Microbiology (2011), 157, 2220–2225
DOI 10.1099/mic.0.049478-0
Organization of ribonucleoside diphosphate
reductase during multifork chromosome replication
in Escherichia coli
Marı́a Antonia Sánchez-Romero,1,2 Felipe Molina1
and Alfonso Jiménez-Sánchez1
Correspondence
Marı́a Antonia Sánchez-Romero
[email protected]
Received 7 March 2011
Revised
23 May 2011
Accepted 27 May 2011
1
Departamento de Bioquı́mica y Biologı́a Molecular y Genética, Universidad de Extremadura,
Badajoz E06080, Spain
2
School of Biosciences, The University of Birmingham, Edgbaston, Birmingham B15 2TT, UK
Ribonucleoside diphosphate reductase (RNR) is located in discrete foci in a number that
increases with the overlapping of replication cycles in Escherichia coli. Comparison of the
numbers of RNR, DnaX and SeqA protein foci with the number of replication forks at different
growth rates reveals that fork : focus ratios augment with increasing growth rates, suggesting a
higher cohesion of the three protein foci with increasing number of forks per cell. Quantification of
NrdB and SeqA proteins per cell showed: (i) a higher amount of RNR per focus at faster growth
rates, which sustains the higher cohesion of RNR foci with higher numbers of forks per cell; and
(ii) an equivalent amount of RNR per replication fork, independent of the number of the latter.
INTRODUCTION
Slowly and rapidly growing bacterial cells face different
organizational challenges during chromosome replication
and segregation. In slowly growing bacteria, the cell cycle is
similar to that of eukaryotic cells, with initiation and
termination of chromosome replication taking place in the
same cycle. In rapidly growing bacteria, replication can be
longer than a generation time, and initiation and termination occur in different cycles. If replication initiates before
the previous round is completed, replication cycles overlap
and chromosomes have multiple replication forks assembled
at different cell cycles (Cooper & Helmstetter, 1968).
Accordingly, chromosome replication is controlled at its
initiation step to ensure that only one initiation event takes
place at each cell cycle, which, independently of the time
required for replicating the entire chromosome, assures that
only one chromosome termination takes place every cell
cycle to allow cell division.
Colocalization of the DNA polymerases from the two forks
originating from a single initiation event was initially
proposed by Dingman et al. (1974). Other authors suggested
that forks are paired for a time and then move in opposite
directions (Bates & Kleckner, 2005; Hiraga et al., 1998;
Reyes-Lamothe et al., 2010). Thus, organization of the
replication apparatus is still controversial. Recent works
Abbreviations: CAA, Casamino acids; HA, haemagglutinin; RNR,
ribonucleoside diphosphate reductase.
A supplementary figure and movie are available with the online version
of this paper.
2220
have reported a dynamic organization of the replication
forks and sister chromosomes: the extent of fork colocalization was found to be dependent on the replication pattern
(Molina et al., 2008; Morigen et al., 2009), and origin
cohesion has been described as being dependent on growth
rate (Adachi et al., 2008; Fossum et al., 2007).
Chromosome replication requires synthesis of the four dNTPs
at a high rate. Ribonucleoside diphosphate reductase (RNR) is
an essential enzyme for biosynthesis of the four dNTPs in all
living cells. The class I RNR, which predominates in
aerobically grown cells, is a tetramer made of two homodimers: subunit R1, encoded by the nrdA gene, and subunit
R2, encoded by the nrdB gene. There has been, for a long time,
a discrepancy between the high demand of dNTPs for every
replication fork and the low pools detected in bacteria. To
resolve this discrepancy, the replication hyperstructure model
postulates that all replicative proteins, including the dNTPbiosynthesis complex, are recruited at the initiation of the
replication process. According to this model, RNR embedded
in a hyperstructure would provide the dNTPs required for its
associated replication fork (Guzmán et al., 2002; Mathews
et al., 1993; Norris et al., 2007).
We have recently published the first evidence for a close
relationship between RNR and the replication proteins in
slow-growing bacteria (Guzmán et al., 2002; Sánchez-Romero
et al., 2010b). Here, we have performed the first study to our
knowledge on the organization of RNR, a replisome protein,
DnaX, and a replication-fork marker, SeqA, in cells with
different patterns of chromosome replication and we have
compared them with the replication forks.
Downloaded from www.microbiologyresearch.org by
049478 G 2011 SGM
IP: 88.99.165.207
On: Sun, 18 Jun 2017 13:10:39
Printed in Great Britain
Organization of RNR during DNA replication
METHODS
Bacterial strains and growth conditions. The strains used were
derived from Escherichia coli K-12 CM735 (metE46, trp3, his4, thi1,
galK2, lacY1, mtl1, ara9, tsx3, ton1, rps8, supE44). CMT931 contains a
36FLAG-tagged nrdB gene and CMT933 contains a haemagglutinin
(HA)-tagged dnaX gene (Sánchez-Romero et al., 2010b). Bacteria
were incubated at 37 uC in M9 minimal medium containing glycerol,
glucose or glucose plus Casamino acids (CAA), and samples were
taken at mid-exponential phase (OD550 0.1).
Cell-cycle parameters. Growth rates of exponentially growing
bacteria were determined by the OD550 of the cultures. Cell-cycle
parameters and the number of forks per cell were obtained for flow
cytometry. For these measurements, cells were treated with 150 mg
rifampicin ml21 and 50 mg cephalexin ml21 to inhibit initiation of
replication and cell division, respectively. After 3 h treatment, cells
were fixed and the number of chromosomes per cell was measured by
flow cytometry using a FACStar DIVA (Becton Dickinson) instrument, essentially as described previously (Molina & Skarstad, 2004).
The time required for chromosome replication, C, and the time from
the termination of replication to cell division, D, were determined
from the number of cells before and after the initiation of replication
in the initiation time equations (Jiménez-Sánchez & Guzmán, 1988).
Immunofluorescence microscopy. Subcellular location of RNR was
samples were taken for flow cytometry and fluorescence
microscopy. Cell-cycle parameters and the distribution of
replication forks per cell were obtained from flow-cytometry
DNA histograms (Fig. 1). As expected, the overlapping of
replication rounds increases as growth rate increases and so
does the number of replication forks. Whilst slowly growing
cells show no overlapping of replication cycles (Fig. 1d), fastgrowing cells show an extensive overlapping of consecutive
replication cycles (Fig. 1f) and carry up to 12 replication
forks for a short period of the cell cycle (Fig. 1i).
Fluorescence microscopy images show that RNR is located in
discrete foci in cells growing at any growth rate (Fig. 2,
Supplementary Fig. S1, Supplementary Movie S1). The
subcellular position of RNR foci in slow-growing cells reveals
that there are cells with a single focus localized at the middle
position of the cell, with two RNR foci localized at the onequarter and three-quarter positions of the cell, with three foci
at one-, two- and three-quarter positions, or with four RNR
foci distributed along the cell length (Fig. 3). This distribution
is essentially the same as has been described for the
replication-fork foci (Hiraga et al., 1998; Ohsumi et al., 2001).
The aim of this work was to investigate whether the
organization of RNR relies on the type of cell cycle, i.e. the
number of replication forks per cell. To this end, we studied
the distribution of the number of replication forks and that
of RNR foci per cell under different growth conditions.
Previous works have reported that as the overlapping of
replication cycles increases, so does the cohesion of
replication-fork and replication-protein foci. To assess
the degree of cohesion of RNR foci under different
replication patterns, the mean number of replication forks
per RNR focus was studied as a cohesion index. As foci
from coupled forks are detected as a single focus,
fork : focus ratio has been used as a way to understand
the organization of fork-associated proteins (Molina et al.,
2008). To compare the cohesion of RNR foci with that of
replication forks and other replication-protein foci, fork :
focus ratios corresponding to cultures with different
replication patterns from previously published data
(Adachi et al., 2008; Bates & Kleckner, 2005; Brendler
et al., 2000; Fossum et al., 2007; Molina & Skarstad, 2004;
Sánchez-Romero et al., 2010b; Sunako et al., 2001) and
from our results were contrasted (Fig. 4). In cells with one
replication cycle per chromosome, i.e. two forks per cell,
the number of RNR foci corresponds to the number of
replication forks throughout the cell cycle. Therefore, the
sister replication forks, each associated with an RNR focus,
move away from each other and show the lowest degree of
cohesion (Fig. 4). An increase in the number of forks per
cell is correlated with an increase in the cohesion degree of
forks and foci, with values even higher than full cohesion
(i.e. more than two forks per focus), suggesting the
transitory grouping of forks from two consecutive
replication cycles. The observed cohesion of RNR foci
shows a high correlation with published data when cells
were grown in glycerol or glucose medium, but this
cohesion was lower than expected for the fast-growing
bacteria in glucose plus CAA (fork : focus ratio of approx.
1.5, when a ratio of approx. 2 would be expected).
E. coli strain CMT931 was grown at 37 uC in medium
containing glycerol, glucose or glucose with CAA and
These results lead us to conclude that the organization of
RNR foci is dependent on the replication pattern and
analysed by using a FLAG-tagged RNR and immunolocalized by
treating ethanol-fixed cells with mouse monoclonal anti-FLAG
M2–Cy3 antibody (Sigma-Aldrich) in CMT931. Mouse monoclonal
anti-HA antibody and secondary anti-mouse antibody conjugated to
FITC were used to immunolocalize the t subunit of DNA polymerase
III (encoded by the dnaX gene) in strain CMT933. Immunostaining of
the SeqA protein was performed using a specific SeqA antibody as
described previously (Fossum et al., 2007; Sánchez-Romero et al.,
2010b). Nucleoids were stained with Hoechst 33258 and micrographs
were obtained by using an Olympus IX-70 Delta Vision fluorescence
microscope equipped with a 1006 UPLS Apo objective. Pictures were
taken using a CoolSNAP HQ/ICX285 camera and deconvolved using
the Delta Vision constrained iterative deconvolution software (Applied
Precision). The deconvolved images were saved in TIF format and
imported into ImageJ software (Wayne Rasband, Research Services
Branch, National Institute of Mental Health, MD, USA). Any
significant differences in focus scoring were found between deconvolved and not-deconvolved images (see Supplementary Fig. S1,
available with the online version of this paper).
Western blot. Whole-cell protein extracts in each medium were
analysed by Western blotting using anti-FLAG–Cy3 mouse monoclonal
antibody (Sigma-Aldrich) for RNR and specific antibodies against the
SeqA and OmpA proteins (gifts from K. Skarstad’s and I. Henderson’s
laboratories, respectively). Western blotting was repeated three times.
RESULTS AND DISCUSSION
Organization of RNR depends on the replication
overlapping
http://mic.sgmjournals.org
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 13:10:39
2221
M. A. Sánchez-Romero, F. Molina and A. Jiménez-Sánchez
Fig. 1. Organization of replication forks in exponentially growing strain CMT931 in minimal medium containing glycerol (a, d, g),
glucose (b, e, h) or glucose plus CAA (c, f, i). (a–c) DNA histograms of exponential cultures (left panels, grey lines) compared
with idealized computer-simulated histograms (black lines). Samples treated with rifampicin and cephalexin for three
generations (right panels, filled histograms) disclose the number of replication origins at the time of adding the drugs.
Experimental histograms survey 10 000 cells. (d–f) Number of replication forks (black triangles) and origins (grey circles) at
different cell ages, calculated by using parameters derived from flow-cytometry histograms. The horizontal black bar
corresponds to the replication period (C) and the horizontal white bar corresponds to the time between termination of
replication and subsequent cell division (D). t, Doubling time (min); C, replication time (min); D, period between the end of
replication and cell division (min) (see http://simon.bio.uva.nl/cellcycle/ for a comprehensive analysis). (g–i) Fraction of cells in
each of the three cell-cycle periods determined by the initiation, i, and termination, t, of replication from cell birth, b, to cell
division, d. The number of predicted replication forks per cell in each period is depicted above the bars.
reproduces that of other replication proteins in slowly
growing cells, whereas in fast-growing cells, the number of
RNR foci is somewhat greater than the number of foci of
replication proteins, which could indicate pre-assembly of
the required proteins before initiation of replication.
Correspondence among the number of foci of
RNR, DnaX and SeqA
To study the relationship between RNR foci and replicationprotein foci, we studied the number of foci of the tagged
replisome protein DnaX, the DNA polymerase III t subunit
and the SeqA protein, a binding protein for hemimethylated
GATC sites on newly replicated DNA, which has often been
used as a marker of replication forks (Molina et al., 2008;
Sánchez-Romero et al., 2010a).
Comparison of the mean number of RNR foci per cell with
those of DnaX and SeqA, and with replication forks shows
2222
that, in cells growing in glycerol, the numbers of foci per cell
of the three proteins are very similar to the number of forks
(Fig. 5). Furthermore, the mean number of foci of all
proteins and forks increases with increasing growth rate,
although not to the same extent (Fig. 5). The higher ratio of
fork number to RNR, SeqA and DnaX foci at high growth
rates suggests an increased grouping or cohesion of the three
protein foci with increasing number of forks per cell.
The amount of RNR protein per fork is constant at
any growth rate
As the number of replication forks per RNR focus was
found to increase with growth rate, we wanted to find out
how the amount of NrdB protein correlates with the
number of RNR foci and replication forks. For this
purpose, the levels of NrdB, SeqA and OmpA (an outermembrane protein) from bacteria growing in medium with
glycerol, glucose or glucose plus CAA were measured. As
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 13:10:39
Microbiology 157
Organization of RNR during DNA replication
Fig. 2. Selected fluorescence microscope images of strain
CMT931 growing in minimal medium containing glycerol (a),
glucose (b) or glucose plus CAA (c). RNR foci were detected by
Cy3 red fluorescence. Nucleoids were stained with Hoechst
33258 (blue). Bar, 1 mm. Constrained iterative deconvolution was
used for these images in order to highlight the position of the
fluorescence. Therefore, intensities are not quantifiable.
expected, the total amount of these proteins per cell
increased with increasing growth rates (Fig. 6a). To
estimate whether the amounts of NrdB and SeqA are
related to the number of forks under different growth
conditions, we analysed protein : fork ratio at different
growth rates. These results show that, essentially, the
amounts of NrdB and SeqA proteins per fork do not vary
with growth rate, whereas OmpA protein : fork ratio
increases with growth rate (Fig. 6b) and, consequently,
the amount of NrdB and SeqA proteins per focus increases.
Conclusions
Results reported in this work suggest a variable cohesion of
RNR foci that depends on the number of forks per cell, in a
similar way as has been reported for other replication proteins.
Comparative analysis of foci from RNR, DnaX (a replisome
protein) and SeqA (a marker for replicating forks) shows
that: (i) at low growth rate, most of the cells have the same
number of RNR foci as DnaX and SeqA protein foci;
(ii) the number of forks per cell increases with growth rate
at a higher level than the number of the protein foci,
denoting a higher cohesion of the foci at increasing
number of forks per cell; and (iii) the numbers of RNR and
DnaX foci are moderately higher than those of SeqA foci,
which might be explained by pre-assembly of the
http://mic.sgmjournals.org
Fig. 3. Subcellular localization of RNR foci in strain CMT931
growing exponentially at 37 6C in minimal medium containing
glycerol. Cells containing one, two, three or four foci (a–d,
respectively) are displayed separately. Each part of the figure is a
superimposition of the positions of the RNR foci from individual
cells.
replication proteins before initiation of replication (den
Blaauwen et al., 2006).
Measurement of the amounts of RNR, of the replicationfork marker SeqA and of OmpA, a protein independent of
replication, demonstrates a constant amount of RNR per
fork, independent of the number of forks per cell and of the
degree of RNR focus cohesion. Furthermore, the increased
amount of both proteins per focus supports the existence
of focus cohesion without disturbance of the composition
of every replication hyperstructure.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 13:10:39
2223
M. A. Sánchez-Romero, F. Molina and A. Jiménez-Sánchez
Fig. 4. Analysis of cohesion of RNR foci (&), SeqA foci (m),
published replication-protein foci (#) and hemimethylated DNA
foci (g) by comparing fork : focus ratio relative to fork : cell ratio.
These conclusions support the organization of the RNR
protein into a higher-order structure, or replication
hyperstructure, that does not depend on the association
between forks. Consequently, each replication fork would
incorporate its own machinery for the synthesis of its
required precursors that would run independently of other
hyperstructures, hence maximizing replication efficiency,
facilitating the cell’s ability to adapt in a changing
environment and leading, eventually, to cell viability.
Fig. 6. Arbitrary mean amounts of RNR (&), SeqA (m) and OmpA
(g) per cell (a), and their ratio to the number of forks (b) of strain
CMT931 growing in minimal medium with glycerol, glucose
or glucose plus CAA (growth rates were 0.73, 1.00 and
1.36 h”1, respectively).
ACKNOWLEDGEMENTS
We are very grateful to J. Jiménez and R. Daga of the Centro Andaluz
de Biologı́a del Desarrollo (Seville, Spain) for the use of their
microscope, to K. Skarstad for the anti-SeqA antibody and the use of
the flow cytometer, and to I. Henderson and D. Leyton for the antiOmpA antibody. We thank Encarna Ferrera for her excellent technical
help. This work was supported by grants from J. Extremadura
(PRI07A001) and MEC (Spain) (BFU2007-63942). M. A. S.-R.
acknowledges a studentship from J. Extremadura.
REFERENCES
Adachi, S., Fukushima, T. & Hiraga, S. (2008). Dynamic events of sister
chromosomes in the cell cycle of Escherichia coli. Genes Cells 13, 181–197.
Bates, D. & Kleckner, N. (2005). Chromosome and replisome
dynamics in E. coli: loss of sister cohesion triggers global chromosome
movement and mediates chromosome segregation. Cell 121, 899–911.
Brendler, T., Sawitzke, J., Sergueev, K. & Austin, S. (2000). A case for
sliding SeqA tracts at anchored replication forks during Escherichia
coli chromosome replication and segregation. EMBO J 19, 6249–6258.
Cooper, S. & Helmstetter, C. E. (1968). Chromosome replication and
the division cycle of Escherichia coli B/r. J Mol Biol 31, 519–540.
den Blaauwen, T., Aarsman, M. E. G., Wheeler, L. J. &
Nanninga, N. (2006). Pre-replication assembly of E. coli replisome
Fig. 5. Number of replication forks (#) and number of foci of RNR
(&), DnaX (h) and SeqA (m) proteins per cell in bacteria growing
in minimal medium containing glycerol, glucose or glucose plus
CAA (m50.73, 1.00 or 1.36 h”1, respectively).
2224
components. Mol Microbiol 62, 695–708.
Dingman, C. W., Fisher, M. P. & Ishizawa, M. (1974). DNA
replication in Escherichia coli: physical and kinetic studies of the
replication point. J Mol Biol 84, 275–295.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 13:10:39
Microbiology 157
Organization of RNR during DNA replication
Fossum, S., Crooke, E. & Skarstad, K. (2007). Organization of sister
Morigen, O., Odsbu, I. & Skarstad, K. (2009). Growth rate dependent
origins and replisomes during multifork DNA replication in Escherichia
coli. EMBO J 26, 4514–4522.
numbers of SeqA structures organize the multiple replication forks in
rapidly growing Escherichia coli. Genes Cells 14, 643–657.
Guzmán, E. C., Caballero, J. L. & Jiménez-Sánchez, A. (2002).
Norris, V., den Blaauwen, T., Doi, R. H., Harshey, R. M., Janniere, L.,
Jiménez-Sánchez, A., Jin, D. J., Levin, P. A., Mileykovskaya, E. &
other authors (2007). Toward a hyperstructure taxonomy. Annu Rev
Ribonucleoside diphosphate reductase is a component of the
replication hyperstructure in Escherichia coli. Mol Microbiol 43,
487–495.
Microbiol 61, 309–329.
Hiraga, S., Ichinose, C., Niki, H. & Yamazoe, M. (1998). Cell cycle-
Ohsumi, K., Yamazoe, M. & Hiraga, S. (2001). Different localization
dependent duplication and bidirectional migration of SeqAassociated DNA–protein complexes in E. coli. Mol Cell 1, 381–387.
of SeqA-bound nascent DNA clusters and MukF–MukE–MukB
complex in Escherichia coli cells. Mol Microbiol 40, 835–845.
Jiménez-Sánchez, A. & Guzmán, E. C. (1988). Direct proce-
Reyes-Lamothe, R., Sherratt, D. J. & Leake, M. C. (2010).
dure for the determination of the number of replication forks and
the reinitiation fraction in bacteria. Comput Appl Biosci 4, 431–433.
Stoichiometry and architecture of active DNA replication machinery
in Escherichia coli. Science 328, 498–501.
Mathews, C. K., Wheeler, L. J., Ungermann, C., Young, J. P. & Ray,
N. B. (1993). Enzyme interactions involving T4 phage-coded
Sánchez-Romero, M. A., Busby, S. J., Dyer, N. P., Ott, S., Millard, A. D.
& Grainger, D. C. (2010a). Dynamic distribution of SeqA protein across
thymidylate synthase and deoxycytidylate hydroxymethylase. Adv
Exp Med Biol 338, 563–570.
Sánchez-Romero, M. A., Molina, F. & Jiménez-Sánchez, A. (2010b).
Molina, F. & Skarstad, K. (2004). Replication fork and SeqA focus
distributions in Escherichia coli suggest a replication hyperstructure
dependent on nucleotide metabolism. Mol Microbiol 52, 1597–1612.
the chromosome of Escherichia coli K-12. MBio 1, e00012-10.
Correlation between ribonucleoside-diphosphate reductase and three
replication proteins in Escherichia coli. BMC Mol Biol 11, 11.
Sunako, Y., Onogi, T. & Hiraga, S. (2001). Sister chromosome
Molina, F., Sánchez-Romero, M. A. & Jiménez-Sánchez, A. (2008).
cohesion of Escherichia coli. Mol Microbiol 42, 1233–1241.
Dynamic organization of replication forks into factories in Escherichia
coli. Process Biochem 43, 1171–1177.
Edited by: W. J. J. Meijer
http://mic.sgmjournals.org
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 13:10:39
2225