Real-time imaging of rotation during synthesis by the replisome
Thomas M. Retzer, Lional T. Rajappa, Masateru Takahashi, Samir M. Hamdan, Karl E. Duderstadt
TL;DR
This study uses a new imaging method to observe how the replisome rotates during DNA replication, revealing how it manages DNA overwinding.
Contribution
A transverse DNA flow-stretching approach was developed to visualize replisome rotation in real time.
Findings
Replisome rotation occurs in bursts to support continuous DNA synthesis.
Excessive rotation increases pausing and the number of polymerases but does not slow replisome speed.
The findings reveal how replisomes overcome DNA topology challenges during replication.
Abstract
During chromosome replication, unwinding by the helicase and synthesis by the polymerases can lead to overwinding and supercoiling of DNA. The mechanical consequences of these events and resulting local dynamics at the replication fork are not well understood. To address these issues, we developed a transverse DNA flow-stretching approach to spatially resolve the parental, leading, and lagging strands in real time. Using bacteriophage T7 as a model system, this approach revealed bursts of high-speed replisome rotation that support continuous DNA synthesis. Unexpectedly, excessive rotation does not reduce replisome speed, but increases pausing, reduces processivity, and increases the number of polymerases. Together, our observations reveal intrinsic pathways to overcome challenges posed by unfavorable DNA topologies during DNA replication. Transverse flow imaging reveals replisome…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5- —http://dx.doi.org/10.13039/100010663H2020 European Research Council
- —http://dx.doi.org/10.13039/100010663H2020 European Research Council
- —http://dx.doi.org/10.13039/501100001659Deutsche Forschungsgemeinschaft
- —Max Planck Society
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsMicrotubule and mitosis dynamics · DNA Repair Mechanisms · Bacterial Genetics and Biotechnology
INTRODUCTION
Throughout the domains of life, DNA is copied using a similar mechanism by replisomes that share a conserved core architecture (1). Unwinding of parental DNA by a helicase is coupled to the synthesis of two daughter strands by DNA polymerases. Because of the antiparallel arrangement of DNA, the leading strand is synthesized continuously whereas the lagging strand is synthesized discontinuously as Okazaki fragments supported by repeated primer synthesis by primases. Coordination of these and many other moving components involved in chromosome replication is frequently challenged by obstacles such as DNA damage (2), chromosome organizational elements (3), and other structurally diverse barriers (4). Among these are chromosomal regions with unexpected DNA topologies (5, 6) that challenge unwinding by the helicase. The mechanical consequences and resulting dynamics of the replisome during collisions with topological barriers have been experimentally challenging to address. As a consequence, it has remained unknown whether replisomes have intrinsic pathways to cope with topological challenges to avoid genome instability.
The double-helical structure of DNA is advantageous for the storage and maintenance of genetic information, but it poses major challenges when the information-rich DNA bases must be accessed during genome duplication. Unwinding by the helicase leads to overwinding in the parental DNA (Fig. 1A). Likewise, superhelical tension forms on the daughter strands as a consequence of DNA synthesis by the polymerases (7) (Fig. 1B). To avoid chromosome damage and genome instability (8–10), these unfavorable DNA topologies must be resolved by topoisomerases, which are a class of enzymes that help ensure chromosomes are maintained during genome compaction, chromosome segregation, DNA replication, and transcription (8, 11–13). Nevertheless, there are many regions on chromosomes where topoisomerases are unable to prevent the accumulation of overwinding. In particular, topoisomerases struggle to keep pace in highly transcribed regions, at topological boundaries, near chromosome ends, and with rapidly moving replication forks (14–20). Replisome rotation has been proposed as an alternative mechanism to cope with overwinding resulting from helicase activity (21). The observation of interwound daughter strands and the formation of precatenanes provides support for this proposal (22–26), but rotational motion of the replisome during synthesis has not been directly observed.
Topological challenges during DNA replication.(A) DNA replication introduces overwinding and positive supercoils. This can lead to fork rotation and the formation of precatenanes on the daughter strands. (B) Structural model of the T7 replisome (28). Parent DNA is unwound by the helicase generating templates (lead and lag) used for daughter-strand synthesis. The DNA polymerases travel on a helical path on leading to rotation of the daughter strands or the formation of superhelical tension.
To address these issues, we used single-molecule fluorescence imaging to directly visualize replisome dynamics during topological challenges. We reconstituted the replisome from bacteriophage T7, which performs DNA replication with a minimal set of components (27, 28). We quantified replication fork progression on DNA molecules with topological barriers introduced in the parental strand. Unexpectedly, we find that the presence of barriers resulted in only a very modest reduction in replisome speed; however, this reduction was accompanied by an increase in pausing frequency and reduced processivity. To characterize the fundamental mechanics of these encounters, we developed a transverse DNA flow-stretching approach that allows for real-time spatial resolution of the parental, leading, and lagging strands. Notably, this approach revealed that high-speed replisome rotation supports bursts of continuous DNA synthesis during encounters with topological barriers. While allowing the replisome to overcome topological barriers, rotation does result in more polymerases residing at the replication fork, consistent with frequent disruptions to helicase-polymerase coordination. Together, our observations reveal that replisome rotation is an intrinsic pathway that supports continued DNA synthesis during topological challenges, but at the cost of detrimental changes in replisome coordination.
RESULTS
Visualization of DNA replication during encounters with topological barriers
We developed a total internal reflection fluorescence (TIRF)–based single-molecule approach to visualize the dynamics of DNA replication during encounters with topological barriers. DNA replication was reconstituted using purified components in a stepwise manner. First, 27-kb-long DNA molecules were surface immobilized on functionalized coverslips via biotin-streptavidin-biotin interactions (fig. S1). A replication fork was introduced at one end of the molecules with a single biotin on the leading strand for surface attachment and one or multiple biotins for surface attachment of the opposite end. We created two types of molecules: unconstrained (using a single biotin and allowing the parental strand to rotate freely) and constrained (using multiple biotins and creating a topological barrier that prevented free rotation of the parental strand) (Fig. 2A). The purified T7 replication proteins gp5-trx (DNA polymerase), gp4 (helicase primase), and gp2.5 (single-stranded DNA binding) were introduced to start the reaction (fig. S2A). Last, flow through the flow cell was stopped and replication fork progression was imaged by fluorescently staining the DNA and monitoring accumulation of the lagging-strand product, visible as a blob moving unidirectionally along the immobilized DNA molecules (Fig. 2B and fig. S2B).
Visualizing replisome progression under topological strain.(A) Schematic of replication assay with unconstrained and constrained molecules. A preformed fork is used to initiate DNA replication using T7 replisome components. Leading and parental strands are tethered either by a single or multiple biotin interactions. Multiple biotin interactions ahead of the replisome create a topological challenge. (B) Representative kymograph showing DNA replication for unconstrained and constrained molecules. (C) Replication burst rate distribution for unconstrained and constrained molecules. (D) Replication processivity distribution for unconstrained and constrained molecules. Inset displays the fraction of molecules that replicated to the end. (E) Replication processivity estimation using a maximum likelihood estimation. (F) Pause probabilities for unconstrained and constrained molecules. (G) Replication blob size for unconstrained and constrained molecules.
The dynamics of replication on the unconstrained DNA were consistent with previous single-molecule observations. Replication kinetics were quantified by performing subpixel localization and tracking of the lagging-strand replication product (fig. S2C). This analysis resulted in a burst replication rate of 99 ± 3 bp/s (mean ± SEM, n = 186 molecules) (Fig. 2C). A total of 75% of the unconstrained DNA molecules were completely copied resulting in 27-kb products (Fig. 2D). To accurately estimate the processivity, we used maximum likelihood estimation (MLE) with a model that combined both replication events observed to stop before the end and those prematurely terminated when they reached the end. Using this approach, the estimated mean processivity was 117 kb (95% confidence interval from 102 to 137, n = 186 molecules) (Fig. 2E), higher than previous reports (29–31). Pausing was observed in 13% of molecules, with 12% of replisomes pausing once and 1% pausing twice (Fig. 2F and fig. S3) at our experimental time resolution.
Several measures were taken to avoid photo damage of DNA that causes nicks, disrupts DNA replication, and prohibits observations of DNA topology dynamics. Optimal conditions were established by quantifying the frequency of nicking using supercoiled DNA molecules attached at each end to the coverslip surface with multiple biotins as previously described (fig. S4) (32). This approach allowed for the establishment of a low light intensity imaging condition in which 95% of the DNA molecules remained intact during the observation time.
Replisomes tolerate topological strain
Having established robust imaging conditions, next we performed replication with the constrained DNA. The presence of multiple biotin attachment sites at the end of the parental strand prohibiting free rotation would be expected to present a major challenge to helicase unwinding and DNA synthesis, which involves rotation along the helical axis of DNA. However, tracking the motion of the lagging-strand replication product (Fig. 2B and fig. S2) unexpectedly revealed a remarkable tolerance by the replisome when faced with the topological constraint. Unexpectedly, the mean burst rate of 96 ± 5 bp/s (mean ± SEM, n = 79 molecules) was within the margin of uncertainty when compared to the unconstrained DNA (Fig. 2C). Nevertheless, the number of molecules that were replicated to completion was reduced to 57%, with a mean processivity of 52 kb (95% confidence interval from 43 to 67, n = 79 molecules) (Fig. 2, D and E). The frequency of molecules with pausing events increased to 22%, and there was an increase in the fraction of molecules exhibiting a second pause to 4% (Fig. 2F and fig. S3). The observation of only modest reductions in processivity suggested that the replisome must be exploiting an alternative pathway to relieve torsional strain. To test whether processivity could be restored by adding topoisomerase, we added a low concentration of DNA gyrase (5 nM) to the reaction. This resulted in a replication burst rate of 103 ± 5 bp/s (mean ± SEM, n = 97 molecules), which was within the margin of uncertainty compared with unconstrained and constrained DNA without gyrase (fig. S5A). The mean processivity increased to 84 (95% confidence interval from 70 to 105, n = 97 molecules), indicating a positive effect on replication at modest concentrations of DNA gyrase (fig. S5, B and C). The frequency of pausing decreased to 10%, with 9% showing a single pause and 1% showing two pauses (fig. S5D).
Visual inspection of kymographs of individual constrained molecules revealed elongation and shape variability in the lagging-strand products (Fig. 2B and fig. S2B). These shape changes were not observed in the unconstrained DNA where the lagging-strand product appears as a single uniform bundle moving in a unidirectional fashion. To quantify the behavior, the kymographs were segmented to determine estimates for the size of the lagging-strand product. Elongation was more pronounced once the products had increased in length toward the end of replication where a mean size difference from 1.1 μm (unconstrained, n = 114 molecules) to 1.5 μm (constrained, n = 52 molecules) was observed. Furthermore, size fluctuations doubled from a standard deviation of 0.2 μm (unconstrained) to 0.4 μm (constrained) (Fig. 2G). This pronounced shape variability hinted at a change in the mechanics of the replisome at the replication fork that could not be fully resolved.
Transverse flow platform to study DNA topology dynamics
To further resolve the spatial dynamics of the lagging-strand products during DNA synthesis, we devised an alternative flow cell geometry in which replication could be monitored in the presence of a transverse flow. This was accomplished using an x-shaped flow cell with two perpendicular flow lanes that crossed each other at the location of imaging (Fig. 3A). DNA replication was reconstituted in a stepwise manner but with different flow lanes used for each step. DNA molecules were surface immobilized on the functionalized coverslips using one flow lane. Next, the other flow lane was used to introduce the replication components using a flow transverse to the DNA molecules. In contrast to the assay presented in Fig. 2, where flow was stopped during imaging, and the lagging-strand product appeared as a compact blob, here transverse flow was continuously applied during imaging to extend the lagging-strand product during synthesis. With the new geometry, the lagging-strand product appeared as a straight line that grew in length over time from the location of the moving replisome at the replication fork (Fig. 3B, fig. S6, and movie S1). On the basis of the direction of movement, molecules were segmented and the length of each arm of the replication fork was precisely measured (Fig. 3C). This approach spatially resolves the parental, leading, and lagging strands in real time during DNA replication.
Visualization of DNA replication using transverse flow.(A) Schematic of transverse flow imaging of DNA replication. After surface immobilization of DNAs using the attachment flow, a perpendicular imaging flow is used to extend the lagging-strand products during DNA replication. (B) Representative kymograph of replication on an unconstrained DNA molecule that appears as an arch with the lagging-strand product extending from the arch over time. (C) Length of parental, leading, and lagging strands as a function of time with a cartoon displaying strand organization. (D) Rate distribution for unconstrained molecules in the linear and transverse configurations. (E) Processivity distribution for unconstrained molecules in the linear and transverse configurations. Inset displays the fraction of molecules that replicated to the end. (F) Distribution of leading- and lagging-strand synthesis rates based on the individual product lengths.
The application of transverse flow did not influence the replication kinetics of the unconstrained DNA. The mean burst rate was 97 ± 7 bp/s (mean ± SEM, n = 24 molecules) and the processivity was 84 kb (95% confidence interval from 60 to 140, n = 24 molecules) (Fig. 3, D and E, and fig. S7) with a confidence interval overlapping the processivity observed for the linear replication assay. The mean burst rate for the constrained DNA was 98 ± 9 bp/s (mean ± SEM, n = 28 molecules) (Fig. 4D), consistent with our findings in the absence of transverse flow. However, we observed a reduction in processivity for the constrained DNA to 29 kb (95% confidence interval from 21 to 46, n = 28 molecules) and the percentage of molecules fully replicated was reduced to 32% (Fig. 4E).
Transverse flow imaging reveals replication fork rotation.(A) Representative kymograph of replication on a constrained molecule that appears as an arch. The lagging-strand product appears as a blob wrapped along the arch. (B) Length of parental, leading, lagging strands over time for the molecule displayed in (A). (C) Cartoon depicting mechanism of fork rotation and arch organization. (D) Rate distribution for unconstrained and constrained molecules using transverse flow imaging. (E) Processivity distribution for unconstrained and constrained molecules imaged using transverse flow. Inset bar graph shows fraction of completely replicated molecules. (F) Lagging-strand length over time for unconstrained and constrained molecules. (G) Variance in lagging-strand length for unconstrained and constrained molecules.
Transverse flow imaging spatially separates each of the strands emerging from the replication fork, revealing the leading- and lagging-strand synthesis dynamics. Measurements of lagging-strand growth revealed a mean burst rate of around 80 ± 7 bp/s (mean ± SEM), which was lower than the leading-strand synthesis rate (Fig. 3F). We attribute this difference to the presence of compacted gp2.5-coated regions undergoing replication that are shorter in length than double-stranded DNA at the applied force (27). Using length measurements of the individual strands of the replication fork, we estimated the applied forces with the worm-like chain (WLC) model. The force on the leading and parental strands of the replication fork was 1.27 ± 0.59 pN (median ± median absolute deviation, n = 24 molecules) (fig. S8A). Force on the lagging strand is expected to increase over time as the lagging-strand product grows and the volume of DNA subjected to viscous drag from flow increases. We estimate a maximal force on the lagging strand of approximately 0.22 ± 0.05 pN (median ± median absolute deviation, n = 24 molecules) (fig. S8B) once a full product has been synthesized. Previous studies have shown that replication is not disrupted by forces in the low-piconewton range used to make observations (27).
Lagging-strand dynamics reveal fork rotation
Transverse flow imaging of the constrained DNA revealed substantial changes in replisome mechanics including bursts of replication fork rotation. For the unconstrained DNA, the lagging-strand product grew as a straight line over time. In contrast, pronounced length fluctuations were observed for the constrained DNA with the lagging-strand product frequently wrapping into a compact blob similar in structure to that observed in the linear replication experiments (Fig. 4A, fig. S9, and movie S2). Segmentation of the individual strands and tracking of the replisome position confirmed that leading-strand synthesis continued during wrapping (Fig. 4B). These observations are consistent with rotation of the replication fork during synthesis and wrapping of the lagging strand around the leading strand (Fig. 4C). In cases where synthesis pauses and then resumes, the extended lagging-strand product adopts a blob consistent with rewrapping as fork rotation continues during synthesis (fig. S10). The burst replication rate of 98 ± 9 bp/s (mean ± SEM, n = 28 molecules) for constrained molecules was nearly identical to 97 ± 7 bp/s (mean ± SEM, n = 24 molecules) for unconstrained molecules (Fig. 4D). In contrast, we observed a 52% reduction in the number of molecules that were completely copied (Fig. 4E).
To quantify variation in the extension of the lagging-strand product over time, length measurements were transformed into a stationary function by taking the difference of consecutive values (Fig. 4F and fig. S11). The mean difference between consecutive length measurements was more than three times greater for the constrained DNA, increasing from 1.8 ± 0.6 kb^2^ (mean ± SD, n = 24 molecules) for the unconstrained DNA to 6.1 ± 5.9 kb^2^ for the constrained DNA (mean ± SD, n = 28 molecules) (Fig. 4G). The larger standard deviation for constrained DNA is consistent with the significant length changes that are visible in the kymographs from individual molecules.
Topological barriers encountered during replication can lead to overwinding of the parental strand to accommodate the helical path of the replisome. At the low force applied by transverse flow, overwinding of the parental strand would be expected to lead to formation of positive supercoils (33). However, we did not observe significant length variation in the parental strand or compaction of the arch structures formed by transverse flow. This indicated no positive supercoils formed on the parental strand during replication and instead only small transient changes in twist may have occurred. Together, these observations suggest that fork rotation is the dominant pathway used by the replisome to overcome acute topological barriers on the parental strand under our experimental conditions.
The transverse flow applied to visualize lagging-strand synthesis and rotational dynamics favors a specific replication fork geometry. Under these conditions, we observe a reduction in processivity to 29 kb (95% confidence interval from 21 to 46, n = 28 molecules) for constrained molecules in comparison to 84 kb (95% confidence interval from 60 to 140, n = 24 molecules) for unconstrained molecules (fig. S7 and table S1). This represents a 65% reduction in processivity, which is greater than the 55% reduction observed for constrained versus unconstrained in the linear replication experiments. This difference may result from the specific replication fork geometry during transverse flow requiring additional torque to wrap the lagging strand. Whether linear or transverse flow are more reflective of the replication fork geometry adopted in cells remains unclear. However, we imagine that the replisome encounters substantial forces when traveling through topologically challenging regions in the chromosome that could be similar if not more substantial than the application of transverse flow. Notably, the motion of the lagging strand is not restricted in our experiments, whereas in cells, the presence of the daughter chromosome would be expected to constrain the motion.
Fork rotation increases the number of polymerases
To determine whether fork rotation leads to changes in replisome composition, we performed replication with fluorescently labeled DNA polymerases (LD655-gp5-trx). Under conditions without flow, we observe polymerase signal moving together with the lagging-strand product (Fig. 5A and fig. S12). To determine polymerase number, the mean fluorescence of single labeled polymerases was estimated before photobleaching (fig. S13). The half-life of labeled polymerases was calculated to be 10-fold longer than the observation time of replication, demonstrating that photobleaching does not affect estimates of polymerase numbers (fig. S14). For the unconstrained DNA under no-flow conditions, we estimate 3.75 ± 0.08 (mean ± SEM, n = 75 molecules) polymerases at the replisome consistent with previous estimates (34). Notably, for the constrained DNA, we observe a significant increase to 11.40 ± 0.22 (mean ± SEM, n = 56 molecules) polymerases at the replisome in the presence of the topological barrier (table S1).
Fork rotation increases the number of polymerases.(A) Representative kymograph of replication of an unconstrained and constrained molecule in the linear configuration. (B) Representative kymograph of replication on an unconstrained and constrained molecule in linear configuration under preassembly conditions with additional polymerases. The top kymograph displays the lagging-strand product and the bottom displays polymerases in (A) and (B). (C) Comparison of polymerase number distribution at the replication fork for unconstrained and constrained molecules for excess protein, preassembly, and preassembly with additional gp5. (D) Violin plot of polymerase numbers for unconstrained and constrained molecules under preassembly conditions, including additional gp5, depending on position split in the first 75% and last 25% of replication substrate. (E) Sliding window analysis of number of polymerases for unconstrained (top) and constrained (bottom), depending on the replisome position on DNA. Gray data points represent the single measurements. Solid points represent the window means, with SEM indicated by the error bars.
The significant increase in polymerases beyond the six binding sites on gp4 (34) prompted us to repeat our experiments using preassembled replisomes (fig. S15). With polymerases omitted after preassembly (fig. S16A), we observe 2.07 ± 0.04 (mean ± SEM) polymerases during replication of unconstrained molecules and 2.92 ± 0.08 (mean ± SEM) polymerases during replication of constrained molecules. Notably, omission of polymerases leads to a reduction in processivity for unconstrained molecules to 44 kb (95% confidence interval from 39 to 52, n = 178 molecules) and a further reduction for constrained molecules to 36 kb (95% confidence interval from 29 to 47, n = 76 molecules) (fig. S16).
Next, we performed experiments with preassembled replisomes with gp5 and gp2.5 included in solution during replication (fig. S17). Under these conditions, we observed 3.61 ± 0.05 (mean ± SEM) polymerases for the unconstrained molecules and 4.91 ± 0.09 (mean ± SEM) polymerases for the constrained molecules. The processivity of the unconstrained molecules increased modestly to 52 kb (95% confidence interval from 46 to 61, n = 194 molecules). In contrast, the processivity for constrained molecules remained unchanged when compared with the condition where polymerases were omitted.
The increase in polymerases we observe for the constrained molecules across all conditions (Fig. 5C) suggests that excessive fork rotation disrupts polymerase-helicase coupling and leads to the recruitment of additional polymerases. Over time, this results in the accumulation of polymerases at the replication fork (Fig. 5, D and E). These observations are consistent with a recent report that excess polymerases in solution help support continued replication under torsion (35). The observed number of polymerases under preassembly conditions are below the six binding sites on gp4, suggesting that the additional polymerases are associated with the replisome. In contrast, the number of polymerases during replication of constrained molecules in the absence of preassembly was beyond the six binding sites available on gp4. This could result from the loading of a second gp4 behind the active replisome that could recruit additional polymerases. Alternatively, lagging-strand polymerases released from the replisome may remain associated for long time periods after completion of synthesis as has been previously reported (31). Because the lagging-strand product resides as a blob at the same location as the replisome, the released polymerases would be included in our estimates.
DISCUSSION
Our observations demonstrate that a rotational mode of replisome operation provides an intrinsic pathway to overcome topological challenges but results in negative consequences for replisome coordination. It remains to be resolved whether replisome rotation serves only as a fail-safe pathway during periods of acute stress or is a mechanical feature of normal replication. Topological barriers frequently form on regions of chromosomes where accumulation of overwinding outpaces topoisomerase activity. This occurs in highly transcribed regions that are rapidly overwound ahead of advancing RNA polymerases (20, 36, 37). During head-on encounters with the replisome, the rate of overwinding is further increased and binding sites available for topoisomerases are reduced (20, 38). Similar challenges occur at topological boundaries and near chromosome ends (6, 7, 14–16). Many common forms of stress can further complicate these issues (8). Therefore, topological barriers are clearly a major frequent challenge to DNA replication. Two recent studies using both optical trapping and angular optical trapping have demonstrated that DNA torsion is a dynamic regulator of DNA replication stalling and reactivation (35, 39). However, live-cell imaging in Escherichia coli has revealed that while topoisomerases accumulate at the replication fork, their number remains insufficient to keep up with the overwinding created by the replisome (14). This latter observation opens up the possibility that replisomes may rotate not only during encounters with acute topological barriers but also during normal operation.
Dynamic exchange of polymerases provides robustness during DNA replication with the frequency adapting in response to environmental challenges (40–47). Changes in polymerase coupling dynamics have been hypothesized to serve as a mechanical response pathway during encounters with torsional strain (7). Normally, lagging-strand synthesis would be expected to continue until encountering the previously synthesized Okazaki fragment or a signal triggers early release (48, 49). In the latter scenario, the lagging-strand polymerase dissociates from its gp4 binding site and continues Okazaki fragment synthesis behind the replisome. This type of premature release pathway was previously proposed as a mechanism to relieve torsional strain during rolling-circle replication (50).
We observe that placement of a topological barrier directly ahead of the replication fork results in polymerase accumulation. This observation is consistent with torsional strain leading to more frequent polymerase dissociation from either the helicase or the template. Previous experimental observations and computational modeling have revealed that dissociation events from either site can facilitate the recruitment of additional polymerases (34, 51, 52). Modeling predicts that polymerases rapidly rebind under preassembly conditions, whereas exchange pathways are activated by polymerases in solution. Our observations are consistent with these predictions. However, further studies will be required to determine the occupancy of specific polymerase binding sites at the replication fork over time.
Eukaryotes have evolved accessory factors, such as the fork protection complex, that provide an early warning system to regulate replisome speed and reduce excessive fork rotation that can lead to uncoupling of daughter-strand synthesis and replication fork collapse (16, 23). Chromatin can also help reduce the topological stress on the replication fork by acting as a shock absorber for moderate increases in overwinding (53). These features could provide eukaryotes with more time for topoisomerases and the replisome to react and ensure fork rotation is limited in duration and rate to reduce the risk of disruption to replisome coordination. Nevertheless, there are many regions on chromosomes where the rate of overwinding can exceed what can be absorbed by chromatin. Stalling and replication fork collapse have been reported under these scenarios (16, 23). This highlights the importance of understanding intrinsic pathways that allow replisomes to cope with topological challenges. Further studies conducted with reconstituted eukaryotic replisomes are needed to determine exactly how nucleosomes modulate fork rotation dynamics.
Transverse flow imaging offers opportunities to discover the dynamics at the replication fork that ensure robust genome duplication. By spatially resolving the parental, leading, and lagging strands in real time during DNA replication, the approach opens up the possibility of time-resolved, strand-specific tracking of not only replication factors but also a wide array of other machineries that must coordinate their activities with the replisome. These events could be further evaluated in the context of diverse strand-specific obstacles frequently encountered on chromosomes. The approach can be easily adapted to studies of eukaryotic replication through small modifications to existing experimental platforms (42). Together, this observation potentially provides a powerful tool to discover the dynamics that support robust chromosome replication throughout the domains of life.
MATERIALS AND METHODS
Gp4 purification
Recombinant gp4 proteins were overexpressed and purified using established protocols with necessary modifications (54–56). E. coli strain HMS 174(DE3), harboring a plasmid encoding the gp4 protein, was grown in lysogeny broth (LB) medium to an optical density at 600 nm (OD_600_) of 1.0. Protein expression was induced by adding isopropyl-β-d-thiogalactopyranoside (IPTG) to a final concentration of 1 mM, and cultures were incubated for an additional 3 hours at 37°C. Cells were harvested by centrifugation and resuspended in lysis buffer (20 mM tris-HCl, pH 7.5, 5 mM EDTA, 0.1 M NaCl, and 1 mM phenylmethylsulfonyl fluoride). Cell lysis was performed using a cell disruptor with a pressure of 25,000 psi (172,368.93 kPa). The lysate was clarified by centrifugation at 22,404g for 30 min, and the supernatant was treated with polyethylene glycol, molecular weight 4000 (PEG-4000) to a final concentration of 10%. The resulting precipitate was collected by centrifugation at 6000g for 20 min, resuspended in binding buffer [20 mM potassium phosphate, pH 6.8, 1 mM EDTA, 1 mM dithiothreitol (DTT), and 10% glycerol], and subjected to phosphocellulose chromatography. Bound proteins were eluted with a gradient of KCl (0.02 to 1 M), and fractions containing gp4 protein were pooled based on SDS–polyacrylamide gel electrophoresis (SDS-PAGE) analysis. To further purify the protein, MgCl_2_ was added to the pooled fractions to a final concentration of 10 mM, and the solution was loaded onto an adenosine triphosphate (ATP)–agarose affinity column. Protein was eluted using a buffer containing 20 mM potassium phosphate, pH 6.8, 20 mM EDTA, 0.5 mM DTT, 10% glycerol, and 0.5 M KCl. The elution fractions from the ATP-agarose column were pooled, and the protein purity was confirmed by SDS-PAGE analysis. The concentration was adjusted to approximately 10 μM to ensure that the final concentration of the hexameric helicase after stock buffer dialysis was maintained above 15 μM, which is optimal for long-term storage. The final solutions were dialyzed against storage buffer (20 mM potassium phosphate, pH 7.5, 0.1 mM DTT, 0.1 mM EDTA, and 50% glycerol) and stored at −20°C until use.
Gp2.5 purification
Recombinant gp2.5 proteins were overexpressed and purified using established protocols with the necessary modifications (57). E. coli BL21(DE3)pLysS cells harboring a plasmid encoding gp2.5 were grown in LB medium to an OD_600_ of 1.0. Cells were harvested by centrifugation and resuspended in lysis buffer (50 mM tris-HCl, pH 7.5, 0.1 mM EDTA, 1 mM DTT, 10% glycerol, and 1 M NaCl) before freezing. Before extraction, frozen cells were thawed on ice in a cold room overnight. Cells were lysed by incubation with lysozyme (1 mg/ml) for 1 hour at 4°C with continuous mixing, followed by cell disruption using a cell disruptor at a pressure of 25,000 psi (172,368.93 kPa). Cellular debris was removed by centrifugation at 22,040g for 30 min. The clarified lysate was further ultracentrifuged at 176,672g for 40 min.
Polyethyleneimine (10% v/v stock solution, pH 7.5) was added to the supernatant to a final concentration of 0.1% v/v, and the solution was stirred at 4°C for 1 hour. The precipitated proteins were pelleted by centrifugation at 21,000g for 30 min. The pellet was resuspended in buffer A (50 mM tris-HCl, pH 7.5, 0.1 mM EDTA, 1 mM DTT, and 10% glycerol) and centrifuged again at 21,000g for 30 min.
Ammonium sulfate was added to the resulting supernatant to reach a final concentration of 80% saturation, followed by stirring for 1 hour at 4°C. The precipitate was collected by centrifugation at 6000g for 30 min, resuspended in buffer A, and subjected to ultracentrifugation at 176,672g for 40 min. The cleared protein solution was filtered through a 0.45-μm filter.
The filtrate was loaded onto an HQ Poros column equilibrated with buffer A supplemented with 50 mM NaCl and eluted using a NaCl gradient (0.05 to 1 M) in buffer B (buffer A supplemented with 1 M NaCl). Fractions containing gp2.5 were identified by SDS-PAGE, pooled, and dialyzed against dialysis buffer (50 mM tris-HCl, pH 7.5, 0.1 mM EDTA, 1 mM DTT, and 50% glycerol). The dialyzed proteins were stored at −80°C until use.
LD655-gp5/trx purification and labeling
The expression and purification of T7 bacteriophage gp5/trx was based on the previously described protocol by Johnson and Richardson (58). Internally YBBR-labeled gp5 was cloned into a pRSFDuet vector. The YBBR tag (DSLEFIASKLA) was introduced between isoleucine-464 and threonine-465. Thioredoxin (trx) was cloned into a pET Duet vector. The plasmids were cotransformed into E. coli BL21 star cells. Cells were grown in 2 liters of TB medium containing Amp, Kan, and K salts at 37°C until reaching an OD_600_ of 1.0. Protein production was induced with 1 mM IPTG and incubated at 37°C for an additional 4 hours. All subsequent purification steps were performed on ice or at 4°C. The cells were harvested by centrifugation (4000g, 15 min), washed with phosphate-buffered saline, and centrifuged again (4000g, 15 min). The resulting cell pellet was frozen in liquid nitrogen. To support cell lysis, the frozen cell pellet underwent a freeze-thaw cycle by thawing in a water bath, refreezing in liquid nitrogen, and thawing again. The pellet was resuspended in lysis buffer [25 mM Hepes-KOH, pH 8.0, 1 mM DTT, 5% (v/v) glycerol, 500 mM KCl, and 20 mM imidazole]. The lysis mix was supplemented with 1× protease inhibitor cocktail, 1× lysozyme, and 1× DNase I. The cells were lysed by three rounds of sonication (5 min, four cycles, ~30% power). The cell lysate was cleared by centrifugation at (27,000g, 30 min).
The cell lysate was applied to a HisTrap (5 ml) column equilibrated in lysis buffer. After sample application, the column was washed sequentially with lysis buffer and wash/desalting buffer [25 mM Hepes-KOH, pH 8.0, 1 mM DTT, 5% (v/v) glycerol, and 200 mM KCl]. The protein was then eluted using a gradient from 20 to 250 mM imidazole [elution buffer: 25 mM Hepes-KOH, pH 8.0, 1 mM DTT, 5% (v/v) glycerol, 200 mM KCl, and 250 mM imidazole]. Peak fractions were pooled based on SDS-PAGE analysis and concentrated using a molecular weight cutoff (MWCO) 10,000 Amicon Ultra Centrifugal Filter unit. The concentrated sample was applied to a HiPrep 26/10 Desalting column equilibrated in wash/desalting buffer to remove imidazole. The fractions were pooled based on SDS-PAGE analysis and incubated overnight at 4°C with Tobacco Etch Virus (TEV) protease at a 1:50 ratio and 1 mg of DNase to cleave the His-tag. The sample was subsequently applied to a HisTrap (5 ml) column equilibrated in wash/desalting buffer, and the flow-through was collected. Next, the sample was loaded onto a HiTrap Heparin (5 ml) column equilibrated in wash/desalting buffer. The protein was eluted using a salt gradient from 200 mM to 1 M KCl. Peak fractions were pooled and concentrated using a MWCO 10,000 Amicon Ultra Centrifugal Filter unit. The sample was then applied to a Superdex 200 Increase 10/300 gel filtration column equilibrated in gel filtration (GF) buffer [25 mM Hepes-KOH, pH 8.0, 1 mM DTT, 10% (v/v) glycerol, 150 mM KCl, and 0.1 mM EDTA]. The protein eluted as a single symmetric peak, corresponding to an approximate molecular weight of 95 kDa (a 1:1 complex of gp5 and trx). Peak fractions were pooled based on SDS-PAGE analysis and spin concentrated.
To produce LD655-gp5 in complex with Trx, YBBR-gp5/Trx was mixed with SFP synthase and LD655-CoA (Lumidyne) at a molar ratio of 1:1:1.5 in GF buffer supplemented with 10 mM MgCl_2_ and incubated at 30°C for 2 hours. The sample was then applied to a Superdex 200 Increase 10/300 gel filtration column equilibrated in GF buffer. Peak fractions were pooled based on SDS-PAGE analysis and concentrated using an Amicon Ultra Centrifugal Filter unit. Aliquots were snap frozen and stored at −80°C. Labeling efficiency was estimated at approximately 94%, based on the extinction coefficients of the gp5/trx complex and LD655. The final protein concentration was determined using a spectrophotometer by measuring absorbance at 280 nm, while protein purity was assessed by the 260/280 absorbance ratio.
DNA handle preparation with multiple biotins
For creating topologically constrained DNA substrates, handles with multiple biotins interacting with the slide surface were assembled. DNA sequence was amplified from lambda DNA via a PCR reaction. Each PCR reaction contained 50 ng of lambda DNA, 6 U of Phusion High-Fidelity DNA polymerase, PCR reaction buffer (1× HF buffer), 200 μM each of dATP, dCTP, dGTP, and dTTP (from the dNTP bundle), 3% DMSO, and the appropriate forward and reverse primers (Not I: oligo1 and oligo2; Xho I: oligo3 and oligo4; table S2).
The PCR product was purified using the QIAGEN QIAquick PCR Purification Kit, following the manufacturer’s instructions. To incorporate biotin molecules into the DNA handle, the concentration of dTTP was reduced and partially replaced with Biotin-11-dUTP. Each PCR reaction contained 50 ng of the purified template from the previous reaction, 15 U of Taq DNA polymerase, 1× ThermoPol reaction buffer, 200 μM of dATP, dCTP, and dGTP, 130 μM dTTP, 70 μM Biotin-11-dUTP, 3% DMSO, and the same forward and reverse primers (Not I: oligo1 and oligo2; Xho I: oligo3 and oligo4; table S2) as in the previous reaction.
The PCR product was purified using the QIAquick PCR Purification Kit, following the manufacturer’s instructions. The handles were digested with either 30 U of Not I–HF or Xho I in 1× rCut Smart Buffer containing 2 μg of DNA per reaction at 37°C for 3 hours, followed by heat inactivation at 65°C for 20 min to stop enzyme activity. The digested DNA was gel purified using gel electrophoresis (0.75% TBE agarose gel, 90 V, 1 hour). The handles were extracted using the QIAquick Gel Extraction Kit. Biotin-labeled handles were freshly prepared before each ligation reaction and kept on ice in a cold room until the ligation step.
Preparation of linear, biotinylated DNA for single-molecule nicking frequency quantification
For single-molecule TIRF assays to quantify nicking frequency, a DNA substrate with topologically constrained ends was generated from pMSuperCos plasmid. Plasmids were isolated from E. coli DH5α using the QIAGEN Plasmid Maxi Kit following the manufacturer’s instructions. Plasmid (125 μg) was digested overnight at 37°C with 125 U of Xho I and Not I–HF in 1× rCutSmart buffer. Before loading the sample onto a Sephacryl S-1000 SF Tricorn 10/300 gel filtration column for size-exclusion chromatography, 0.1% SDS was added to stop the digestion reaction. The column was equilibrated with 10 mM tris-HCl (pH 8.0), 300 mM NaCl, and 1 mM EDTA (S-1000 Buffer). Desired fractions containing the Xho I–Not I fragment were pooled, and precipitated with ethanol at −20°C. The linearized plasmid (Xho I–Not I fragment) was resuspended in 10 mM tris-HCl (pH 8.0) and stored at 4°C until the ligation step.
To construct a torsionally constrained linear DNA substrate, handles containing multiple Biotin-11-dUTP were ligated onto the linearized DNA. The handles were prepared as described in the previous section. The reaction mixture contained 18 μg of linearized DNA, 2000 U of T4 DNA Ligase, 1× T4 DNA ligase buffer, 2 mM ATP, 2 μg of the Not I handle, and 2 μg of the Xho I handle. The reaction was incubated overnight at 16°C. To remove excess handles from the final DNA product, the reaction mixture was loaded onto a Sephacryl S-1000 SF Tricorn 10/300 gel filtration column equilibrated with S-1000 buffer. Peak fractions were pooled, precipitated with ethanol, and resuspended in 10 mM tris-HCl, pH 8.0, and 0.1 mM EDTA. Aliquots of the final DNA substrate were snap frozen in liquid nitrogen and stored at −80°C.
Preparation of forked, biotinylated DNA for single-molecule replication assays
For single-molecule TIRF assays, involving unconstrained and constrained substrates, a preprimed DNA forked substrate was created to bind to the surface through streptavidin-biotin interactions. Depending on the number of biotin molecules present in a DNA substrate, the substrate was free to rotate around a single biotin interaction (unconstrained) or was fixed on the surface (constrained). Both substrates were created from the same 27-kb plasmid backbone in addition to different DNA modules. Similar approaches have been used previously to investigate DNA replication (42, 59) using fluorescence microscopy. A method for investigation of helicase unwinding on long DNA substrates containing one end with multiple digoxigenins has likewise been reported (60).
Plasmids (pMSuperCos) were isolated from E. coli DH5α using the QIAGEN Plasmid Maxi Kit. The purified plasmid was then digested with Xba I and Xho I. The digestion reaction, which contained 125 μg of plasmid DNA, 125 U each of Xho I and Xba I, and 1× rCutSmart buffer, was performed at 37°C overnight. The reaction was supplemented with 0.1% SDS to stop the digest reaction. Xho I–Xba I fragments were purified by gel filtration using a Sephacryl S-1000 SF Tricorn 10/300 gel filtration column equilibrated with S-1000 buffer. Ethanol precipitation was then performed, and DNA pellet was resuspended in tris-HCl, pH 8.0, and stored on ice until the ligation.
Preprimed DNA fork containing Xba I matching sequence was created by annealing oligos (oligo7, oligo8, and oligo9; table S2) in a ratio of 1:6:60 in Duplex Buffer (30 mM Hepes-KCl, pH 7.5, and 100 mM KOAc). The mixture was heated to 95°C for 5 min and then cooled to 4°C at 0.5°C/min. The end piece with a single biotin containing Xho I matching sequence was created by annealing oligos (oligo5 and oligo6; table S2) in a ratio of 6:1 and performing the same annealing procedure described for the fork. The biotin handle with multiple biotins matching Xho I was described in the “DNA handle preparation with multiple biotins” section.
To create unconstrained and constrained substrates, a ligation reaction was performed containing 1× T4 DNA ligase buffer with 2000 U of T4 DNA ligase, 18 μg of Xho I–Xba I fragments, and 2 mM ATP at 16°C overnight. To produce the unconstrained substrate, the ligation reaction additionally included the end piece with a single biotin (oligo5 and oligo6; table S2) and the preprimed fork (oligo7, oligo8, and oligo9; table S2) at 140 nM each. For the constrained substrate, the Xho I biotin handle (2 μg in total) and the single biotin fork (oligo7, oligo8, and oligo9; final concentration, 140 nM; table S2) were additionally included in the reaction. The steps following the ligation reaction were identical for the following steps onward. Excess DNA handles were removed by a Sephacryl S-1000 SF Tricorn 10/300 gel filtration column equilibrated with the S-1000 buffer. The peak fractions containing the ligated substrate were pooled, precipitated with ethanol, and reconstituted in tris-HCl, pH 8.0, and 0.1 mM EDTA. Aliquots of the final DNA substrates were snap frozen in liquid nitrogen and stored at −80°C.
Single-molecule assays
PEG-biotin microscope slide preparation
Glass coverslips (22 mm by 22 mm, Marienfeld) were first cleaned with a Zepto plasma cleaner, then transferred to a glass container filled with acetone containing 2% (v/v) 3-aminopropyltriethoxysilane and incubated for 1.5 min. The reaction was quenched by adding an excess of deionized water (ddH_2_O) and the coverslips were subsequently rinsed with ddH_2_O, dried with compressed air, and baked at 110°C for 1 hour. For functionalization, coverslips were covered with a fresh solution of 0.6% (w/v) Biotin-PEG-succinimidyl carbonate (MW 5000) and 15% (w/v) mPEG-succinimidyl carbonate (MW 5000) in 0.1 M NaHCO_3_ and incubated overnight at room temperature. After incubation, the coverslips were rinsed with ddH_2_O, dried with compressed air, and incubated again with a fresh Biotin-PEG/mPEG solution as described above. Functionalized PEF-Biotin microscope slides were stored under vacuum.
Flow cell preparation
A functionalized PEG-Biotin microscope slide was covered with streptavidin (0.2 mg/ml) in blocking buffer [20 mM tris-HCl, pH 7.5, 50 mM NaCl, 2 mM EDTA, bovine serum albumin (BSA; 0.2 mg/ml), and 0.005% (v/v) Tween 20] for 30 min. On top of the previously washed and dried slide, a polydimethylsiloxane (PDMS) block was placed to assemble a flow cell. The PDMS block had either the linear or the transverse flow cell configuration. The flow channels had a height of 0.1 mm and a width of 0.5 mm. Polyethylene tubes with an inner diameter of 0.58 mm were placed in the inlets and outlets of the PDMS. The flow channel was flushed with blocking buffer and incubated for 15 min.
The DNA tethering process varied depending on the flow cell configuration. For the linear flow cell configuration, 5 pM of forked, biotinylated DNA was introduced at a flow rate of 17 μl/min for 24 min in reaction buffer [40 mM tris-HCl, pH 7.4, 50 mM potassium glutamate, 10 mM magnesium chloride, BSA (0.1 mg/ml), and 10 mM DTT] supplemented with 200 μM chloroquine. Unbound DNA was then washed out using reaction buffer supplemented with 300 μM ATP/CTP, 600 μM dNTPs, and 150 nM SYTOX Orange at a flow rate of 20 μl/min for 10 min. For the preassembly condition, the reaction buffer was supplemented with 300 μM ATP/CTP; 600 μM dATP, dTTP, and dGTP; and 150 nM SYTOX Orange.
For the transverse flow cell configuration, inlet 2 and outlet 2 were clipped off with metal clips to prevent side flow. Then, 5 pM of forked, biotinylated DNA was introduced at a flow rate of 20 μl/min for 24 min in reaction buffer. Unbound DNA was washed out with reaction buffer containing 300 μM ATP/CTP, 600 μM dNTPs, and 150 nM SYTOX Orange at a flow rate of 20 μl/min for 10 min, while inlet 2 and outlet 2 remained clipped. The clips were then removed from inlet 2 and outlet 2, and inlet 1 and outlet 1 were clipped. Washing buffer was introduced at a flow rate of 20 μl/min for 5 min to ensure an even distribution of SYTOX Orange throughout the flow cell.
Single-molecule photoinduced nicking assay
To introduce negative supercoils in the linear, biotinylated DNA, the DNA application process in the transverse flow configuration was slightly modified. The workflow for inducing negative supercoils in DNA was adapted from the protocol by Ganji et al. (32). During the DNA tethering step, the linear, biotinylated DNA was mixed in reaction buffer containing 500 nM SYTOX Orange. The DNA was applied following the same procedure described for the transverse flow cell configuration, with the SYTOX Orange concentration reduced to 150 nM during the subsequent wash steps, which facilitated the formation of plectonemes. Different imaging conditions with varied laser powers were tested and compared.
Single-molecule replication assay in linear configuration
To initiate the T7 replication reaction in the linear configuration, a solution containing 2.5 nM gp4 (hexamer), 20 nM LD655-gp5/trx, and 750 nM gp2.5 in reaction buffer supplemented with 300 μM ATP/CTP, 600 μM dNTPs, and 150 nM SYTOX Orange was introduced into the prepared flow cell at a flow rate of 150 μl/min for 1.5 min. For experiments with DNA gyrase, 5 nM was added to the T7 replication mix. The inlet and outlet tubes were then clipped to prevent any flow fluctuations or external force. The proteins were allowed to incubate for 0.5 min before starting image collection.
For the preassembly condition, a solution containing 2.5 nM gp4 (hexamer), 20 nM LD655-gp5/trx, and 750 nM gp2.5 in reaction buffer supplemented with 300 μM ATP/CTP; 600 μM dATP, dTTP, and dGTP; and 150 nM SYTOX Orange was introduced into the prepared flow cell at a flow rate of 150 μl/min for 1.5 min. The inlet and outlet tubes were then clipped, and the flow cell was incubated for 1.5 min. Depending on the preassembly condition, the flow cell was washed with replication buffer containing 300 μM ATP/CTP, 600 μM dNTPs, and 150 nM SYTOX Orange, or with replication buffer supplemented with an additional 20 nM gp5/trx and 750 nM gp2.5. The solution was flushed into the flow cell at 150 μl/min. The inlet and outlet tubes were clipped, and data collection was started.
Imaging conditions
Single-molecule experiments were performed using a micromirror TIRF microscope from Mad City Labs (MCL, Madison, WI, USA) with custom modifications. The microscope was equipped with an Apo N TIRF 60× oil-immersion TIRF objective (numerical aperture 1.49, Olympus). All experiments were performed in a temperature-controlled room at 22.5° ± 0.5°C. SYTOX Orange and LD655 dyes were excited with a 532- and 637-nm laser (OBIS 532 nm LS 120 mW and OBIS 637 nm LX 100 mW, Coherent). An emission filter (ZET532/640 m, Chroma) was used to remove residual scattered light from excitation and to separate signals. Emission light was split at 635 nm (T635lpxr, Chroma) and collected on a Photometrics PrimeBSI sCMOS camera and, later, on an Andor iXon Life 888 EMCCD camera with comparable specifications. Protein or DNA was visualized sequentially every 5 to 10 s with a 100-ms integration time for 20 to 30 min. All microscope parts were controlled by Micromanager v2.0.0 (61, 62) and custom BeanShell scripts.
Single-molecule data analysis
Image processing
All single-molecule raw data were processed in Fiji (63) using Molecule Archive Suite (Mars) commands (64). Stuck dots on the slide surface were used to correct for stage drift. For the linear configuration, a Single Molecule Archive was generated by tracking with subpixel resolution individual lagging-strand products moving along the DNA. To generate a DNA Molecule Archive, individual DNA molecules were fit and checked for colocalization with individual molecule trajectories from the previous step. The DNA blob position on DNA versus time was fitted with a kinetic change-point algorithm (65) by assigning individual regions to distinguish between reaction and stalling.
Nick event analysis, quantification, and fitting
To analyze nicking events, supercoiled DNA molecules were fitted using the “Object Tracker” tool from Mars. Upon nick introduction, the area of the DNA molecule increases, which could be analyzed using the kinetic change-point step fitting. In addition, each molecule was inspected manually. Survival curves of the molecules were fitted using Python.
To create survival curves, the percentage of unnicked molecules at each time point was calculated by dividing the number of unnicked molecules at a specific time point by the total number of constrained molecules present at the start of the measurement. This percentage was then plotted against time. The half-life was determined by fitting the curves with an exponential decay model, where b represents the half-life of the curve and y0 denotes the initial quantity
Spatial-temporal protein dynamics and kinetics in linear flow cell configuration
To analyze replication events in the linear flow cell configuration, the lagging-strand product of active molecules was tracked with “Peak Tracker,” creating a Single Molecule Archive. Tracking was corrected for stage drift by identifying immobile peaks and using the coordinates as a reference. DNA molecules were fitted and tracking coordinates were transformed onto the reference frame of individual DNAs, creating a DNA Molecule Archive allowing the description of kinetics in terms of base pairs. An activity region was defined for each molecule, and kinetics were fitted with the kinetic change point. Molecules were sorted in subgroups. Molecules that exceeded the set time point for the 95% confidence interval were excluded. Rates lower than 15 bp/s (sixfold rate reduction compared to the mean rate) and durations >25 s were considered pausing events. For the processivity, the endpoint of the final segment fitted by the kinetic change point was considered as the final position.
Processivity estimation using MLE
Polymerase processivity was estimated using MLE for an exponential survival model with right censoring. The method treats polymerase dissociation as a stochastic process where the probability of dissociation follows an exponential distribution characterized by a rate parameter λ. Observed dissociation events before the template end contribute to the likelihood through their probability density function, while polymerases completing the full template length are treated as right-censored observations contributing through their survival probability. The rate parameter λ is estimated by minimizing the negative log-likelihood function, and the mean processivity is calculated as 1/λ. Confidence intervals are derived using Fisher information, and the model’s goodness of fit is assessed by comparing predicted versus observed completion fractions at the template cutoff length.
DNA product shape analysis in the linear flow cell
A kymograph of each molecule was created using the “DnaArchiveKymographBuilder” tool from Mars-kymograph by averaging 5 pixels on each side of the marked DNA molecule for every time step. Pixel values were saved in a table. Results were smoothed with a sliding window (window size, 2 pixels). The “threshold_otsu” function from the scikit-image package was used to find the global threshold for each kymograph. Pixels below the threshold were set to 0 (black) while those above result in a value of 1 (white). Blob size of the DNA was the number of white pixels. The average size was calculated from five time points at the end of the reaction. The 5 time points were chosen from the interval between 10 and 5 time points preceding the final time point of the reaction.
Quantification of labeled gp5 protein and polymerase estimation
The average intensity of a labeled protein gp5 was quantified by analyzing bleaching steps in surface-immobilized molecules. The step size for each experiment was determined separately. Bleaching steps were fitted using kinetic change point step fitting. Images were beam profile corrected. Protein spots were tracked using Peak Tracker and simultaneously integrated (inner radius 3 and outer radius 12) to measure protein signal. The resulting intensity value was divided by the estimated step size to calculate the number of polymerases.
Rate and processivity analysis for transverse flow
Transverse flow replication data were imported into Labkit (66) and replication molecules were manually segmented. Strands of “leading,” “lagging,” and “parental” were traced in each frame. Segmentation files were imported into Mars using a custom importer (“Transverse Flow Archive Builder”). Lengths of each segment are saved in pixels and converted to base pairs for each time point. After import into Mars, replication rates and processivities were determined in a manner similar to that used for the linear flow cell configuration. Kinetic change point analysis was applied to extract replication rates for the leading, lagging, and parental strands. Manual segmentation of the individual strands was more challenging during fork rotation events because it was more difficult to accurately identify the location of the replisome without the three-way junction being visible. This can result in more variability in the rate distribution as displayed in Fig. 4D.
Analysis of lagging-strand arm length dynamics
Lagging-strand arm length over time was first tested using the Dickey-Fuller unit root test (67) to examine whether trace is nonstationary. Time traces were transformed into a stationary function by taking the difference between length at a time point t x(t) and length at previous time point x(t − 1). Traces were tested with the augmented Dickey-Fuller unit root test to confirm stationarity. The variance of the length changes of the stationary function was calculated.
Estimating forces using the WLC model
Force was estimated using the WLC model, where kB is the Boltzmann constant, T is the temperature (296.15 K), L is the contour length, A is the persistence length (50 nm) (68, 69), and z the current extension of the DNA for each time step
To estimate the force on the DNA arch consisting of leading and parental strand, contour length L was calculated by multiplying DNA length in base pairs times 0.34 nm (length per base pair) and compared to the measured arch length. To estimate the force on the replisome, the extension of the lagging-strand product was compared to the expected length based on the leading-strand extension for the same time point. Forces were grouped into bins of 3 kb, and a moving average with a step size of 1.5 kb was applied. Values below 5 kb were excluded from analysis. The median force for each bin was calculated and plotted. The resulting data were fitted with a line plot to illustrate the trend in force.
Quantification and statistical analysis
The number of observations (n) is indicated in the figure or figure legends. Errors in this study represent either the SEM or the SD, as indicated. Python packages NumPy, pandas, matplotlib, and seaborn were used for all statistical analysis.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1N. Y. Yao, M. E. O’Donnell, Evolution of replication machines. Crit. Rev. Biochem. Mol. Biol. 51, 135–149 (2016).27160337 10.3109/10409238.2015.1125845 PMC 4979536 · doi ↗ · pubmed ↗
- 2J. L. Sparks, G. Chistol, A. O. Gao, M. Raschle, N. B. Larsen, M. Mann, J. P. Duxin, J. C. Walter, The CMG helicase bypasses DNA-protein cross-links to facilitate their repair. Cell 176, 167–181.e 21 (2019).30595447 10.1016/j.cell.2018.10.053PMC 6475077 · doi ↗ · pubmed ↗
- 3D. T. Gruszka, S. Xie, H. Kimura, H. Yardimci, Single-molecule imaging reveals control of parental histone recycling by free histones during DNA replication. Sci. Adv. 6, eabc 0330 (2020).32948589 10.1126/sciadv.abc 0330 PMC 7500940 · doi ↗ · pubmed ↗
- 4M. K. Zeman, K. A. Cimprich, Causes and consequences of replication stress. Nat. Cell Biol. 16, 2–9 (2014).24366029 10.1038/ncb 2897 PMC 4354890 · doi ↗ · pubmed ↗
- 5L. Postow, N. J. Crisona, B. J. Peter, C. D. Hardy, N. R. Cozzarelli, Topological challenges to DNA replication: Conformations at the fork. Proc. Natl. Acad. Sci. U.S.A. 98, 8219–8226 (2001).11459956 10.1073/pnas.111006998 PMC 37424 · doi ↗ · pubmed ↗
- 6M. S. Guo, R. Kawamura, M. L. Littlehale, J. F. Marko, M. T. Laub, High-resolution, genome-wide mapping of positive supercoiling in chromosomes. e Life 10, e 67236 (2021).34279217 10.7554/e Life.67236 PMC 8360656 · doi ↗ · pubmed ↗
- 7C. J. Ullsperger, A. V. Vologodskii, N. R. Cozzarelli, “Unlinking of DNA by topoisomerases during DNA replication,” in Nucleic Acids and Molecular Biology, F. Eckstein, D. M. J. Lilley, Eds. (Springer Berlin Heidelberg, 1995), pp. 115–142.
- 8D. Branzei, M. Foiani, Maintaining genome stability at the replication fork. Nat. Rev. Mol. Cell Biol. 11, 208–219 (2010).20177396 10.1038/nrm 2852 · doi ↗ · pubmed ↗