In all domains of life, DNA synthesis occurs bidirectionally from replication origins. Despite variable rates of replication fork progression, fork convergence often occurs at specific sites1. Escherichia coli sets a ‘replication fork trap’ that allows the first arriving fork to enter but not to leave the terminus region2,3,4,5. The trap is set by oppositely oriented Tus-bound Ter sites that block forks on approach from only one direction3,4,5,6,7. However, the efficiency of fork blockage by Tus–Ter does not exceed 50% in vivo despite its apparent ability to almost permanently arrest replication forks in vitro8,9. Here we use data from single-molecule DNA replication assays and structural studies to show that both polarity and fork-arrest efficiency are determined by a competition between rates of Tus displacement and rearrangement of Tus–Ter interactions that leads to blockage of slower moving replisomes by two distinct mechanisms. To our knowledge this is the first example where intrinsic differences in rates of individual replisomes have different biological outcomes.
Subscribe to Journal
Get full journal access for 1 year
only $3.90 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Rent or Buy article
Get time limited or full article access on ReadCube.
All prices are NET prices.
Protein Data Bank
Atomic coordinates and structure factors for the reported crystal structures have been deposited at the Protein Data Bank under accession codes 4XR0 (Tus–UGLT fork), 4XR1 (Tus–TGTA fork), 4XR2 (H144A–WT fork) and 4XR3 (Tus-UGLC).
Dalgaard, J. Z. et al. Random and site-specific replication termination. Methods Mol. Biol. 521, 35–53 (2009)
Hill, T. M., Henson, J. M. & Kuempel, P. L. The terminus region of the Escherichia coli chromosome contains two separate loci that exhibit polar inhibition of replication. Proc. Natl Acad. Sci. USA 84, 1754–1758 (1987)
Coskun-Ari, F. F. & Hill, T. M. Sequence-specific interactions in the Tus-Ter complex and the effect of base pair substitutions on arrest of DNA replication in Escherichia coli. J. Biol. Chem. 272, 26448–26456 (1997)
Neylon, C., Kralicek, A. V., Hill, T. M. & Dixon, N. E. Replication termination in Escherichia coli: structure and antihelicase activity of the Tus-Ter complex. Microbiol. Mol. Biol. Rev. 69, 501–526 (2005)
Duggin, I. G., Wake, R. G., Bell, S. D. & Hill, T. M. The replication fork trap and termination of chromosome replication. Mol. Microbiol. 70, 1323–1333 (2008)
Mulcair, M. D. et al. A molecular mousetrap determines polarity of termination of DNA replication in E. coli. Cell 125, 1309–1319 (2006)
Kaplan, D. L. & Bastia, D. Mechanisms of polar arrest of a replication fork. Mol. Microbiol. 72, 279–285 (2009)
Valjavec-Gratian, M., Henderson, T. A. & Hill, T. M. Tus-mediated arrest of DNA replication in Escherichia coli is modulated by DNA supercoiling. Mol. Microbiol. 58, 758–773 (2005)
Duggin, I. G. & Bell, S. D. Termination structures in the Escherichia coli chromosome replication fork trap. J. Mol. Biol. 387, 532–539 (2009)
Coskun-Ari, F. F., Skokotas, A., Moe, G. R. & Hill, T. M. Biophysical characteristics of Tus, the replication arrest protein of Escherichia coli. J. Biol. Chem. 269, 4027–4034 (1994)
Kamada, K., Horiuchi, T., Ohsumi, K., Shimamoto, N. & Morikawa, K. Structure of a replication-terminator protein complexed with DNA. Nature 383, 598–603 (1996)
Lee, J. B. et al. DNA primase acts as a molecular brake in DNA replication. Nature 439, 621–624 (2006)
Tanner, N. A. et al. Single-molecule studies of fork dynamics in Escherichia coli DNA replication. Nat. Struct. Mol. Biol. 15, 170–176 (2008)
Jergic, S. et al. A direct proofreader-clamp interaction stabilizes the Pol III replicase in the polymerization mode. EMBO J. 32, 1322–1333 (2013)
Tanner, N. A. et al. Real-time single-molecule observation of rolling-circle DNA replication. Nucleic Acids Res. 37, e27 (2009)
Hamdan, S. M., Loparo, J. J., Takahashi, M., Richardson, C. C. & van Oijen, A. M. Dynamics of DNA replication loops reveal temporal control of lagging-strand synthesis. Nature 457, 336–339 (2009)
Yao, N. Y., Georgescu, R. E., Finkelstein, J. & O’Donnell, M. E. Single-molecule analysis reveals that the lagging strand increases replisome processivity but slows replication fork progression. Proc. Natl Acad. Sci. USA 106, 13236–13241 (2009)
Pham, T. M. et al. A single-molecule approach to DNA replication in Escherichia coli cells demonstrated that DNA polymerase III is a major determinant of fork speed. Mol. Microbiol. 90, 584–596 (2013)
Moreau, M. J. & Schaeffer, P. M. Differential Tus-Ter binding and lock formation: implications for DNA replication termination in Escherichia coli. Mol. Biosyst. 8, 2783–2791 (2012)
Neylon, C. et al. Interaction of the Escherichia coli replication terminator protein (Tus) with DNA: a model derived from DNA-binding studies of mutant proteins by surface plasmon resonance. Biochemistry 39, 11989–11999 (2000)
Bastia, D. et al. Replication termination mechanism as revealed by Tus-mediated polar arrest of a sliding helicase. Proc. Natl Acad. Sci. USA 105, 12831–12836 (2008)
Ribeck, N., Kaplan, D. L., Bruck, I. & Saleh, O. A. DnaB helicase activity is modulated by DNA geometry and force. Biophys. J. 99, 2170–2179 (2010)
Kim, S., Dallmann, H. G., McHenry, C. S. & Marians, K. J. Coupling of a replicative polymerase and helicase: a τ-DnaB interaction mediates rapid replication fork movement. Cell 84, 643–650 (1996)
Finkelstein, I. J. & Greene, E. C. Molecular traffic jams on DNA. Annu. Rev. Biophys. 42, 241–263 (2013)
Oakley, A. J. et al. Flexibility revealed by the 1.85 Å crystal structure of the β sliding-clamp subunit of Escherichia coli DNA polymerase III. Acta Crystallogr. D 59, 1192–1199 (2003)
Marians, K. J. ϕX174-type primosomal proteins: purification and assay. Methods Enzymol. 262, 507–521 (1995)
McPhillips, T. M. et al. Blu-Ice and the Distributed Control System: software for data acquisition and instrument control at macromolecular crystallography beamlines. J. Synchrotron Radiat. 9, 401–406 (2002)
Otwinowski, Z. & Minor, W. Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol. 276, 307–326 (1997)
Vagin, A. & Teplyakov, A. MOLREP: an automated program for molecular replacement. J. Appl. Crystallogr. 30, 1022–1025 (1997)
Murshudov, G. N., Vagin, A. A. & Dodson, E. J. Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr. D 53, 240–255 (1997)
Winn, M. D., Isupov, M. N. & Murshudov, G. N. Use of TLS parameters to model anisotropic displacements in macromolecular refinement. Acta Crystallogr. D 57, 122–133 (2001)
Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development of Coot. Acta Crystallogr. D 66, 486–501 (2010)
The PyMOL Molecular Graphics System, Version 126.96.36.199, Schrödinger, LLC
Pandey, M. et al. Two mechanisms coordinate replication termination by the Escherichia coli Tus-Ter complex. Nucleic Acids Res. 43, 5924–5935 (2015)
van Oijen, A. M. et al. Single-molecule kinetics of lambda exonuclease reveal base dependence and dynamic disorder. Science 301, 1235–1238 (2003)
Bustamante, C., Marko, J. F., Siggia, E. D. & Smith, S. Entropic elasticity of lambda-phage DNA. Science 265, 1599–1600 (1994)
We thank A. van Oijen for critical comments and the groups of N. Dekker and S. Patel for helpful discussions. This research was supported by the King Abdullah University of Science and Technology through core funding (to S.M.H.) and a Faculty Initiated Collaborative Award (to S.M.H. and N.E.D.), and by the Australian Research Council (DP0877658 to N.E.D. and A.J.O.; DP0984797 to N.E.D.), including an Australian Professorial Fellowship to N.E.D. and a Future Fellowship (FT0990287) to A.J.O. X-ray crystallographic data were collected at the Australian Synchrotron, Victoria, Australia.
The authors declare no competing financial interests.
Extended data figures and tables
a, A schematic representation of the 13.7 kb DNA substrate construct. The substrate contains a biotinylated fork at one end to attach it to the streptavidin-coated glass coverslip and a digoxigenin moiety at the other end to attach it to a 2.8 μm diameter anti-digoxigenin-coated paramagnetic bead. A single insert of TerB site is located at 3.6 kb from the biotinylated fork. b, Oligonucleotides used to assemble wild-type and variants of TerB substrates for their ligation to the 3.6 kb EcoRI and 10.1 kb ApaI λ DNA fragments34. Native TerB residues are highlighted in yellow except C6 that is in red. Non-native (modified) residues in TerB are highlighted in grey. Native TerH residues are highlighted in orange. Leading and lagging DNA strands as well as permissive (P) and non-permissive (NP) faces of Ter when bound to Tus are denoted. Directionality of translocation of DnaB that encircles the lagging strand as it unwinds dsDNA during leading strand DNA synthesis by Pol III holoenzyme is denoted by arrows.
Extended Data Figure 2 Examples of trajectories for leading-strand synthesis upon encountering Tus bound to non-permissive Ter sites.
The location of the TerB site at 3.6 ± 0.1 kb is indicated by the dashed lines. The rates of leading strand synthesis were calculated by fitting the slopes of the trajectories by linear regression using a least-squares approach. The replisomes displayed heterogeneity in rates of DNA synthesis. a, Trajectories where forks stopped at the NPTerB site. The average stoppage time captured within our acquisition time was 9.7 ± 1 min (uncertainty is the standard error) as illustrated for the top trajectory. b, Trajectories where forks displaced Tus and bypassed the NPTerB site without displaying any transient stoppage. c, Fate of the replication fork upon encountering CG(6)-NPTerB. Examples of trajectories for leading-strand synthesis upon encountering Tus bound to CG(6)-NPTerB showing transient stoppage at CG(6)-NPTerB, followed by resumption of DNA synthesis; 56% of the restarted events displayed DNA synthesis with disrupted behaviour (top row) while 44% showed normal behaviour (bottom row). We attributed the disrupted restart of DNA synthesis in some of the trajectories to the replisome losing some components other than DnaB during stalling.
Extended Data Figure 3 Effect of TerB site alone and nonspecifically DNA-bound Tus on DNA synthesis.
a, Probability of termination of DNA synthesis at 0.2 kb intervals (spatial resolution of the assay) along the 13.7 kb NPTerB in the absence of Tus, showing stops at TerB (3.5–3.7 kb, denoted by black arrow) occur randomly with a 3% probability when all events were considered, in contrast to 5% when only events that reached TerB (≥3.5 kb) were taken into account. b, Processivity of DNA synthesis on the NPTerB substrate in the absence of Tus. The processivity distribution is fit with an exponential decay (N = 88) and uncertainty corresponds to the standard error, illustrating the random stoppage behaviour of the replisome during synthesis. c, Rate of leading strand synthesis using the 13.7 kb force-calibrated DNA construct (NPTerB in this case) in the absence (left panel; N = 94) or presence of Tus (right panel; N = 69). The rate distributions were fit with a Gaussian distribution. The fit lines are shown and the uncertainties correspond to the standard error. The rate agrees with our previously reported rate using force-calibrated λ DNA constructs14, demonstrating the accurate force calibration of the 13.7 kb substrate.
Extended Data Figure 4 Linear fitting of the rate of leading-strand synthesis is appropriate for deriving the correlation between rate of DNA synthesis and stalling activity at the NPTerB site.
a, Rate dependence of fork arrest at NPTerB. A scatter plot of forks that stopped (N = 31) or bypassed (N = 32) Tus bound to NPTerB; rates were calculated by fitting the DNA shortening phase of the entire trajectory in cases of events that bypassed and up to the stoppage point in events that stopped/restarted (histograms are shown in Fig. 1g). A significant correlation between fork progression rate and fork bypass at NPTerB is observed using a one sided Pearson’s correlation test at the 0.05 level of significance (the calculated correlation coefficient (r) was 0.62). The Pearson’s correlation coefficient was calculated using the equation . b, Scatter plot (left) and rate distributions (right) of leading-strand synthesis for events that bypassed (grey bars) (N = 32) or stopped/restarted (blue bars) (N = 31) at NPTerB when the rate was estimated from fitting the slope of the three data points before the TerB site (acquisition time is 0.5 s per data point). The rates were fit with a Gaussian distribution and uncertainty corresponds to the standard error. The calculated average rates for events that bypassed or stopped/restarted at NPTerB are similar to those calculated when the rates were fit using the DNA shortening phase of the entire trajectory in cases of events that bypassed and up to the stoppage point in events that stopped/restarted (shown in Fig. 1g in the main text), underscoring the suitability of linear fitting of the rate. Furthermore r2 from linear regression fits was 0.95 ± 0.05. c, The correlation between apparent fluctuation in rate of DNA synthesis within individual DNA molecules and their corresponding Brownian motion (N = 23). d, The individual trajectories displayed apparent fluctuation in rate of DNA synthesis as illustrated in a representative trajectory where we zoomed in at the DNA shortening phase and fit the rate linearly to intervals of three consecutive data points. The percentage of apparent fluctuation in rate of DNA synthesis within individual DNA molecules was calculated by dividing the standard deviation of the average of interval rates over the average rate. The standard deviation of the average of Brownian motion of each individual DNA molecule was calculated from the fluctuation of the DNA before and after being replicated. The percentage of apparent fluctuation in rate of individual DNA molecules displayed a strong positive correlation with their corresponding Brownian motion when analysed by two-sided Pearson’s correlation test at the 0.05 level of significance (r = 0.81, panel c). e, The correlation between the percentage of apparent fluctuation in rate and the average rate of individual molecules. The percentage of apparent fluctuation in rate of individual molecules was calculated as described in d and for the same 23 replisomes. There was no correlation between the average rate of individual DNA molecules and their corresponding percentage of apparent fluctuation in rate; the Pearson’s correlation coefficient was –0.18. The results from c–e demonstrate that one strong factor behind the apparent fluctuation in rates within our individual 13.7 kb molecules under our spatial and temporal resolution is the Brownian motion of the DNA and that this apparent fluctuation in rate does not bias the estimates of speed of the replisomes.
The sequences of oligonucleotides used for each complex are shown at the bottom of each panel; nucleotides for which electron density could not be interpreted are highlighted. a–d, Complexes of Tus proteins with forked Ter sites. The C6-binding pocket is shown in the circle, with key residues Ile79, Phe140 and His144 in the binding pocket, and Arg198 shown in stick form. a, The wild-type Tus–Ter lock (PDB code: 2I06), with C6 located in the binding pocket, and the TA(7) base pair melted. Arg198 is positioned to interact with the 5′-phosphate of T7. b, Complex of wild-type Tus with a forked oligonucleotide that has C6 substituted by a mispaired G (UGLT: upper G, lower T; PDB code: 4XR0); G6 does not occupy the pocket nor does it make any new specific interactions with Tus, and Arg198 no longer interacts with the 5′-phosphate of T7. c, Further extension of the mismatched region in b to include A7 (TGTA: mispaired TGTA on the lower strand; PDB code: 4XR1) does not enable G6 to occupy the C6-binding pocket or form any new specific interactions. d, Tus(H144A) in complex with the normal Tus–Ter lock oligonucleotide (PDB code: 4XR2), showing the mispaired C6 does not occupy the cytosine-binding pocket or form any new interactions with Tus. e, f, Potential interactions of Arg198 in crystal structures of Tus complexes with fully base-paired Ter oligonucleotides. Only nucleotides in base pairs 5 and 6 are shown, and they are colour-coded to match the stick representations of them in the figures. Arg198 is shown in yellow stick representation. e, Structure of the wild-type Tus–TerA (GC(6)) complex (PDB code: 2I05). Arg198 is positioned potentially to make H-bonding interactions with the A5, G6 and T5 bases and the deoxyribose ring oxygen of G6, as well as electrostatic interactions with the 5′-phosphate of A5, as suggested previously11 and demonstrated by molecular dynamics simulations (A.J.O., unpublished observations). f, Structure of the complex with a GC(6)-flipped version of the TerA oligonucleotide (UGLC: upper G, lower C; PDB code: 4XR3) showing an alternate major conformation of the Arg198 side-chain that has lost all base-specific interactions; only the interaction with the sugar ring oxygen of the substituted C6 is maintained.
a, Examples of trajectories of replication forks that transiently stopped at Tus bound to the bubble template with C6 switched to G6 (5-mismatch G6-NPTerB). b, The distribution of the pause durations fit with a single exponential decay. The fit line is shown in black and the uncertainty corresponds to the standard error (N = 20).
Extended Data Figure 7 Fate of the replication fork upon encountering NPTerB sites with swapped sequences in the first five base pairs, TA(6)-NPTerB and NPTerH.
Rate dependence of replication fork arrest at Tus bound to: a, TA(5)-NPTerB (N = 25); b, swapped F4n GC(6)-NPTerB (N = 29); c, swapped F5n GC(6)-NPTerB (N = 36). The rate distributions of leading-strand synthesis for events that bypassed (grey bars) or stopped/restarted (blue bars) at these sequences. d, Examples of trajectories of leading-strand synthesis that transiently stopped at Tus bound to TA(6)-NPTerB. 75% of the restarted events displayed DNA synthesis of normal behaviour (left traces) while 25% showed disrupted behaviour (right trace). e, The distribution of the pause durations at TA(6)-NPTerB fit with a single exponential decay (N = 8). f, The rate distribution of events that bypassed (N = 30; grey bars) or stopped/restarted (N = 11; blue bars) at TA(6)-NPTerB. g, Examples of trajectories of leading-strand synthesis that transiently stopped at Tus bound to NPTerH. The average pause duration was 180 ± 26 s (N = 4). The uncertainty is the standard error. h, The rate distribution of leading-strand synthesis for events that bypassed (N = 26; grey bars) or stopped/restarted (N = 11; blue bars) at NPTerH. The histograms in a–c, e, f and h were fit to Gaussian distributions, the fit lines are shown, and the uncertainties correspond to the standard error.
Extended Data Figure 8 SPR assessment of Tus–TerB interactions: whereas Tus and Tus(R198A) are capable of forming a lock, Tus(H144A) is not.
ProteOn sensorgrams show association and dissociation phases of Tus−TerB interactions at ranges of Tus concentrations (as specified in g) of serially-diluted samples of Tus proteins. Curves, shown in colours, were fit simultaneously (black curves) to various binding models (see Methods). a, Wild-type Tus and dsTerB. Considering that the ka > 1 × 106 M–1s–1 suggests significant mass transport limitations, the LMT model was used to fit the data with Rmax constrained to 700 RU. The derived kinetic parameters were used to simulate sensorgrams devoid of mass transfer limitation using the L model (inset). b, Wild-type Tus−forked TerB interaction; Rmax was constrained to 775 RU. The fit kd is in good agreement with the value of (5.20 ± 0.00) × 10−5 s–1 obtained from an independent experiment where dissociation was monitored over 50,000 s (not shown). c, Tus(R198A)−dsTerB interaction. Binding kinetics parameters were obtained using the HLPR model. The sum of fit Rmax1 (543 ± 9) and Rmax2 (54 ± 5 RU) values were in reasonable agreement with the expected value of ∼700 RU. Only the relevant ka and kd values of the predominant (based on Rmax1) interaction are presented in g. For assessment of the fitting procedure, responses at equilibrium were fit using the L model (inset). The derived KD was within the factor of two of the calculated KD obtained from kinetic parameters (kd/ka). The Rmax value of 816 ± 32 RU was slightly higher than theoretical (700 RUs), probably owing to some non-specific binding in the high range of Tus concentration. d, Tus(R198A)−forked TerB interaction. The L model was used to fit the data with Rmax constrained to 775 RU. The fit kd was within a factor of two of the value, (5.70 ± 0.00) × 10−5 s–1, derived from an independent experiment where dissociation was monitored over 50,000 s (not shown). e, Tus(H144A)−dsTerB interaction. Binding kinetic parameters were obtained using the HLPR model. The sum of fit Rmax1 (537 ± 1) and Rmax2 (31 ± 0 RU) values were in reasonable agreement with the expected value of ∼700 RU. Only the relevant ka and kd values of the predominant interaction (Rmax1) are presented in g. For assessment of the fitting procedure, responses at equilibrium were fit using the L model (inset). The derived KD was within a factor of 1.5 of KD obtained from the kinetic parameters. In addition, the fit Rmax value of 621 ± 10 RU compares reasonably to the expected value of ∼700 RU. f, Tus(H144A)−forked TerB interaction. Binding kinetics parameters were obtained using the HLPR model. The sum of fit Rmax1 (879 ± 4) and Rmax2 (65 ± 1) values were somewhat high compared to the expected value of ∼775 RU. Only the relevant ka and kd values of the predominant reaction are presented in g. Responses at equilibrium were fit using the L model (inset). Derived KD was within the factor of 2 of the calculated KD obtained from (kd/ka). In addition, fit Rmax value of 1,040 ± 50 RU was slightly higher than theoretical. g, Summary of binding parameters for Tus–Ter interactions. All uncertainties are standard errors in parameters from fitting of complete data sets to appropriate binding models as described in the Methods. Data are representative of those from two technical replicates using different instruments (BiaCore T200 and ProteOn XPR-36).
Rate dependence of fork arrest. The rate distribution of leading-strand synthesis for events that bypassed (N = 18; grey bars) or stopped/restarted (N = 15; blue bars) at NPTerB fit with Gaussian distributions. The fit lines are shown and the uncertainties correspond to the standard error.
About this article
Cite this article
Elshenawy, M., Jergic, S., Xu, Z. et al. Replisome speed determines the efficiency of the Tus−Ter replication termination barrier. Nature 525, 394–398 (2015). https://doi.org/10.1038/nature14866
Probabilistic model based on circular statistics for quantifying coverage depth dynamics originating from DNA replication
Scientific Reports (2020)
Chemical Reviews (2020)
BMB Reports (2019)
FEBS Letters (2019)