.
Bahk, J.-D., Kioka, N., Sakai, H., Komano, T., 1988. Runawayreplication plasmid pSY343 contains two ssi signals. Plasmid
20, 266270.
Bazaral, M., Helinski, D.R., 1970. Replication of a bacterial
plasmid and an episome in Escherichia coli. Biochemistry 9,
399406.
Bernander, R., Krabbe, M., Nordstrom, K., 1992. Mapping of the
in vivo start site for leading strand DNA synthesis in plasmid
R1. EMBO J. 11, 44814487.
Blohm, D., Goebel, W., 1978. Restriction map of the antibiotic
resistance plasmid R1 drd-19 and its derivatives pKN102 R1
(drd-19B2) and R1 drd-16 for the enzymes BamH1, HindIII,
EcoRI and SalI. Mol. Gen. Genet. 167, 119127.
Blomberg, P., Wagner, E.G.H., Nordstrom, K., 1990. Control of
replication of plasmid R1: the duplex between the antisense
RNA antisense RNA, CopA, and its target, CopT, is processed
specifically in vivo and in vitro by RNase III. EMBO J. 9,
23312340.
Blomberg, P., Nordstrom, K., Wagner, E.G.H., 1992. Control of
replication of plasmid R1: RepA synthesis is regulated by
CopA RNA through a leader peptide translation. EMBO J. 11,
26752683.
Blomberg, P., Engdahl, H.M., Malmgren, C., Romby, P., Wagner,
E.G.H., 1994. Replication control of plasmid R1: disruption of
an inhibitory RNA structure that sequesters the repA ribosomebinding site permits tap-independent RepA synthesis. Mol.
Microbiol. 12, 4660.
Bogan, J.A., Grimwade, J.E., Thornton, M., Zhou, P., Denning,
G.D.C., Helmstetter, C.E., 2001. P1 and NR1 plasmid
replication during the cell cycle of Escherichia coli. Plasmid
45, 200208.
Brady, G., Frey, J., Danbara, H., Timmis, K.N., 1983. Replication
control mutations of plasmid R6-5 and their eects on
interactions of the RNA-I control element with its target. J.
Bacteriol. 154, 429436.
Bramhill, D., Kornberg, A., 1988. A model for initiation at origins
of DNA replication. Cell 54, 915918.
23
Diaz, R., 1987. Identification of

y system of plasmid R1, ParD,
eplication of this plasmid, Mol.
96. Modulation of chemical

eters of the cell by growth rate.
cherichia coli and Salmo-nella:
y. ASM Press, Wash-ington, DC,
8. Stability of pBR322-derived
0.
., 1981. The structure of R1 drdthe plasmid. Mol. Gen. Genet.
68. Chromosome replication and

erichia coli B/r. J. Mol. Biol. 31,
of F lac replication and chroatl. Acad. Sci. USA 69, 2706
make big splash. Science 298,
004. RNaseIII enzymes and the

ature Struct. Mol. Biol. 11, 214
is, J.K., Timmis, K.N., 1981.

on: target determinant of the
of plasmid R6-5 lies within a
Natl. Acad. Sci. USA 78, 4699
1965. Penicillinase synthesis

ors in Enterobacteriaceae. Nature
973. Control of F lac replication

eriol. 114, 294 299.
Echevarra, M.J., Espinosa, M.,

cation and control of circular
Mol. Rev. 62, 434 464.
on of plasmid R12 replication in

scription by RNA polymerase.
191.
2. Origin and direction of miniin cell extracts of Escherichia
084.
enbauer, W., 1981. Plasmid R1
in synthesis in cell-free extracts
9, 10771079.
. The incompatibility product of
gene expression in the plasmid
152, 829839.
Gerdes, K., 2005. Partitionused by random assortment of
crobiol. 56, 14301440.
-i., 1991. Antisense RNA. Annu.
Ehrenberg, M., Sverredal, A., 1995. A model for copy number

control of the plasmid R1. J. Mol. Biol. 246, 472485.
Eliasson, A., Bernander, R., Dasgupta, S., Nordstrom, K., 1992.
Direct visualisation of plasmid DNA in bacterial cells. Mol.
Engberg, B., Nordstrom, K., 1975. Replication of the R-factor R1
in Escherichia coli K-12 at dierent growth rates. J. Bacteriol.
123, 179186.
Funnell, B.E., 2005. Partition-mediated plasmid pairing. Plas-mid
53, 119125.
Funnell, B.E., Baker, T.A., Kornberg, A., 1986. Complete
enzymatic replication of plasmids containing the origin of the
Escherichia coli chromosome. J. Biol. Chem. 261, 5616 5624.
Gerdes, K., Molin, S., 1985. Partitioning of plasmid R1. Structural

and functional analysis of the parA locus. J. Mol. Biol. 190,
269279.
Gerdes, K., Larsen, J.E.L., Molin, S., 1985. Stable inheritance of
plasmid R1 requires two dierent loci. J. Bacterol. 161, 292
298.
Gerdes, K., Rasmussen, P.B., Molin, S., 1986. Unique type of
plasmid maintenance function: postsegregational killing of
plasmid-free cells. Proc. Natl. Acad. Sci. USA 83, 3116 3120.
Gerdes, K., Gultyaev, A.P., Franch, T., Pedersen, K., Mikkel-sen,
N.D., 1997. Antisense RNA-regulated programmed cell death.
Annu. Rev. Genet. 31, 131.
Givskov, M., Molin, S., 1984. Copy mutants of plasmid R1:
eects of base pair substitutions in the copA gene on the
replication control system. Mol. Gen. Genet. 194, 286 292.
Givskov, M., Stougaard, P., Light, J., Molin, S., 1987.
Identification and characterization of mutations responsible for
a runaway replication phenotype of plasmid R1. Gene 57, 203
211.
Gordon, G.S., Wright, A., 2000. DNA segregation in bacteria.
Annu. Rev. Microbiol. 54, 681708.
Gustafsson, P., Nordstrom, K., 1975. Random replication of the
stringent plasmid R1 in Escherichia coli K-12. J. Bacteriol.
123, 443448.
Gustafsson, P., Nordstrom, K., 1980. Control of replication of
plasmid R1. Kinetics of replication in shifts between dierent
copy numbers. J. Bacteriol. 141, 106110.
Gustafsson, P., Nordstrom, K., Perram, J.W., 1978. Selection and
timing of replication of plasmids R1drd-19 and F lac in
Escherichia coli. Plasmid 1, 187203.
Helmstetter, C.E., Cooper, S., 1968. DNA synthesis during the cell
division cycle of three substrains of rapidly growing
Escherichia coli B/r. J. Mol. Biol. 31, 507518.
Hill, T.M., Pelletier, A.J., Tecklenburg, M.L., Kuempel, P.L., 1988.
Identification of the DNA sequence from the E. coli terminus
region that halts replication forks. Cell 55, 459 466.
Hjalt, T.A.H., Wagner, E.G.H., 1992a. The eect of loop size in

antisense and target RNAs on the eciency of antisense RNA
control. Nucleic Acids Res. 20, 67236732.
24
Hjalt, T.A.H., Wagner, E.G.H., 1992b. Bulged-out nucleotides

protect an antisense RNA from RNase III cleavage. Nucleic
Acids Res. 23, 571579.
Hjalt, T.A.H., Wagner, E.G.H., 1995. Bulged-out nucleotides in an
antisense RNA are required for rapid target RNA binding in
vitro and inhibition in vivo. Nucleic Acids Res. 23, 580587.
Ho, T.Q., Zhong, Z.P., Aung, S., Pogliano, J., 2002. Compat-ible
bacterial plasmids are targeted to independent cellular locations
in Escherichia coli. EMBO J. 21, 18641872.
Itoh, T., Tomizawa, J.-i., 1980. Formation of an RNA primer for
initiation of replication of ColE1 DNA by ribonuclease H.
Proc. Natl. Acad. Sci. USA 77, 24502454.
Izant, J.-G., Weintraub, H., 1984. Inhibition of thymidine kinase
gene expression by antisense RNA: a molecular approach to
genetic analysis. Cell 36, 10071015.
Kollek, R., Oertel, W., Goebel, W., 1978. Isolation and
characterization of the minimal fragment required for
autonomous replication of a copy mutant (pKN102) of the
antibiotic resistance factor R1. Mol. Gen. Genet. 162, 5158.
Koppes, L., Nordstrom, K., 1986. Insertion of an R1 plasmid into
the origin of replication of the E. coli chromosome: random
timing of replication of the hybrid chromosome. Cell 44, 117
124.
Koppes, L.J., Woldringh, C.L., Nanninga, N., 1999. Esche-richia
coli contains DNA replication compartments in the cell center.
Biochimie 81, 803810.
Krabbe, M. 1995. In vivo studies of Escherichia coli. Bacteriophage T4 endonuclease II-dependent restriction and initia-tion
of plasmid R1 replication. Thesis, Uppsala University.
Krabbe, M., Zabielski, J., Bernander, R., Nordstrom, K., 1997.
Inactivation of the replication-termination system aects the
replication mode and causes unstable maintenance of plasmid
R1. Mol. Microbiol. 24, 723735.
Larsen, J.E.L., Gerdes, K., Light, J., Molin, S., 1984. Low-copynumber plasmid-cloning vectors amplifiable by dere-pression
of an inserted foreign promoter. Gene 28, 4554.
Lee, S.B., Bailey, J.E., 1984a. A mathematical model for kdv
plasmid replication: analysis of wild-type plasmid. Plasmid 11,
151165.
Lee, S.B., Bailey, J.E., 1984b. A mathematical model for kdv
plasmid replication: analysis of copy number mutants. Plasmid
11, 166177.
Lemon, K.P., Grossman, A.D., 1999. Localization of bacterial
DNA polymerase: evidence for a factory model of replica-tion.
Science 282, 14301431.
Lemon, K.P., Grossman, A.D., 2000. Movement of replicating
DNA through a stationary factory. Mol. Cell. 6, 13211330.
Lemon, K.P., Grossman, A.D., 2001. The extrusion-capture model
for chromosome partitioning in bacteria. Genes Dev. 15, 2031
2041.
Levin-Zaidman, S., Reich, Z., Wachtel, E.J., Minsky, A., 1996.
Flow of structural information between four DNA conformational levels. Biochemistry 35, 29852991.
Light, J., Molin, S., 1981. Replication control functions of plasmid
R1 act as inhibitors of expression of a gene required for
replication. Mol. Gen. Genet. 184, 5661.
Light, J., Molin, S., 1982. The sites of action of the two copy
number control functions of plasmid R1. Mol. Gen. Genet.
187, 486493.
Light, J., Molin, S., 1983. Post-transcriptional control of
expression of the repA gene of plasmid R1 mediated by a small
RNA molecule. EMBO J. 2, 9398.
Light, J., Riise, E., Molin, S., 1985. Transcription and its
regulation in the basic replicon of plasmid R1. Mol. Gen.
Genet. 198, 503508.
Lbner-Olesen, A., 1999. Distribution of minichromosomes in
individual Escherichia coli cells: implications for replication
control. EMBO J. 18, 17121721.
Maas, R., Wang, C., 1997. Role of the RepA1 protein in RepFIC
plasmid replication. J. Bacteriol. 179, 21632168.
Maas, R., Wang, C., Maas, W.K., 1997. Interactions of the RepA1
protein with its replicon targets: two opposing roles in control
of plasmid replication. J. Bacteriol. 179, 38233827.
Marians, K.J., Minden, J.S., Parada, C., 1986. Replication of
superhelical DNAs in vitro. Prog. Nucleic Acids Res. Mol.
Biol. 33, 111140.
Masai, H., Arai, K., 1987. RepA and DnaA proteins are required
for initiation of R1 plasmid replication in vitro and interact
with the oriR sequence. Proc. Natl. Acad. Sci. USA 84, 4781
4785.
Masai, H., Arai, K.-I., 1988. RepA protein- and oriR-dependent
initiation of R1 plasmid replication: identification of rhodependent transcription terminator required for cis-action of
repA protein. Nucleic Acids Res. 16, 64936514.
Masai, H., Arai, K., 1989. Leading strand synthesis of R1 plasmid
replication in vitro is primed by primase alone at a specific site
downstream of oriR. J. Biol. Chem. 264, 8082 8090.
Masai, H., Kaziro, Y., Arai, K., 1983. Definition of oriR, the
minimum DNA segment essential for initiation of R1 plasmid
replication in vitro. Proc. Natl. Acad. Sci. USA 80, 68146818.
Meinhardt, H., de Boer, P., 2001. Pattern formation in Escherichia
coli: a model for the pole-to-pole oscillations of Min proteins
and the localization of the division site. Proc. Natl. Acad. Sci.
USA 98, 1420214207.
Meselson, M., Stahl, F.C., 1958. The replication of DNA in
Escherichia coli. Proc. Natl. Acad. Sci. USA 44, 671682.
Meynell, E., Datta, N., 1966. The relation of resistance transfer
factors to the F-factor (sex-factor) of Escherichia coli K12.
Genet. Res. 7, 134140.
Miki, T., Easton, A.M., Rownd, R.H., 1980. Cloning of replication,
incompatibility and stability functions of R plasmid NR1. J.
Bacteriol 141, 8799.
Miyazaki, C., Kawai, Y., Ohtsubo, H., Ohtsubo, E., 1988.
Unidirectional replication of plasmid R100. J. Mol. Biol. 204,
331343.
Morris, C.F., Hashimoto, H., Mickel, S., Rownd, R., 1974. Round
of replication mutant of a drug resistance factor. J. Bacteriol.
118, 855866.
Mller-Jensen, J., Jensen, R.B., Lowe, J., Gerdes, K., 2002.
Prokaryotic DNA segregation by an actin-like filament. EMBO
J. 18, 17121721.

Morrison, P.F., Chattoraj, D.K., 2004. Replication of a unit-copy
plasmid F in the bacterial cell cycle: a replication rate function
analysis. Plasmid 552, 1330.
Nielsen, P.F., Molin, S., 1984. How the R1 replication control
system responds to copy number deviations. Plasmid 11, 264
267.
Nomura, N., Masai, N., Inuzuka, M., Miyazaki, C., Ohtsubo, E.,
Itoh, T., Sasamoto, S., Matsui, M., Ishizaki, R., Arai, K.-I.,
1991. Identification of eleven single-strand initiation sequences
(ssi) for priming of DNA replication in the F, R6K, R100 and
ColE2 plasmids. Gene 108, 1522.
Nordstrom, K., 1969. A mutated R-factor mediating increased
resistance to several antibiotics. J. Gen. Microbiol. 57, xxxi
xxxii.
Nordstrom, K., 1971. Increased resistance to several antibiotics by
one mutation in an R-factor. J. Gen. Microbiol. 66, 205 214.
Nordstrom, K., 1983. Replication of plasmid R1: Meselson Stahl
density shift experiments revisited. Plasmid 9, 218 221.
Nordstrom, K., 1985. Control of plasmid replication: theoret-ical
considerations and practical solutions. In: Helinski, D.R.,
Cohen, S.N., Clewell, D.B., Jackson, D.A., Holla-ender, A.
(Eds.), Plasmids in Bacteria. Plenum Publishing Corporation,
pp. 189214.
Nordstrom, K., 1990. Control of plasmid replicationhow do
DNA iterons set the replication frequency? Cell 63, 1121
1124.
Nordstrom, K., 2002. Replication and stable maintenance of
plasmid R1. Recent Res. Develop. Plasmid Biol. 1, 7593.
Nordstrom, K., Aagaard-Hansen, K., 1984. Maintenance of
bacterial plasmids: comparison of theoretical calculations and
experiments with plasmid R1. Mol. Gen. Genet. 197, 1 7.
Nordstrom, K., Austin, S., 1989. Mechanisms that contribute to
the stable segregation of plasmids. Annu. Rev. Genet. 23, 37
69.
Nordstrom, K., Dasgupta, S., 2001. Partitioning of the Escherichia coli chromosome: superhelicity and condensation.
Biochimie 83, 4148.
Nordstrom, K., Gerdes, K., 2003. Clustering versus random
segregation of plasmids lacking a partitioning function: a
plasmid paradox? Plasmid 50, 95101.
Nordstrom, K., Uhlin, B.E., 1992. Runaway-replication plas-mids
as tools to produce large quantities of proteins from cloned
genes in bacteria. Bio/Technol. 10, 661666.
Nordstrom, K., Wagner, E.G.H., 1994. Kinetic aspects of control
of plasmid replication by antisense RNA. Trends Biochem. Sci.
19, 294300.
Nordstrom, K., Ingram, L.C., Lundback, A., 1972. Mutations in
R-factors of Escherichia coli causing an increased number of
R-factor copies per chromosome. J. Bacteriol. 111, 562 569.
Nordstrom, K., Molin, S., Aagaard-Hansen, H., 1980a. Partitioning of plasmid R1 in Escherichia coli. I. Kinetics of loss of
plasmid derivatives deleted of the par region. Plasmid 4, 215
227.
25
Nordstrom, K., Molin, S., Aagaard-Hansen, H., 1980b.

Partitioning of plasmid R1 in Escherichia coli. II. Incompatibility properties of the partitioning system. Plasmid 4, 332
349.
Nordstrom, K., Molin, S., Light, J., 1984. Control of replica-tion
of bacterial plasmids: genetics, molecular biology, and
physiology of the plasmid R1 system. Plasmid 12, 7190.
Novick, R.P., Hoppensteadt, F.C., 1978. On plasmid incompatibility. Plasmid 1, 421434.
Olsson, J. 2003. Control of chromosome and plasmid replica-tion
in Escherichia coli. Thesis, Uppsala University.
Olsson, J., Nordstrom, K., Berg, O., Dasgupta, S., 2002. Eclipse
period without sequestration in Escherichia coli. Mol.
Microbiol. 44, 1142911440.
Olsson, J.A., Berg, O.G., Dasgupta, S., Nordstrom, K., 2003.
Eclipse period during replication of plasmid R1: contribu-tions
from structural events and from the copy-number control
system. Mol. Microbiol. 50, 291301.
Olsson, J.A., Paulsson, J., Nordstrom, K., 2004. Eect of the
CopB auxiliary replication control system on stability of
+
maintenance of Par plasmid R1. J. Bacteriol. 186, 207 211.
Onogi, T., Niki, T., Hiraga, S., 2002. Behavior of sister copies of
mini-F plasmid after synchronized plasmid replication in
Escherichia coli. J. Bacteriol. 184, 31423145.
Paulsson, J., Ehrenberg, M., 2000. Molecular clocks reduce
plasmid loss rates: the R1 case. J. Mol. Biol. 297, 179192.
Paulsson, J., Ehrenberg, M., 2001. Noise in a minimal regulatory
network: plasmid copy number control. Quar-terly Revs
Biophys. 34, 159.
Paulsson, J., Nordstrom, K., Ehrenberg, M., 1998. Require-ments
for rapid plasmid ColE1 copy number adjustments: a
mathematical model of inhibition modes and RNA turnover
rates. Plasmid 39, 215234.
Persson, C., Wagner, E.G.H., Nordstrom, K., 1988. Control of
replication of plasmid R1: kinetics of in vivo interaction
between the antisense RNA, CopA, and its target, CopT.
EMBO J. 7, 32793288.
Persson, C., Wagner, E.G.H., Nordstrom, K., 1990a. Control of
replication of plasmid R1: formation of an initial transient
complex is rate-limiting for antisense RNA/target RNA pairing
and the Appendix by Kiselman, K., Nord-strom, K. A
mathematical analysis of the kinetics of duplex formation.
EMBO J. 9, 37773785.
Persson, C., Wagner, E.G.H., Nordstrom, K., 1990b. Control of
replication of plasmid R1: structures and sequences of the
antisense RNA, CopA, required for its binding to the target,
CopT. EMBO J. 7, 37673775.
Pogliano, J., Ho, T.Q., Zhong, Z.P., Helinski, D.R., 2001.
Multicopy plasmids are clustered and localized in Esche-richia
coli. Proc. Natl. Acad. Sci. USA 98, 44864491.
Pritchard, R.H., 1978. Control of DNA replication in bacteria. In:
Molineux, I., Kohiyama, M. (Eds.), DNA Synthesis Present
and Future. Plenum, New York, pp. 126.
Pritchard, R.H., Barth, P.T., Collins, J., 1969. Control of DNA
synthesis in bacteria. Symp. Soc. Gen. Microbiol. 19, 263
297.
26
Reich, Z., Wachtel, E.J., Minsky, A., 1994a. Liquid-crystalline

mesophases of plasmid DNA in bacteria. Science 264, 1460
1463.
Reich, Z., Levin-Zaidman, S., Gutman, S.B., Arad, T., Minsky, A.,
1994b. Supercoiling-regulated liquid-crystalline packag-ing of
topologically-constrained, nucleosome-free DNA molecules.
Biochemistry 33, 1417714184.
Riise, E., Molin, S., 1986. Purification and characterization of the
CopB replication control protein, and precise mapping of its
target site in the R1 plasmid. Plasmid 15, 163171.
Riise, E., Stougaard, P., Bindslev, B., Nordstrom, K., Molin, S.,
1982. Molecular cloning and functional characterization of a
copy number control gene (copB) of plasmid R1. J. Bacteriol.
151, 11361145.
Rosen, J., Ryder, T., Inokuchi, H., Ohtsubo, H., Ohtsubo, E., 1980.
Genes and sites involved in replication and incom-patibility of
an R100 plasmid derivative based on nucleotide sequence
analysis. Mol. Gen. Genet. 179, 527537.
Rosenfeld, R., Grover, N.B., 1993. Control of mini-R1 replication: a computer simulation. Plasmid 29, 94116.
Rownd, R., 1969. Replication of a bacterial episome under relaxed
control. J. Mol. Biol. 44, 387402.
Sawitzke, J., Austin, S., 2001. An analysis of the factory model for
chromosome replication and segregation in bacteria. Mol.
Schwartzman, J.B., Stasiak, A., 2004. A topological view of the
replicon. EMBO Reports 5, 256261.
Sharp, P.A., Cohen, S.N., Davidson, N., 1973. Electron microscope
heteroduplex studies of sequence relations among plasmids of
Escherichia coli. II Structure of drug-resistance (R) factors and
F factors. J. Mol. Biol. 73, 235 255.
Simons, R.W., Kleckner, N., 1983. Translational control of IS10
transposition. Cell 34, 683691.
Soderbom, F., Wagner, E.G.H., 1998. Degradation pathway of
CopA, the antisense RNA that controls replication of plasmid
R1. Microbiology 144, 19071917.
Soderbom, F., Binnie, U., Masters, M., Wagner, E.G.H., 1997.
Regulation of plasmid R1 replication: PncB and RNaseE
expedite the decay of the antisense RNA, CopA. Mol.
Stougaard, P., Molin, S., Nordstrom, K., 1981a. RNA mole-cules
involved in copy number control and incompatibility of
plasmid R1. Proc. Natl. Acad. Sci. USA 78, 60086012.
Stougaard, P., Molin, S., Nordstrom, K., 1981b. The nucleotide
sequence of the replication control region of the resistance
plasmid R1 drd-19. Mol. Gen. Genet. 181, 116122.
Stougaard, P., Light, J., Molin, S., 1982. Convergent transcrip-tion
interferes with expression of the copy number control gene,
copA, from plasmid R1. EMBO J. 1, 323328.
Summers, D.K., Sherratt, D.J., 1984. Multimerisation of high copy
number mutants causes instability: ColE1 encodes a
determinant essential for plasmid monomerisation and stability.
Cell 36, 10971103.
Summers, D.K., Beton, C.W., Withers, H.L., 1993. Multicopy
plasmid instability: the dimer catastrophe hypothesis. Mol.
Microbiol. 8, 10311038.
Tomizawa, J.-i., 1984. Control of ColE1 replication: the process of

binding of RNA I to the primer transcript. Cell 38, 861 870.
Tomizawa, J.-i., 1986. Control of ColE1 replication: binding of
RNA I to RNA II and inhibition of primer formation. Cell 47,
8997.
Tsutsui, H., Matsubara, K., 1981. Replication control and switcho function as observed with a mini-F factor plasmid. J.
Bacteriol. 147, 509516.
Uhlin, B.E., Nordstrom, K., 1977. R plasmid gene dosage eects
in Escherichia coli: copy mutants of the R plasmid R1 drd-19.
Plasmid 1, 17.
Uhlin, B.E., Nordstrom, K., 1978. A runaway-replication mutant
of plasmid R1 drd-19: temperature-dependent loss of copy
number control. Mol. Gen. Genet. 165, 167179.
Uhlin, B.E., Molin, S., Gustafsson, P., Nordstrom, K., 1979.
Plasmid cloning vehicles with temperature-dependent copy
number control: amplification of genes and gene products.
Gene 6, 91106.
Vogel, J., Argaman, L., Wagner, E.G.H., Altuvia, S., 2004. The
small RNA IstR inhibits synthesis of an SOS-induced toxic
peptide. Curr. Biol. 14, 22712276.
Wagner, E.G.H., Nordstrom, K., 1986. Structural analysis of an
RNA molecule involved in replication control of plasmid R1.
Nucleic Acids. Res. 14, 25232538.
Wagner, E.G.H., Simons, R.W., 1994. Antisense RNA control in
bacteria, phages, and plasmids. Annu. Rev. Microbiol. 48, 713
742.
Wagner, E.G.H., Altuvia, S., Romby, P., 2002. Antisense RNAs in
bacteria and their genetic elements. Adv. Genet. 46, 361398.
Wang, J.D., Rokop, M.E., Barker, M.M., Hanson, N.R., Grossman,
A.D., 2004. Multicopy plasmids aect replisome partitioning
in Bacillus subtilis. J. Bacteriol. 186, 70847090.
Weitao, T., Dasgupta, S., Nordstrom, K., 2000a. Plasmid R1 is
present as clusters in the cells of Escherichia coli. Plasmid 43,
200204.
Weitao, T., Dasgupta, S., Nordstrom, K., 2000b. Role of the mukB
gene in chromosome and plasmid partition in Escherichia coli.
Mol. Microbiol. 38, 392400.
Womble, D.D., Rownd, R.H., 1986. Regulation of IncFII plasmid
DNA replication. A quantitative model for control of plasmid
NR1 replication in the bacterial cell division cycle. J. Mol.
Biol. 192, 529548.
Womble, D.D, Sampathkumar, P., Easton, A.M., Luckow, V.A.,
Rownd, R.H., 1985. Transcription of the replication control
region of the IncFII R-plasmid NR1 in vivo and in vitro. J.
Mol. Biol. 181, 395410.
Wu, L.J., 2004. Structure and segregation of the bacterial nucleoid.
Curr. Opin. Genet. Dev. 14, 126132.
Yoshikawa, M., 1974. Identification and mapping of the replication
genes of an R-factor, R100-1, integrated into the host
chromosome of Escherichia coli K-12. J. Bacteriol. 118, 1123
1131.
Zeuthen, J., Pato, M.L., 1971. Replication of the F lac sex factor in
the cell cycle of Escherichia coli. Mol. Gen. Genet. 111, 242
255.

Articulo Micro

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Articulo Micro

Uploaded by

Copyright:

Available Formats

Plasmid 55 (2006) 126

Plasmid R1Replication and its control

Fax: +46 18 53 03 96.

K. Nordstrom / Plasmid 55 (2006) 126

Fig. 1. Genetic map of plasmid R1, showing the basic replicon

2. The basic R1 replicon

RepA protein is a typical ini-tiator protein

K. Nordstrom / Plasmid 55 (2006) 126

1998; Masai and Arai, 1987; Masai et al., 1983) in a

Fig. 2. The basic replicon of plasmid R1 consists of an origin of

33,000 (Del Solar et al., 1998; Rosen et al., 1980;

Fig. 3. The start region for leading strand synthesis of plasmid R1

the DnaA protein to the chromosomal origin, oriC

K. Nordstrom / Plasmid 55 (2006) 126

4. Replication of plasmid R1 is random in time and

studies and showed that selection of plasmid NR1 for

K. Nordstrom / Plasmid 55 (2006) 126

has two heavy strands (HH), after replication HL

Relative amount of HL-DNA x2

HLmax 0.5 Ds=2.

It should also be stressed that the eclipse is an

Relative amount of LL-DNA x3

Plasmids are present at defined copy numbers that

K. Nordstrom / Plasmid 55 (2006) 126

number of R1 copies per oriC (origin of the chromosome) is about 1.

Fig. 4. Kinetics of control of plasmid replication (Nordstrom et

sures its concentration and uses this measurement to

K. Nordstrom / Plasmid 55 (2006) 126

almost a step function (Tsutsui and Matsubara,

Fig. 5. Secondary structure of CopA-RNA and the locations of cop

The sequence of the CopA-RNA (identical in R1,

K. Nordstrom / Plasmid 55 (2006) 126

8. CopA-RNA inhibits translation of the repA

turned out that RepA synthesis is translationally

K. Nordstrom / Plasmid 55 (2006) 126

daughters containing both plasmid derivatives. If the

The equation is approximate, because all cells at

Fig. 7. Incompatibility due to random selection for replication and

Mutations in the copA/copT region of plasmid R1

K. Nordstrom / Plasmid 55 (2006) 126

Wild-type CopA/mutant CopT

Mutant CopA/wild-type CopT

Wild type lost

Miss-matched base pairs are shown in bold letters.

there are unpaired bases in both combinations with

Fig. 8. Compatible R1 cop mutants due to cis-acting Rep proteins.

cis case, since the RepA molecules formed do not belong to a

K. Nordstrom / Plasmid 55 (2006) 126

Fate of daughter cells

(Table 2). However, if the cells harbours another

10. Parameters that determine the copy number of

Fig. 9. Schematic diagram showing the reactions that deter-mine

in the R1 copy number (Soderbom and Wagner,

K. Nordstrom / Plasmid 55 (2006) 126

11. Runaway replication

Fig. 10. Eect of convergent transcription on the copy number of

The kpL promoter is much stronger than the pcopB

mutant is much higher than that of the

K. Nordstrom / Plasmid 55 (2006) 126

12. The eclipse is determined by structural factors

Fig. 11. Transcription pattern of a runaway-replication mutant of

lation. After shift from 30 to 42 LC, the plasmid

13. Copy-number distributions

K. Nordstrom / Plasmid 55 (2006) 126