Dynamics of the PriA Helicase at Stalled DNA Replication Forks

Abstract

The DNA helicase PriA is a key protein for restarting stalled DNA replication forks in bacteria. With 3′ to 5′ helicase activity, PriA is important in primosome assembly. We used atomic force microscopy (AFM) and specifically employed time-lapse AFM to visualize the interaction of PriA with two DNA substrates. The results show that most of the PriA molecules are observed bound at the fork. However, PriA is capable of translocating over distances of about 400 bp. There is a preference for the long-range translocation of PriA depending on the fork type. For a fork with the nascent leading strand as single-stranded DNA (ssDNA; F4 substrate), PriA translocates preferentially on the parental arm of the fork. For the substrate F14, which contains an additional ssDNA segment between the parental and lagging arms (5 nt gap), PriA translocates on both the parental and lagging strand arms. These data suggest that transient formation of the single-stranded regions during the DNA replication can change the selection of the DNA duplex by PriA. Translocation of the helicase was directly visualized by time-lapse AFM imaging, which revealed that PriA can switch strands during translocation. These novel features of PriA shed new light on the mechanisms of PriA interaction with stalled replication forks.

INTRODUCTION

The inherently accurate and highly processive DNA replication machinery is essential for genome duplication.¹⁻³ However, the DNA replication machinery often stalls due to roadblocks, such as tightly bound protein complexes or DNA lesions.⁴⁻⁶ The stalled replication fork needs to be resurrected to avoid genome instability, cell death, and, in higher organisms, cancer.⁷

The DNA helicase PriA is one of the key proteins required for replication restart in bacteria.⁸⁻¹¹ It binds to repaired forks and facilitates the loading of the replicative helicase DnaB onto the lagging-arm template.¹² In addition, PriA is a 3′−5′ DExH helicase classified in Superfamily 2 that can also remodel DNA replication fork structures using its helicase activity to unwind the lagging arm preferentially.^13,14 The helicase is composed of two domains: an N-terminal DNA binding domain and a C-terminal helicase domain (HD). PriA binds to DNA in a manner either dependent on or independent of 3′-terminus recognition.¹⁵ Crystallographic data for PriA and PriA/DNA replication fork reveal the structural mechanism by which PriA recognizes and processes branched DNA replication forks.^11,16 In a structure-specific manner, PriA recognizes the abandoned DNA replication fork and is then proposed to remodel the stalled fork to form a single-strand DNA (ssDNA) loop by pulling DNA across its HD.^11,16 In the presence of ATP, PriA binds and interacts with the three arms of a DNA replication fork. The movements of the helicase lobe of PriA have been proposed to enable PriA to “pull in” that arm to form Dloops.¹⁶⁻¹⁸ In the pathway of PriA-dominated replication restart, PriA only needs to unwind few base pairs of the lagging arm to promote the replication restart.^11,19,20 However, there are still some unknown features of PriA: does the helicase remain bound at the fork position during translocation? How does the ssDNA at fork position regulate the enzyme’s translocation activity?

In our previous study, we used atomic force microscopy (AFM) to investigate the structure-dependent, fork-recognition pattern of PriA fork in the absence of ATP. We found that the PriA binds preferentially to F4 fork substrates, which have a gap in the nascent leading strand, compared to the other forked or tailed DNA substrates.²¹ Here, we extend our previous studies by visualizing PriA-fork dynamics in the presence of ATP using time-lapse AFM. The results show that most of the PriA molecules are observed bound at the fork. However, PriA is capable of translocating over distances as large as several hundred base pairs. For a fork with ssDNA as the nascent leading arm, PriA translocates preferentially on the parental duplex region of the fork. When there is a 5 nt gap placed between the lagging and parental arms, PriA translocates on both the parental and lagging arms. Time-lapse AFM imaging revealed long-range translocation of PriA and a previously undiscovered property of PriA—the switching of DNA strands during translocation on DNA duplexes. These novel features of PriA shed light on the mechanisms of PriA interaction with stalled replication forks.

METHODS

Protein Preparation.

Purification of the PriA protein followed the method described previously.^21,22 We briefly describe the protocol here. The his-PriA protein was purified by ammonium sulfate precipitation followed by affinity chromatography using HisTrap FF crude column, SP Sepharose column [equilibrated with 20 mM potassium phosphate, pH 7.6, 150 mM KCl, 0.1 mM ethylenediaminetetraacetate (EDTA), and 1 mM dithiothreitol (DTT); eluted with a linear 150−500 mM KCl gradient], and Heparin column [equilibrated with 20 mM Tris-OAc, pH 7.5, 0.1 mM EDTA, 1 mM DTT, 10% (v/v) glycerol, and 100 mM KCl; Eluted with a linear KCl gradient of 100−600 mM]. Fractions containing PriA were pooled and dialyzed overnight against storage buffer [20 mM Tris-HCl (pH 7.5), 1 mM DTT, 400 mM KCl, and 50% (v/v) glycerol]. The PriA concentration was determined using an extinction coefficient of 104,850 M⁻¹ cm⁻¹.²³

DNA Substrate Preparation.

The preparation of F4 was described previously.^21,24−26 Briefly, the substrate was assembled from two duplexes and the core fork segment, similar to our previous methodology. The constructs of the two duplexes were precisely the same as the one we used previously. The core fork segment was assembled from the four ssDNA oligos O30, O31, O32, and O33 (Table 1), which were mixed in equimolar ratios and then annealed by decreasing the temperature from 95 °C to room temperature. The two duplexes and the core fork segment were ligated together in a ratio of 1:1:1 at 16 °C overnight. The final products were purified with HPLC using a TSKgel DNA-STAT column. All oligonucleotides were bought from IDT (Integrated DNA Technologies, Inc. Coralville, Iowa, USA). The construction of F14 was the same as for F4 but with O31 replaced with oligo O46 (Table 1).

Table 1.

ssDNA Oligomers Used for DNA Substrate Preparation

oligomer	sequence
O30	TCATCTGCGTATTGGGCGCTCTTCCGCTTCCTATCT
O31	TCGTTCGGCTGCGGCGAGCGGTATCAGCTCACTCATA
O32	GCTTATGAGTGAGCTGATACCGCTCGCCGCAGCCGAACGACCTTGCGCAGCGAGTCAGTGAGATAGGAAGCGGAAGAGCGCCCAATACGCAGA
O33	CACTGACTCGCTGCGCAAGGCTAACAGCATCACACACATTAACAATTCTAACATCTGGGTTTTCATTCTTTGGGTTTCACTTTCTCCAC
O46	CGGCTGCGGCGAGCGGTATCAGCTCACTCATA

Preparation of DNA−Protein Complexes.

Fork DNA construct (F4 or F14) was mixed with the PriA separately in a molar ratio of 1:2, and incubated in 10 μL of binding buffer [10 mM Tris-HCl (pH 7.5), 50 mM NaCl, 5 mM MgCl₂, 1 mM DTT, and 1 mM ATP] for 10 min. The complexes were then ready for deposition on the AFM substrates. The final concentration was 2 nM for the fork DNA and 1 nM for dsDNA.

AFM Dry Sample Imaging and Data Analysis.

Imaging.

1-(3-Aminopropyl)silatrane (APS)-functionalized mica was used as the AFM substrate for all experiments.²⁷ Briefly, freshly cleaved mica was incubated in 4 mL APS (167 μM) in a cuvette for 30 min and then rinsed with double-distilled water (ddH₂O) thoroughly, as described in ref 26. Ten microliters of the sample were deposited onto the APS mica for 2 min, rinsed with ddH₂O, and dried with a gentle Argon gas flow. Images were acquired using tapping mode in the air on a MultiMode 8, Nanoscope V system (Bruker, Santa Barbara, CA) using TESPA probes (320 kHz nominal frequency and a 42 N/m spring constant) from the same vendor.

Data Analysis.

The dry sample AFM images were analyzed using the FemtoScan Online software package (Advanced Technologies Center, Moscow, Russia). The positions of PriA were measured from the end of the short arm on the DNA substrate to the center of the protein. The contour lengths of the DNA were then measured continuously from the center of the protein to the end of the long arm of the DNA. The yields of protein−DNA complexes were calculated from the ratio of the counts of the complexes divided by the total number of DNA molecules.

Time-Lapse Imaging in Aqueous Solutions with Time-Lapse AFM.

The preparation of samples for time-lapse AFM is similar to that reported in our previous research.^24,28,29 Briefly, a mica disk with 1.5 mm in diameter was glued to the glass cylinder and then attached to the time-lapse AFM stage. Then, the mica was cleaved with tape and functionalized with 167 μM APS. The sample (2.5 μL) was deposited on the APS mica and incubated for 2 min. The sample was then rinsed with 20 μL of binding buffer. Time-lapse images were acquired using a commercial time-lapse AFM instrument (RIBM Co. Ltd., Tsukuba, Japan) using custom-built, high-aspect-ratio, and high-frequency carbon probes manufactured as described in ref 30 (based on BL-AC10DS, Olympus Corp., Tokyo, Japan). The image size was usually set to 300 nm × 300 nm, and the scan rate corresponding to the data acquisition was 600 ms/frame. The time-lapse AFM movies were read with Falconview plugin in Igor software (kindly provided by T. Ando) and saved as regular image files. The images were then analyzed using the FemtoScan.

RESULTS

Stalled Fork DNA Constructs.

Fork constructs F4 and F14 used in this study have a 69 nt ssDNA corresponding to the template leading strand arm of the fork and are shown schematically in Figure 1. This ssDNA is flanked by a 280 bp parental duplex and a 396 bp lagging strand arm, creating fork F4, which is the same as our earlier design.^21,25 F14 differs from F4 by the incorporation of a 5-nucleotide ssDNA at the fork between the lagging strand arm and the parental duplex. The full lengths of F4 and F14 substrates were measured and the data are shown as the histograms in Figure S1. The histograms were fitted with Gaussians in Figure S1A,B for F4 and F14 constructs, respectively. The maxima were centered at 660 ± 28 bp (F4 substrate) and 669 ± 31 bp (F14 substrate), which match the expected lengths of both DNA substrates (680 bp).

Figure 1.

Fork substrates to visualize ATP-dependent PriA dynamics. In fork F4, the single-stranded leading strand (69 nt) is sandwiched between two duplex regions corresponding to a shorter parental duplex (280 bp) and a longer lagging arm (393 bp). Fork F14 is similar to F4 but contains a 5 nt ssDNA gap on the lagging arm (388 bp) at the fork position. The gap position is marked in the scheme of F14.

AFM Imaging of Complexes of PriA with Fork DNA Constructs.

To visualize ATP-dependent PriA dynamics, the protein was mixed with F4 or F14 DNA separately in the presence of ATP and incubated for 10 min at 23 °C. Aliquots were taken and the samples were prepared for AFM, as described in the Methods. Typical AFM images are shown in Figure 2. Complexes of PriA with DNA appear as bright features on the DNA molecules. A few of them are indicated with black arrows in both images. The yields of complexes of PriA in the presence of ATP were 11.3 ± 1.4 and 14.3 ± 1.2% for substrates F4 and F14, respectively. Representative zoomed-in images are presented to the right of the large-scale AFM images and show the colocalized PriA at the fork position as expected [image (i) for F4 and image (iii) for F14]. PriA was also observed in positions distal to the forks in F4 and F14 indicated with green arrows in images (ii) and (iv), respectively.

Figure 2.

ATP-dependent PriA translocation. AFM images of PriA bound to forks F4 and F14 in the presence of ATP: (A) fork F4; (B) fork F14 typical complexes of PriA with fork DNA are directed with black arrows in the larger images in each panel. Some PriA binds to the fork position and some bind to the duplex arm. The bar size is 300 nm. The zoomed images to the right of each panel (300 nm × 300 nm) show selected PriA-DNA molecules. The PriA position is indicated by the black arrows and the position of the fork is indicated by green arrows.

To graphically characterize the PriA localization pattern on the DNA substrates, we mapped the positions of PriA on each substrate, and the data are shown in Figure 3. Similar to our previous papers,^21,24,26 the PriA position on the DNA (indicated by the blue triangles) was measured from the end of the parental duplex to the center of the protein. The length of the DNA (in bp) is shown on the y-axis, with a schematic of each substrate to the right of the panel. For both substrates, the parental duplex is the shorter flank of the duplex region, the end of which corresponds to the zero position on the y-axis. Results show that ∼80% of the PriA molecules are bound to the fork position in substrate F4, with the remaining ∼20% bound to duplex regions of parental and lagging arm of the fork (Figure 3A). The histogram in Figure 3B confirms this with a sharp peak centered at 276 ± 25 bp, corresponding to the designed fork position (280 bp). There is a tail to the left of the peak, indicating PriA localization to, predominantly, the parental duplex, compared to the lagging strand of F4. Counting the number of PriA molecules to the left and right of the main peak, the data show that 80% of nonfork-associated PriA molecules are bound to the parental duplex ahead of the fork. The control experiments (Figure S2) show that PriA does not bind to the duplex DNA (1 kb) neither in the absence (Figure S2A) nor presence of ATP (Figure S2B). Furthermore, another set of control experiments reveal that, as expected, PriA binds to the fork and does not translocate in the absence of ATP (Figures S3 and S4). Therefore, these data suggest that in the presence of ATP, PriA is preferentially bound to the fork and is capable of translocation into the duplex arms, with the preference to the parental arm of F4.

Figure 3.

Mapping of the ATP-dependent PriA dynamics. Mapping of PriA on F4 (A) and F14 (C) in the presence of ATP, respectively. A schematic of each substrate is present to the right of each graph. Each DNA was aligned to the end of the parental strand (short duplex). Individual PriA molecules are represented as blue triangles. When a second PriA is present on a DNA molecule, it is colored green. (B,D) Distributions of PriA translocation position relative to the end of the parental duplex for F4 and F14, respectively. The data were fit with a Gaussian.

The results with substrate F14 show some similarities to those observed with F4, but also have significant differences. The inspection of the graph in Figure 3C shows that for F14, the majority of PriA molecules are bound at the fork. Indeed, 60% of the PriA molecules are bound at the fork, as shown by the Gaussian distribution with the peak centered at 276 ± 21 bp (Figure 3D). In contrast to F4, the nonfork-bound locations of PriA on F14 (40% of molecules examined) were distributed equally between the parental duplex and lagging strand arms of the F14. Furthermore, the distribution of the nonfork positions of the protein is broad, which is consistent with ATP-dependent translocation of PriA.

In addition to DNA substrates with a single PriA molecule bound, a small fraction of DNA molecules contained two PriA molecules (Figure 3A, molecules #22, 29, 50, and 64 and Figure 3C, molecules #12, 19, and 43; positions indicated by green triangles). These likely are the result of two PriA molecules binding sequentially, followed by translocation over the duplex arms. Control experiments show that PriA does not bind to duplex DNA directly (Figure S2).

Direct Visualization of PriA Translocation in the Presence of ATP.

The mapping data suggest that in the presence of ATP, PriA can move away from the fork by distances above one hundred base pairs. To directly visualize PriA translocation, we performed time-lapse AFM experiments with the high-speed AFM using approaches described in our previous publications.^24,28,29,31 In these experiments performed with the F14 DNA substrate, PriA was mixed with DNA in a buffer containing ATP for 10 min. The mixture was then deposited onto APS mica and scanned with HS-AFM continuously without drying the sample. Images for each frame are assembled as movies. Each frame in all of the movies takes 600 ms to capture. All these experiments show that PriA translocates on the DNA arms over large distances. One of such experiments is shown as Movie S1. A few frames from this movie file are shown in Figure 4A, and the traces of complexes are placed below each frame for clarity (Figure 4B). These traces were also assembled as a cartoon movie file (Movie S2). The data demonstrate the PriA translocation to the lagging strand arm in a direction away from the fork position and ended with the protein dissociation from the DNA. Graphically, the position of PriA measured from the end of the parental strand is shown in Figure 4C (brown curve). It demonstrates that PriA translocated gradually on the lagging strand arm about 100 bp between frames 6 and 11. It then appeared to stay motionless between frames 11 and 17, followed by a rapid translocation during frames 17−20 (∼100 bp). After that, PriA dissociated from DNA. Similar data were obtained when the PriA position was measured relative to the free end of the lagging strand arm (Figure 4C, gray curve). To determine whether the change in length of the DNA accounts for the change in position of PriA instead of active translocation, we measured the DNA length in each frame (Figure 4C, blue curve). These data show that the length of DNA remains constant, except for frames 17 and 18. The decrease in the contour length measurement in these frames suggests that the end of the lagging arm DNA is partially floating up from the mica surface. Indeed, the PriA position measured from the parental end did not change from frame 16 to 17, which supports the explanation for the dissociation of the lagging arm DNA. As the length of the DNA remains constant and PriA is observed in different positions, we conclude that the change in position was due to ATP-dependent PriA translocation on the lagging strand arm of the fork.

Figure 4.

Time-lapse AFM data of PriA translocation over F14 substrate. (A) Selected frames taken from Movie S1. Each frame takes 600 ms to capture. Scale bar size is 50 nm. Black arrows point to the PriA locations along F14. Green arrows point to the fork when it is visible. (B) Traces of the complexes shown in (A) corresponding to each frame. DNA and PriA are colored with blue and green, respectively. The stars mark the end of the parental arm. (C) Graph shows the position of PriA on F14 dependent on the frame number. The orange dots mark the position of fork position when it is visible. The blue, brown, and gray curves show the full length of DNA, the PriA position measured from the end of parental and lagging arms, respectively.

Translocation of PriA in the parental duplex arm of F14 was observed in another experiment assembled as Movie S3. A few frames from this dataset are shown in Figure 5A and the traces of complexes are placed below each frame for clarity (Figure 5C). Initially (frames 1−16), PriA remained bound at a position of 100 bp from the end of the parental duplex. A burst in translocation occurs between frames 16−17, resulting in the protein translocation by ∼100 bp (Figure 5E). PriA then remained motionless until it dissociated. However, the major difference between this experiment and the previous one was the translocation direction of PriA. In Figure 4, PriA moves away from the fork, but in Figure 5, the translocation occurs toward the fork.

Figure 5.

Time-lapse AFM data of PriA translocation toward the fork of F14 substrate. (A) and (B) Selected frames taken from Movies S3 and S4. Each frame takes 600 ms to capture. Black arrows point to the PriA locations along F14. Green arrows point to the fork when it is visible. Scale bar size is 50 nm. (C,D) Schematics showing the relative position of PriA and DNA in the frames (A,B). The stars mark the end of the parental arm. (E,F) Graphs showing the position of PriA as a function of time in (A,B). Explanations of DNA length, fork position, parental, and lagging arm measurements are in the legend to Figure 4.

Similar data were obtained in another experiment with the data assembled as Movie S4. A few frames and traces of complexes are shown in Figure 5B,D, with the graph of the PriA position measured from both ends of DNA in Figure 5F. The position of PriA fluctuates initially at ∼200 bp from the end of the parental duplex (frames 1−6), followed by a short burst in translocation for 85 bp toward the fork (frames 6 and 7) and protein dissociation. Importantly, the DNA length in each of these movies sometime changes due to the thermal motion and cannot account for the changes in the enzyme position. Consequently, the only way for PriA to change the position is if it was translocating in an ATP-dependent manner toward the fork. Thus, the time-lapse experiments provide direct evidence for the long-range PriA translocation, and also revealed that the direction of translocation can be different.

We were able to directly visualize strand switching that resulted in a change in the direction of PriA translocation, which is described below. The data in Figure 6A show a few frames from the data set assembled in Movie S5 illustrating the direction change of PriA translocation. The traces of the complex are placed below each frame for clarity (Figure 6B). Images in frames 35−47 correspond to translocation of PriA away from the fork, whereas snapshots between 51 and 56 correspond to the movement of PrA in the opposite direction toward the fork position. To quantitatively assess PriA movement, we measured the position of PriA from both ends of the fork substrate and plotted these data in Figure 6C. Similar to Figures 4 and 5, the brown graph in this figure corresponds to the measurements of PriA distances to the end of the parental DNA arm. The gray curve shows the data for the PriA distances measured from the end of the lagging arm of F14 construct. The blue curve corresponds to the contour length measurements of the entire construct, which shows rather low time-dependent variability of the DNA length. Indeed, in this movie, the DNA length does not change significantly and the DNA length distribution shows that the average DNA length is 669 ± 35 bp (Figure S5). At the same time, the position of PriA changes nonmonotonously. From the beginning of the observation, PriA does not change its position between frames 1 and 38 and translocates over a distance of ∼180 bp between frames 38 and 46. Then, PriA moves in the opposite direction toward the fork for a distance of ∼190 bp (frames 47−55). Between frames 56 and 80, there is no net translocation but PriA moves away from the fork for a distance of 200 bp (frames 80−90). At frame 100, the enzyme changes the direction again and translocates for 100 bp back toward the fork. PriA fluctuates on lagging arm at this position and moves back to the fork between frames 116 and 119. Similar multiple strand-switching events were observed in other experiments (data not shown).

Figure 6.

Time-lapse AFM data for reversible PriA translocation over the lagging strand of the F14 substrate. (A) Selected frames taken from Movie S5. Scale bar size is 50 nm. Each frame takes 600 ms to capture. Black arrows point to the PriA locations on F14. Green arrows point to the fork position when it is visible. (B) Traces of complexes shown in (A) corresponding to each frame. DNA and PriA are colored blue and green, respectively. The stars mark the end of the parental arm. (C) Positions of PriA on F14 as a function of frame number. Explanations of DNA length, fork position, parental, and lagging arm measurements are in the legend to Figure 4.

Given that PriA unwinds DNA in the 3′ to 5′ direction;³² therefore, the change in the translocation direction suggests that PriA is capable of switching between DNA strands in the duplex. We discuss this novel property of PriA below.

DISCUSSION

AFM studies revealed three novel properties of PriA in complexes with stalled replication fork DNA in the presence of ATP. These are (1) the specificity of PriA in binding to the different fork DNA molecules with and without gaps at the fork position; (2) the ATP-dependent dynamics of PriA at the fork and the long-range translocation of PriA mediated by ATP hydrolysis; and (3) DNA strand switching by PriA during translocation. Each of these findings is discussed below.

Specificity of PriA in Binding to Fork DNA.

According to our recent paper, in the absence of ATP, PriA binds to the fork present in substrates F4 and F14 with no detectable binding to the duplex DNA.²¹ This is consistent with previous studies showing that fork binding is an essential first step to replication restart and is supported by our control experiments shown in Figures S3 and S4, which show PriA located exclusively at the fork position.^{17,30,33−36} There are two conceivable ways for PriA to bind to these fork substrates. First, PriA binds to the fork by placing the HD on the 69 nt single-stranded leading strand arm, positioning it to translocate in the 3′−5′ direction from the fork position over the parental arm on the substrates as drawn in Figure 1. Second, PriA binds to the fork with the HD positioned on the lagging strand duplex arm instead so that it would translocate along the lagging arm of F14, away from the fork.¹⁷ The second way is on F14 only because F4 does not have an ssDNA region on the lagging strand that is required for PriA translocation. Additional interactions involving the winged-helix domain and the 3′-DNA binding domain further stabilize the binding of the helicase to the fork.³⁷ In the presence of ATP, PriA unwinds a few base pairs of the lagging duplex to facilitate the binding of DnaB and the further loading of the rest of the replication machinery.¹⁶ However, the expected effect (up to 5 bp) is too small to be detected in our system. This follows because the data in Figure 3B,D show PriA bound to the fork on both F4 and F14.

At the same time, the data in Figure 3 reveal PriA locations far away from the fork position, which can be as distant as several hundred base pairs. The overall populations of such nonfork location events are 20% for F4 and 40% for F14. No such complexes have been found for control experiments performed in the absence of ATP (Figure S3). These data are consistent with ATP-dependent translocation of PriA, resulting in fork-distal positions.

ATP-Dependent Translocation and Dynamics of PriA.

The translocation of PriA is supported by the static results in Figure 3, and all time-lapse AFM experiments, including the data shown in Figures 4−6, directly visualize the ATP-dependent translocation of PriA. The AFM data in Figure 3 revealed the partition of PriA translocation on the parental and lagging arms of the fork. Importantly, this effect is dictated by the fork structure. Indeed, for the F4 substrate (Figure 3A), PriA prefers the parental arm to the lagging strand (80%:20%), whereas there is almost no preference for PriA selection of the DNA duplexes for the F14 substrate (Figure 3C). This selectivity of strands can be explained by the affinity of PriA for the ssDNA segments at the fork. Compared to a duplex DNA with an ssDNA tail, when ssDNA is placed in the middle of the duplex (an F4-like substrate), the affinity of PriA for DNA increased 250-fold.³⁴ Control experiments show that PriA does not bind to duplex DNA (Figure S2). Therefore, the 69 nt ssDNA is essential for the binding of PriA to the fork substrate. When PriA is bound to the 69 nt ssDNA of fork F4, the enzyme translocates preferentially onto the parental duplex arm.

As a 3′-to-5′ DNA helicase, PriA also requires at least 3-nt single-stranded DNA at the fork to initiate unwinding.^5,11,36,37 On F4, the HD of PriA binds the 69 nt ssDNA at the fork, leading to translocation onto the parental duplex ahead of the fork 80% of the time. However, there are still 20% of the cases, where PriA translocates on the lagging strand arm. This can be explained by the DNA breathing at the fork, which provides some affinity of PriA to the lagging arm.³⁸ This result is similar to an earlier finding, in which on the F4-like fork with only shorter duplexes arms,³⁶ PriA also unwinds mostly to the parental arm (90%).

The F14 substrate contains an extra 5 nt ssDNA region on the lagging strand arm immediately adjacent to the fork (Figure 1). Previous studies show that 5 nt ssDNA is critical for the initiation of the PriA unwinding activity.^5,36 This allows PriA to bind F14 in two different modes. In the first mode, the HD of PriA binds to the 3′−69 nt ssDNA, orienting the enzyme to translocate toward the fork and onto the parental duplex arm. In the second mode, the HD of PriA binds to the parental duplex, orienting the enzyme in the opposite direction so that translocation on the lagging strand arm is observed. Because PriA translocates on the parental and lagging arms of F14 in an almost equal ratio, it suggests that PriA originally initiates unwinding with the same preference to the 5 nt and 69 nt ssDNA gaps at the fork. This is consistent with the data of PriA unwinding fork DNA in bulk,³⁶ in which similar unwinding efficiency of PriA on the parental and lagging arms has been shown (52% on parental, 46% on lagging, and 2% on both arms).

A recent model proposed where PriA remains at the fork position and translocates DNA toward itself.^18,39 These studies proposed the formation of D-loops containing single-stranded DNA during PriA helicase activity. This model is in line with our data in Figure 3 according to which the fork location is preferable for PriA. At the same time, our data clearly show PriA in positions distal to the fork, consistent with ATP-dependent long-distance translocation on duplex DNA.

DNA Strand Switching of PriA.

Time-lapse AFM revealed a previously undiscovered property of PriA−−the capability to alter the direction of translocation. The time-lapse data in Figure 4 (Movie S1) and Figure 5 (Movies S3 and S4) reveal that PriA can move toward and away from the fork. The data in Figure 6 (Movie S5) show that the PriA can change the translocation direction, so it gradually moves away from the fork position and then moves in the opposite direction, toward the fork. Moreover, these data also show that PriA alters the translocation direction multiple times. Because PriA translocates in the 3′-to-5′ direction, and the enzyme does not bind to dsDNA, the changes in the translocation direction could only occur as a result of strand switching. However, another possible reason for PriA changing the direction could be due to backtracking along the same strand of DNA, as shown for HIM.⁴⁰ However, backtracking normally occurs over a few unwinding steps. In this study, PriA moves away from and toward the fork over a distance of ∼200 base pairs (Figure 6). Therefore, the changes in the translocation direction of PriA are due to the strand-switching property of the enzyme. The strand switching is not without precedent. It has been observed at the single-molecule level for UvrD, which is also a helicase that unwinds the DNA duplex in the 3′−5′ direction.⁴¹

CONCLUSIONS

In this paper, we revealed several novel properties of PriA. We have shown that the fork position is the preferable location for the initial binding of PriA, consistent with what is known about the enzyme.⁸ However, its motor function drives the enzyme away from the fork translocating over distances as large as 400 bp. Importantly, PriA can translocate on the parental and lagging strand DNA arms of the fork, with the final direction dictated by the fork structure. Finally, we found that PriA can switch DNA strands, allowing the enzyme to change the translocation direction. As a result, PriA is capable of moving toward and away from the fork location. It is conceivable that strand switching serves to redirect the enzyme back to the fork so that the resumption of DNA replication can be directed to the lagging strand template arm of the fork. In addition, it may also serve to clear the DNA in the immediate vicinity of the fork, thereby ensuring unobstructed reloading of the replisome once the lagging strand arm has been unwound.