Shuttling of RecQ helicase HIM-6 coordinates iterative cycle of unwinding, pulling, and backsliding

Abstract

RecQ DNA helicases are a highly conserved family of proteins essential for maintaining genome stability. Despite smFRET studies on repetitive DNA unwinding by RecQ using fluorophore-labeled DNA substrates, the domains controlling this behavior, direct visualization of RecQ movement, and in vivo factors such as protein partners and nucleotide modifications affecting it have remained elusive. Using fluorescently labeled HIM-6 fragments and various DNA substrates, we present the sequential functional activities of HIM-6 and its shuttling along DNA. The helicase domain and zinc-binding element together constitute a minimal functional unit responsible for repetitive unwinding, whereas larger fragments containing additional domains exhibited enhanced DNA unwinding activity and additionally acquired a new strand-pulling activity. During strand pulling, HIM-6 remains stationary and subsequently undergoes backsliding. These three activities occur in an iterative manner and coordinates shuttling of HIM-6 along the DNA. Notably, upon encountering a single ribonucleotide in DNA, HIM-6 paused unwinding, transitioned to a pulling mode, and subsequently pulled in the displaced strand, representing a novel and previously unrecognized trigger for an activity switching. Together, these findings provide new insights into the dynamic behavior of RecQ helicase in regulating genome maintenance.

Graphical Abstract

Graphical Abstract.

Introduction

RecQ helicases are a highly conserved family of DNA enzymes essential for genome stability from bacteria to humans [1]. They function in diverse DNA metabolic processes, including replication, recombination, repair, and telomere maintenance. RecQ helicases translocate along DNA in the 3′ to 5′ direction, catalyzing the unwinding of a variety of structured DNA substrates, including replication forks, Holliday junctions, and G-quadruplexes [2]. In humans, five RecQ helicases-RECQL1, BLM, WRN, RECQL4, and RECQL5-have been identified. Mutations in BLM, WRN, and RECQL4 are associated with distinct genetic disorders, Bloom syndrome (BLM), Werner syndrome (WRN), and Rothmund–Thomson syndrome (RECQL4), respectively, characterized by genome instability, cancer predisposition, and premature aging [3–5]. These disorders underscore the vital role of RecQ helicases in protecting cells from DNA damage and genomic instability. Given their central role in DNA metabolism and their implication in genome instability syndromes and cancer, RecQ helicases continue to be subjects of intense research.

Notably, several members of the RecQ family displayed a distinctive behavior known as repetitive DNA unwinding, characterized by cycles of partial duplex separation, re-annealing, and re-unwinding. Interestingly, some of them also revealed single-stranded DNA (ssDNA) reeling activity [6–13]. These dynamic behaviors likely enable RecQ to resolve transient DNA structures, avoiding excessive DNA destabilization. RecQ helicases typically contain the helicase (Hlc) domain and one or more accessory domains such as zinc-binding (ZB) element, the winged-helix (WH) domain, and the Helicase and RNase D C-terminal (HRDC) domain. However, the specific contributions of these domains to these activities remain incompletely understood.

Repetitive unwinding has been demonstrated using single-molecule Förster resonance energy transfer (smFRET) measurements using two-dye-labeled DNA substrates [6–10]. Such measurements provide only indirect evidence for back-and-forth movement of RecQ helicases, as they monitor changes in DNA conformation that correspond to repeated cycles of short-range duplex DNA unwinding and strand rezipping. Direct visualization of this movement is required to confirm whether RecQ shuttling is accompanied by cycles of repetitive DNA unwinding.

Previous studies have reported three different modes underlying DNA rewinding. Firstly, strand switching and translocating back, in which helicases transfer from one strand to the other and actively translocate in ATP-dependent. Escherichia coli UvrD, human BLM, Arabidopsis thaliana AtRECQ2, and E. coli RecQ helicases use this mode [6, 11, 14, 15]. Secondly, switching and sliding back, which allows helicases to switch strands but to return passively. This mode was observed for A. thaliana AtRECQ3, human BLM, and Saccharomyces cerevisiae (S. cerevisiae) Pif1 [11, 13, 16]. Lastly, sliding back on the same ssDNA without strand switching: Gallus gallus (G. gallus) WRN, Caenorhabditis elegans (C. elegans) HIM-6 and WRN-1, human WRN, and S. cerevisiae Pif1, and Schizosaccharomyces pombe (S. pombe) Pfh1 operate in this mode [7, 8, 10, 17, 18]. The mechanism underlying these modes remain a matter of debate and the factors regulating the modes remain unexplored.

HIM-6 (High Incidence of Males-6), the C. elegans orthologue of the human BLM helicase [1]. Loss of HIM-6 function leads to increased genome instability, sensitivity to DNA-damaging agents, and defects in meiotic chromosome disjunction, indicating a critical role in accurate meiotic recombination and the maintenance of genome integrity. [19, 20]. Biochemical studies showed that HIM-6 exhibits characteristic RecQ helicase activities, including ATP-dependent 3′ to 5′ DNA unwinding, with substrate preferences for forked duplexes, D-loops, and recombination intermediates [21]. An smFRET study has dissected the distinct stages of HIM-6 in repetitive DNA unwinding and proposed a molecular-level mechanism underlying its activity [7]. Such activity may generate transient ssDNA intermediates that facilitate protein binding while limiting prolonged ssDNA exposure. This mode of unwinding may enable the helicase to repeatedly probe DNA structure, promoting localized destabilization. Understanding how HIM-6 coordinates dynamic switching between an ATP-dependent unwinding mode and a backsliding mode can resolve the mechanistic ambiguities of RecQ helicases during dynamic DNA processing.

In this study, we examined the repetitive unwinding of HIM-6 by employing its truncated constructs and fluorescently labeling them for the single-molecule measurements. We identified specific functional domains responsible for repetitive DNA unwinding. Beyond unwinding, HIM-6 exhibited a dynamic range of activities including shuttling along the DNA substrate and pulling-in ssDNA. Remarkably, HIM-6 actively pulls in ssDNA when its translocation is restricted by a single ribonucleotide on the tracking strand. These observations collectively provide novel insights into the iterative activities of HIM-6 in genome maintenance.

Materials and methods

Plasmid construction for truncated HIM-6 proteins

DNA fragments corresponding to amino acid residues 205–898 (spanning Hlc to HRDC), 205–788 (spanning Hlc to WH), and 205–650 (spanning Hlc to ZB) of him-6 gene were amplified by polymerase chain reaction (PCR) using the plasmid pDEST17-HIM-6 (encoding amino acid residues 21–988) as a template [21]. A Forward primer was designed to include a PstI restriction site, while a reverse primer contained a 6× His-tag sequence followed by an XhoI site. Amplified products were gel-purified and cloned into the pSNAP-tag(T7)-2 vector (NEB) using the NEBuilder HiFi DNA Assembly Master Mix (NEB), following the manufacturer’s protocol. All recombinant plasmids were verified by DNA sequencing. The resulting expression vectors encoded N-terminally SNAP-tagged and C-terminally 6× His-tagged HIM-6 protein truncations. HIM-6(205–788) fragment lacking the SNAP-tag was constructed by PCR-mediated deletion of the SNAP-tag from the plasmid pSNAP-HIM-6^Hlc_WH with a 5′-phosphorylated forward primer (5′-CACGGCAGATTCCGAGGATT) and a 5′-phosphorylated reverse primer (5′- CATATGTATATCTCCTTCTTAAAGTTAAAC). Amplified products were gel-purified and ligated. The recombinant plasmid was verified by by DNA sequencing. These expression vectors were transformed into E. coli BL21-AI cells (Invitrogen) for protein production.

Purification and fluorescent labeling of proteins

Expression of recombinant SNAP-tagged HIM-6 truncated proteins carried out in E. coli BL21-AI cells. Cultures were grown in 200 ml of Terrific Broth medium supplemented with 2 mM MgSO₄, 20 mM sodium succinate, and 100 µg/ml ampicillin at 37°C until an optical density at 600 nm (OD₆₀₀) of 1.2 was reached. Protein expression was induced by adding 1 mM isopropyl-β-d-thiogalactoside and 0.05% (w/v) L-arabinose (Sigma–Aldrich) and the cells were incubated further at 15°C for 20 h. Cells were then collected by centrifugation at 8000 rpm for 5 min and resuspended in ice-cold lysis buffer [20 mM Tris–HCl, pH 7.4, 300 mM NaCl, 5 mM β-mercaptoethanol, 15% (v/v) glycerol, 0.5 mM phenylmethylsulfonyl fluoride (PMSF), 1× Xpert Protease Inhibitor Cocktail (GenDEPOT), and 0.25 mg/ml lysozyme]. Following a 30-min incubation on ice, cells were lysed by sonication and the lysates were clarified by centrifugation at 20 000 rpm for 30 min at 4°C. The clarified supernatant was applied to a 1-ml HisTrap HP column (Cytiva) pre-equilibrated with buffer B500 [20 mM Tris–HCl, pH 8.0, 500 mM NaCl, 5 mM β-mercaptoethanol, 0.1% Triton X-100, 15% (v/v) glycerol, and 15 mM imidazole]. The column was washed sequentially with 30 column volumes (CV) of buffer B500, 10 CV of buffer B150 [20 mM Tris–HCl, pH 8.0, 150 mM NaCl, 5 mM β-mercaptoethanol, 0.1% Triton X-100, 15% (v/v) glycerol, 15 mM imidazole], and 30 CV of buffer W50 [20 mM Tris–HCl, pH 8.0, 150 mM NaCl, 5 mM β-mercaptoethanol, 0.1% Triton X-100, 15% (v/v) glycerol, 50 mM imidazole] to remove nonspecifically bound protein. Bound protein was eluted with elution buffer [20 mM Tris–HCl, pH 8.0, 150 mM NaCl, 5 mM β-mercaptoethanol, 0.1% Triton X-100, 15% (v/v) glycerol, 200 mM imidazole]. Protein-containing fractions were pooled, desalted using Amicon Ultra-30K centrifugal filters, and further purified by batch binding to HIS-Select Nickel Magnetic Agarose Beads (Millipore). Beads were washed with B150 buffer and eluted with the same elution buffer. Eluted fractions were analyzed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and protein concentrations were measured using the Bradford protein assay (Bio-Rad). These fragments were purified to >95% purity as assessed by SDS–PAGE and Imperial Protein Stain. Purified SNAP-tagged HIM-6 constructs were fluorescently labeled following the protocol provided by in New England Biolabs (NEB) for in vitro labeling SNAP-tagged proteins. Briefly, purified proteins were incubated with a two-fold molar excess of DY-549P1-benzylguanine (SNAP-Surface^® 549, NEB) in labeling buffer [20 mM Tris–HCl, pH 8.0, 50 mM NaCl, 15% (v/v) glycerol, 1 mM dithiothreitol, 1× Xpert Protease Inhibitor Cocktail] at 4°C for 1 h with gentle rotation. The labeling mixture was incubated with HIS-Select Nickel Magnetic Agarose Beads at 4°C for 2 h with gentle rotation to recapture the His-tagged protein. The resin was washed extensively with buffer B150 and labeled protein was eluted with elution buffer (as described above). Labeling efficiencies were assessed by measuring the absorbance at 560 nm using an extinction coefficient of 150 000 M⁻¹ cm⁻¹ and ranged between 85% and 95%.

Preparation and labeling of DNA/RNA substrates

DNA and RNA oligomers were synthesized by Integrated DNA Technologies, with substrate modifications detailed in Supplementary Table S1. Cy3 and Cy5 NHS esters (GE healthcare) were conjugated to thymidine (dT) residues of DNA (or RNA) via a C6 amino linker. For annealing, biotinylated strands (final concentration 1 µM) were mixed with nonbiotinylated strands (1.2 µM) in annealing buffer (10 mM Tris–HCl, pH 8.0, 1 mM EDTA, pH 8.0, 50 mM NaCl). The mixture was heated to 95°C for 5 min and gradually cooled to 4°C at a rate of −1°C per minute.

smFRET assay

All smFRET measurements were performed using a home-built prism-based total internal reflection fluorescence microscopy. Pre-cleaned quartz slides and glass coverslips were coated with a mixture of polyethylene glycol (m-PEG-SVA-5000; Laysan Bio, USA) and biotinylated-PEG (biotin-PEG-SC-5000; Laysan Bio) at a ratio of 80:1. Surface immobilization of the biotinylated DNA was mediated by NeutrAvidin (Thermo Scientific, USA). The standard reaction/imaging buffer contained 50 mM HEPES (pH 7.5), 20 mM KCl, 1 mM MgCl₂, 2 mM DTT, 0.1 mg/ml bovine serum albumin (Sigma), and saturated Trolox (~3 mM), with an oxygen scavenging system [0.1 mg/ml glucose oxidase, 0.6% (w/v) glucose and 0.06 mg/ml catalase] added prior each measurement. Proteins were added to the slide chamber (flat surface) and unbound proteins were washed twice using 15 µl of imaging buffer for each wash. Reaction was started by adding imaging buffer with ATP (at indicated concentrations). All smFRET experiments were recorded by SmCamera software at room temperature (23°C) with exposure time of 30–100 ms.

Binding kinetics assay

The slide chamber was seeded with 10 pM FK-50^A25 substrate and incubated for 2 min. Unbound substrate was removed by extensive washing with imaging buffer followed by the addition of various SNAP-Surface549-labeled HIM-6 constructs while single-molecule fluorescence signals were recorded. FK-50^A25 was used to prevent photobleaching of SNAP549 dye by lowering the FRET level. In cases of high affinity (⁵⁴⁹HIM-6^Hlc_HRDC and ⁵⁴⁹HIM-6^Hlc_WH), 1200-s long movies were recorded with 200-ms time resolution and a 90% reduced laser power (to prevent photobleaching). For detecting the binding and dissociation events of ⁵⁴⁹HIM-6^Hlc_ZB protein, 60-s long movies were recorded with 100-ms time resolution. The binding time (t_bound) and dissociation time (t_unbound) were manually analyzed from individual FRET time traces using custom scripts written in MATLAB. Data were fitted to a single exponential decay function using Origin.

Photobleaching assay

The slide chamber was seeded with 10 pM FK-50 substrate and incubated for 2 min. Unbound substrate was removed by extensive washing with imaging buffer, followed by the addition of various SNAP-Surface549-labeled HIM-6 constructs at concentrations of 1 or 10 nM. After incubation, unbound proteins were removed by washing, and image acquisition was initiated immediately. Imaging was performed with a time resolution of 100 ms. To achieve complete fluorescence photobleaching within 100 s, laser power was increased three-fold. FRET time traces were manually analyzed using custom scripts written in MATLAB.

Analysis of smFRET data

E _FRET was calculated from the fluorescence intensities of donor (D) and acceptor (A) using the formula E = (A – aD)/(A + D) where a = 0.06 represents donor leakage correction. FRET histograms were generated either from traces displaying a single pair of donor and acceptor, acquiring the frames before photobleaching or by averaging the initial 10 frames of FRET traces (~5000 molecules) obtained from 20–30 movies. Data analyses were performed by custom scripts written in MATLAB. smFRET time traces that displayed one-step photobleaching of labeled SNAP549 proteins (indicating monomer binding) were selected for further analyses. Experiments were repeated three times to obtain error ranges as the standard error of the mean.

Alphafold-based structure prediction

Structural models of HIM-6 were predicted using AlphaFold3 [22]. Models with the highest confidence scores were selected for further structural analysis and visualization using UCSF ChimeraX [23]. Representative structures were shown in the supplementary figures.

Results

WH domain confers enhanced unwinding activity to HIM-6 fragment containing only the helicase domain and zinc-binding element

HIM-6 has been reported to repetitively unwind dsDNA [7]. However, the functional unit responsible for its repetitive unwinding has remained unclear. Since HIM-6 consists of multiple domains, we first predicted the structure of HIM-6 using AlphaFold3 (the “Materials and methods” section; Supplementary Fig. S1). Based on the predicted structure, we designed three HIM-6 fragments: HIM-6(205–650) spanning amino acids 205–650, which includes the Hlc domain and a ZB element, HIM-6(205–788) extending to the WH domain, and HIM-6(205–898) extending to the HRDC domain. The three fragments were tagged with 6× His at the C-terminus for affinity chromatography and with a SNAP polypeptide (199 amino acids) at the N-terminus for fluorescent dye labeling (Fig. 1A). The constructs were designated as follows: ^SNAPHIM-6(205–898) as HIM-6^{Hlc_HRDC, SNAP}HIM-6(205–788) as HIM-6^Hlc_WH, and ^SNAPHIM-6(205–650) as HIM-6^Hlc_ZB. All fragments were purified to >95% purity as assessed by SDS–PAGE and Imperial Protein Stain (Supplementary Fig. S2A).

Figure 1.

Helicase activity of truncated HIM-6 constructs. (A) Domain organization of full-length HIM-6 and truncated HIM-6 constructs fused to SNAP and 6× His tags. 6-histidine tag is denoted as “6xHis” and SNAP tag is indicated by SNAP. Protein lengths are shown as numbers, with distinct color codes and symbols representing the domains. Hlc, helicase domain; ZB, zinc-binding element; WH, winged-helix domain; HRDC, helicase and RNase D-like C-terminal domain. (B) Schematic of smFRET-based DNA unwinding assay using HIM-6 fragments. Forked DNA substrate, FK-18^D0, composed of 18-bp duplex with 3′ oligo-dT₃₀ tail and 5′ of oligo-dT₂₄ tail. Tracking strand (3′-tailed) was labeled with Cy3 (D) at ss/dsDNA junction, while the complementary strand was labeled with Cy5 (A) at 7th nucleotide from fork junction toward the 5′ end. (C) Unwinding activity of HIM-6 fragments. HIM-6 fragments (10 nM) were incubated with substrate and unbound protein was washed away with imaging buffer. FRET time traces were analyzed. Representative time traces (left panel) show fluorescence intensities of Cy3 (green) and Cy5 (magenta) and corresponding E_FRET of FK-18^D0 alone. Panels (middle and right) display representative time traces of single-step full unwinding events of FK-18^D0 by HIM-6^Hlc_HRDC and HIM-6^Hlc_WH, respectively. Dotted lines in the FRET traces indicate the time for introducing imaging buffer containing 1 mM ATP-Mg²⁺. (D) Fraction of FK-18^D0 substrate fully unwound over time by HIM-6 fragments. Remaining Cy3 spots were counted to calculate unwinding fraction. Data are shown for HIM-6^Hlc_HRDC (gray), HIM-6^Hlc_ZB/ATP (green), HIM-6^Hlc_WH/ATP (cyan), and HIM-6^Hlc_HRDC/ATP (magenta). Remaining Cy3 spot counts collected from 50–70 movies; data from three independent experiments. (E) Schematic of smFRET-based DNA unwinding assay using fluorescently labeled ⁵⁴⁹HIM-6 fragments (green). Forked DNA substrate FK-18^A18, composed of 18-bp duplex with 3′ oligo-dT₃₀ tail and 5′ oligo-dT₂₄ tail. Cy5 dye was attached to 5′ end of tracking strand, while 3′ end of complementary strand was immobilized on slide surface. (F) Representative time traces showing fluorescence intensities of ⁵⁴⁹HIM-6 construct (green) and Cy5 (magenta), along corresponding E_FRET (left panel). Middle and right panels display representative time traces of single-step full unwinding events of FK-18^A18 by ⁵⁴⁹HIM-6^Hlc_HRDC/ATP and ⁵⁴⁹HIM-6^Hlc_WH/ATP, respectively. Dotted lines in FRET traces indicate time for introducing imaging buffer containing 1 mM ATP-Mg²⁺. (G) Fraction of FK-18^A18 substrate fully unwound over time by HIM-6 fragments. Remaining Cy5 spots were counted to calculate unwinding fraction. Data are shown for ⁵⁴⁹HIM-6^Hlc_HRDC (gray), ⁵⁴⁹HIM-6^Hlc_ZB/ATP (green), ⁵⁴⁹HIM-6^Hlc_WH/ATP (cyan), and ⁵⁴⁹HIM-6^Hlc_HRDC/ATP (magenta). Counts of remaining Cy5 spots from 50–70 movies; data from three independent experiments.

We first examined the unwinding of a forked DNA substrate to measure the helicase activity of the three HIM-6 fragments using smFRET assay. For this, we designed a forked DNA substrate, FK-18^D0, which consists of 18 bp of dsDNA with a 3′ oligo-dT₃₀ tail and a 5′ oligo-dT₂₄ tail (Fig. 1B-first diagram, and Supplementary Table S1). The donor fluorophore Cy3 (D) was labeled at the fork junction on the 3′-tailed strand (the tracking strand) and the location of Cy3 at the fork junction was denoted as D0. The acceptor fluorophore Cy5 (A) was placed 7 nt into the 5′-tailed strand (the displaced strand). The displaced strand of FK-18^D0 was immobilized on the PEGylated quartz slide via neutravidin–biotin interaction. HIM-6 fragment was added to the chamber and allowed to bind to the substrate for 5 min, followed by washing out of unbound protein (the “Materials and methods” section; Fig. 1B-second diagram). Unwinding reaction was then initiated by flowing imaging buffer containing ATP and Mg²⁺ (ATP-Mg) into the chamber (Fig. 1B-wash/ATP step).

FK-18^D0 alone yielded a stable FRET efficiency (E_FRET) of ~0.9 (Fig. 1C, first panel for FK-18^D0 only). Upon the addition of ATP-Mg (vertical dotted lines, Fig. 1C), HIM-6^Hlc_HRDC and HIM-6^Hlc_WH displayed two distinct types of FRET traces. One type showed a single drop in E_FRET followed by the disappearance of fluorescence signals (Fig. 1C), while the other exhibited repeated E_FRET fluctuations (Supplementary Fig. S3A). In the first type, once HIM-6^Hlc_HRDC and HIM-6^Hlc_WH began unwinding, the process continued, resulting in the complete displacement of the Cy3-labeled strand. In the second type, E_FRET fluctuated between 0.9 and 0.2 before the disappearance of Cy3 signal (Supplementary Fig. S3A), indicating that the HIM-6 fragment unwound the substrate repetitively before complete unwinding. To measure the rate of full unwinding, Cy3 spots per imaging area were counted and the fraction of fully unwound DNA molecules was plotted as a function of time (Fig. 1D). The fraction of fully unwound DNA by HIM-6^Hlc_WH or HIM-6^Hlc_HRDC was larger than that by HIM-6^Hlc_ZB, suggesting that the WH domain promotes DNA unwinding. Measurements with both the unlabeled and labeled HIM-6 fragments showed that ⁵⁴⁹HIM-6^Hlc_HRDC exhibited both one-step and repetitive unwinding to a similar extent, while HIM-6^Hlc_WH preferentially exhibited repetitive unwinding (Supplementary Fig. S3A). Although HIM-6^Hlc_ZB has weak helicase activity, it also displayed similar traces (Supplementary Fig. S3B).

We next monitored whether HIM-6 fragments could translocate and reach the end of the duplex region before they fully unwind the forked substrate. For this experiment, we labeled HIM-6 fragments with SNAP-Surface^® 549 (referred to as SNAP549). An SDS–PAGE analysis showed a single fluorescent band, confirming the proper labeling of the HIM-6 fragments and the labeling efficiency ranged from 85% to 95% (the “Material and methods” section; Supplementary Fig. S2B). The labeled HIM-6 fragments were denoted as ⁵⁴⁹HIM-6^{Hlc_ZB, 549}HIM-6^Hlc_WH, and ⁵⁴⁹HIM-6^Hlc_HRDC. The number 549 in ⁵⁴⁹HIM-6 indicates that SNAP-tag was labeled with SNAP549. To monitor FRET between the labeled HIM-6 fragment and DNA, we designed a forked DNA substrate in which the duplex 5′ end of the tracking strand was labeled with Cy5, and the 5′ end of the nontracking was immobilized on the surface, as in Fig. 1B (Fig. 1E, first diagram-FK-18^A18). If the labeled HIM-6 fragment initiates unwinding and reaches the end of the forked DNA, an increase in E_FRET is expected (Fig. 1E, mid FRET). The disappearance of Cy5 and SNAP549 signals indicates full unwinding of the forked DNA (Fig. 1E, Cy5 disappearance).

In the absence of ATP-Mg, a stable E_FRET ~0.2 was observed (Fig. 1F, first panel). In the presence of ATP-Mg_, the FRET traces again exhibited the rise in E_FRET followed by the simultaneous and rapid loss of both Cy5 and SNAP549 signals. This indicates that ⁵⁴⁹HIM-6^Hlc_WH and ⁵⁴⁹HIM-6^Hlc_HRDC reached the duplex terminus, completely unwound the Cy5-labeled strand, and dissociated from the DNA immediately afterwards. The fraction of fully unwound DNA was calculated, revealing comparable contrasts among ⁵⁴⁹HIM-6^{Hlc_HRDC, 549}HIM-6^Hlc_WH, and ⁵⁴⁹HIM-6^Hlc_ZB (Fig. 1G) as determined from measurements of the unlabeled HIM-6 fragments under the same unwinding direction as that used in Fig. 1B.

A single monomer of HIM-6 is preferentially localized at the fork junction

Fluorescently labeled protein enables determination of the binding species, the kinetics of the interaction, and the DNA-binding sites. We first assessed determined the oligomeric states of ⁵⁴⁹HIM-6 fragments during DNA binding. Since photobleaching of individual fluorophores occurs stochastically, the number of discrete photobleaching steps reflects the number of fluorophores, and hence the number of labeled protein molecules bound to the DNA [24–26]. After binding ⁵⁴⁹HIM-6 fragment to the unlabeled forked DNA substrate (FK-50), they were excited with a strong laser power to photobleach the bound ⁵⁴⁹HIM-6 fragments. Drops of SNAP549 signal were observed (Fig. 2A). For ⁵⁴⁹HIM-6^Hlc_HRDC and ⁵⁴⁹HIM-6^Hlc_WH equilibrated at 1 nM, the majority (~90%) of the traces exhibited single-step photobleaching events. Although ⁵⁴⁹HIM-6^Hlc_ZB bound weakly to the forked DNA, nearly 90% of the traces also showed single-step photobleaching (Fig. 2B, beige bars). Less than 10% of the traces exhibited multistep photobleaching events for all the HIM-6 fragments (Fig. 2B, green bar). These results indicate that under low concentration conditions, all three HIM-6 fragment bind to the ss/dsDNA junction as a single monomer. At 10 nM of ⁵⁴⁹HIM-6^Hlc_HRDC or ⁵⁴⁹HIM-6^Hlc_WH single-step photobleaching event were observed in nearly 70%, while single-step photobleaching events were observed for ⁵⁴⁹HIM-6^Hlc_ZB showed in only 40% (Fig. 2B, hatched beige bar). These results suggest that ⁵⁴⁹HIM-6^Hlc_HRDC and ⁵⁴⁹HIM-6^Hlc_WH bind mainly as a single monomer, while ⁵⁴⁹HIM-6^Hlc_ZB can bind as multiple monomers.

Figure 2.

HIM-6 preferentially binds to ss/dsDNA junction. (A) Representative fluorescence time traces for DNA-bound HIM-6 from photobleaching analysis. ⁵⁴⁹HIM-6 fragment was incubated with FK-50 substrate, followed by washing away with imaging buffer. Imaging acquisition was initiated with laser power three times higher than standard conditions to photobleach ⁵⁴⁹HIM-6 fragments. Representative fluorescence time traces: Single-step photobleaching event (upper trace) and multistep photobleaching event (lower trace). (B) Fractions of single monomer and multiple monomers bound to FK-50 substrate. Single-step photobleaching indicates single monomer bound, while multiple-step photobleaching indicates multiple monomers bound to the substrate. Photobleaching events were counted and converted into fractions of DMA-bound single monomer and DNA-bound multiple monomers. Bar graphs show fractions measured with ⁵⁴⁹HIM-6^{Hlc_HRDC, 549}HIM-6^Hlc_WH, or ⁵⁴⁹HIM-6^Hlc_ZB at 1 or 10 nM. Solid beige and green bars show the fractions at 1 nM protein, while hatched beige and green bars show fractions at 10 nM protein. Fraction of single monomer at 1 nM protein: ⁵⁴⁹HIM-6^Hlc_HRDC (91%, data from n = 217 traces), ⁵⁴⁹HIM-6^Hlc_WH (90%, n = 159), ⁵⁴⁹HIM-6^Hlc_ZB (95%, n = 59). Fraction of multiple monomers at 10 nM protein: ⁵⁴⁹HIM-6^Hlc_HRDC (70%, data from n = 675 traces), ⁵⁴⁹HIM-6^Hlc_WH (67%, n = 549), ⁵⁴⁹HIM-6^Hlc_ZB (43%, n = 252). (C) Dwell time analysis of HIM-6 fragments to FK-50^A25 DNA substrate. Left panel: Cy5-labeled DNA substrate. Right panel: Representative time traces showing fluorescence intensities of ⁵⁴⁹HIM-6 (green) and Cy5 intensities, along corresponding E_FRET in the presence of ⁵⁴⁹HIM-6^Hlc_HRDC (upper trace, 1 nM), ⁵⁴⁹HIM-6^Hlc_WH (middle trace, 1 nM), or ⁵⁴⁹HIM-6^Hlc_ZB (lower trace, 1 nM). Time traces were recorded at 100 ms frame rate for ⁵⁴⁹HIM-6^Hlc_ZB, and at 200 ms for ⁵⁴⁹HIM-6^Hlc_HRDC and ⁵⁴⁹HIM-6^Hlc_WH. Increases in ⁵⁴⁹HIM-6 fluorescence intensity and E_FRET indicate protein-binding event. Dwell times in bound (t_bound) and unbound (t_unbound) states were extracted from traces. (D) Kinetic analysis of dissociation constants for HIM-6 fragments. Binding (k_on) and dissociation (k_off) rate constants for each HIM-6 fragment were determined from dwell-time histograms: (i) ⁵⁴⁹HIM-6^Hlc_HRDC (n = 187 traces), (ii) ⁵⁴⁹HIM-6^Hlc_WH (n = 80), (iii) ⁵⁴⁹HIM-6^Hlc_ZB (n = 312). Data were fitted using single-exponential functions. Error bars represent standard deviation of at least three independent replicates. Dissociation constant (K_d) was calculated as k_off/k_on. (E) Schematic of smFRET-based DNA binding assay using labeled ⁵⁴⁹HIM-6 constructs. FK-50^An substrates (where n = 0, 8, 18, or 25), composed of 50-bp duplex region with 3′ oligo-dT₃₀ tail and 5′ oligo-dT₂₄ tail. Cy5 dyes (A) were placed at multiple positions along tracking strand, including ss/dsDNA junction (FK-50^A0) and downstream within duplex region-8th (FK-50^A8), 18th (FK-50^A18), and 25th nucleotides (FK-50^A25). 5′ end of unlabeled complementary strand was immobilized on slide surface. (F) Representative fluorescence images showing labeled ⁵⁴⁹HIM-6^Hlc_HRDC with DNA substrates. ⁵⁴⁹HIM-6^Hlc_HRDC (1 nM) was incubated and then washed away. Image is captured. Left panel: Fluorescently labeled ⁵⁴⁹HIM-6^Hlc_HRDC fragment; middle panel: Cy5-labeled DNA substrate; right panel: Merged image illustrating colocalization. In merged image, red circles represent DNA, green circles represent ⁵⁴⁹HIM-6^Hlc_HRDC, and yellow circles indicate DNA molecules colocalized with ⁵⁴⁹HIM-6^Hlc_HRDC. (G) Histograms of FRET efficiencies for DNA binding by HIM-6 fragments. Indicated ⁵⁴⁹HIM-6 fragments (1 nM) were incubated each FK-50^An substrates. E_FRET values of colocalized molecules were analyzed. Panel (i) shows Cy5-labeled forked DNA substrates used for DNA binding assay. Panels (ii–iv) display histograms corresponding data from FRET traces: (ii) ⁵⁴⁹HIM-6^Hlc_HRDC (n = ~4900 per substrate), (iii) ⁵⁴⁹HIM-6^Hlc_WH (n = ~4300 per substrate), (iv) ⁵⁴⁹HIM-6^Hlc_ZB (n = ~4000 per substrate). Red curves represent Gaussian fits to the data. (H) FRET peak value from panel (G) plotted at different Cy5 positions.

We next examined the binding kinetics of interaction between the ⁵⁴⁹HIM-6 fragments and DNA. On-and-off-DNA binding times of ⁵⁴⁹HIM-6 fragments for the FK-50^A25 were measured t_bound and t_unbound, respectively from FRET traces (Fig. 2C). The association rate constant (k_on) and dissociation rate constant (k_off) were determined from the on and off time distributions and subsequently dissociation constant (K_d) was calculated as a ratio between two constants (Fig. 2D). ⁵⁴⁹HIM-6^Hlc_HRDC and ⁵⁴⁹HIM-6^Hlc_WH exhibited K_d values of 0.7 nM (Fig. 2D-i) and 0.5 nM (Fig. 2D-ii), respectively, while ⁵⁴⁹HIM-6^Hlc_ZB exhibited a much higher K_d of 6.1 nM (Fig. 2D-iii). These results suggest that ⁵⁴⁹HIM-6^Hlc_HRDC and ⁵⁴⁹HIM-6^Hlc_WH exhibit higher affinity for the substrate than ⁵⁴⁹HIM-6^Hlc_ZB, and that ⁵⁴⁹HIM-6^Hlc_ZB rapidly associates and dissociates, reflecting its lower stability on the substrate.

We next investigated the DNA-binding regions preferred by ⁵⁴⁹HIM-6 fragments using colocalization between fluorescently labeled protein and the DNA [26–29]. A forked DNA substrate, consisting of a 50 bp of duplex, a 3′ oligo-dT₃₀ tail, and a 5′ oligo-dT₂₄ tail, was labeled with Cy5 at various positions within the duplex region of the tracking strand (Fig. 2E, left diagram-FK-50^An). In the FK-50^A0 substrate, Cy5 was placed at the fork junction, whereas in the FK-50^A18 substrate, Cy5 was located 18 nt from the junction within the duplex. The association and dissociation of ⁵⁴⁹HIM-6 fragments with FK-50^An substrates were tracked by single molecule imaging. Following the addition of ⁵⁴⁹HIM-6 fragments (1 nM) at which they bound to DNA as a single monomer (Fig. 2B), fluorescent spots appeared in SNAP549 and Cy5 channels, respectively (Fig. 2F), and colocalized spots were indicated in yellow circles (Fig. 2F, colocalization). Since E_FRET is strongly dependent on the inter-dye distance, the efficiency is expected to increase as the distance between ⁵⁴⁹HIM-6 fragment and Cy5 decreases. The interaction of ⁵⁴⁹HIM-6^Hlc_HRDC with FK-50^A0 exhibited a major population at E_FRET ~0.45, and E_FRET ~0.55 with FK-50^A8 (Fig. 2G-ii and H). ⁵⁴⁹HIM-6^Hlc_WH also exhibited comparable levels of E_FRET ~0.45 with FK-50^A0 and E_FRET ~0.5 with FK-50^A8 (Fig. 2G-iii and H). However, for the FK-50^A18 and FK-50^A25 substrates, both fragments showed marked decrease in FRET levels below 0.3, indicating that ⁵⁴⁹HIM-6^Hlc_HRDC and ⁵⁴⁹HIM-6^Hlc_WH preferentially bind to the fork junction (Fig. 2G and H). In the structural model (Supplementary Fig. S4A and B), the fluorophore in SNAP-tag appears to be closer to A8 (67.6 Å) than A0 (75.2 Å) (Supplementary Fig. S4C and D), consistence with the above results. In contrast, ⁵⁴⁹HIM-6^Hlc_ZB lacking the WH domain exhibited a major FRET population less than E_FRET 0.1 for all four FK-50^An substrates (Fig. 2G-iv and H), indicating that the fork junction is unlikely a preferential binding site for ⁵⁴⁹HIM-6^Hlc_ZB. However, we cannot rule out the possibility that bound molecules may be underestimated due to its short lifetime at the fork junction (see binding kinetics data).

These results suggest that the WH domain plays a crucial role in positioning HIM-6 to the fork junction and increasing its bound lifetime. As a negative control, when protein was added to the chamber lacking immobilized DNA substrates (Supplementary Fig. S5A), a few fluorescence spots were detected, indicating that the protein did not strongly bind to the PEGylated surface (Supplementary Fig. S5B). However, upon the FK-50^A25 substrate alone was immobilized (Supplementary Fig. S5C), only Cy5 signal was detected (Supplementary Fig. S5D).

A minimal functional unit for repetitive DNA unwinding

Due to weak unwinding activity of HIM-6^Hlc_ZB on the forked substrate with an 18-bp duplex (Fig. 1D and E), its repetitive DNA unwinding could not be clearly determined. To identify the minimal functional unit responsible for the repetitive unwinding, a long forked DNA, FK-50, was designed, comprising a 50-bp of duplex, a 3′ tail of oligo-dT₃₀, and a 5′ tail of oligo-dT₂₄. Cy3 was placed at the fork junction (D⁰) on the tracking strand and Cy5 (A) was placed 7 nt into the 5′ tail of the displaced strand [FK-50(D⁰A); Fig. 3A-i and Supplementary Table S1]. This structure alone exhibited E_FRET ~0.9 (Fig. 3A-i). Unwinding reaction was then initiated by adding each HIM-6 fragment in the presence of ATP-Mg. For all HIM-6 fragments, FRET levels were fluctuated between E_FRET 0.9 and 0.2 (Fig. 3A-ii–iv), but no significant disappearance of Cy5 spots was observed. These data suggest that the HIM-6 fragments did not fully unwind the forked DNA, but rather repeatedly performed partial unwinding events. FRET level distributions revealed a larger mid-FRET population (Fig. 3A-v). Taken together, these results indicate that the helicase domain and ZB element constitute the minimal functional unit required for the repetitive unwinding of dsDNA.

Figure 3.

Domains are responsible for repetitive DNA unwinding. (A) Schematic of forked DNA FK-50(D⁰A) substrate (left panel), composed of 50-bp duplex region with 3′ oligo-dT₃₀ tail and 5′ oligo-dT₂₄ tail. Cy3 (D) and Cy5 (A) fluorophores were labeled at ss/dsDNA junction. 5′ end of 5′-displaced strand was immobilized on slide surface. Indicated proteins (10 nM) are incubated for 5 min, unbound protein was washed away before introducing imaging buffer containing 100 μM ATP. Representative time traces display the fluorescence intensities of Cy3 (green) and Cy5 (magenta), along with the corresponding E_FRET for the following conditions: FK-50(D⁰A) only (i), FK-50(D⁰A) in the presence of HIM-6^Hlc_HRDC/ATP (100 μM) (ii), HIM-6^Hlc_WH/ATP (iii), and HIM-6^Hlc_ZB/ATP (iv). Data were acquired at 100 ms frame rate. (v) smFRET histograms compare FK-50(D⁰A) alone (n = 105 traces) (gray) with FK-50(D⁰A) in the presence of HIM-6^Hlc_ZB/ATP (n = 147) (green), HIM-6^Hlc_WH/ATP (n = 168) (cyan), and HIM-6^Hlc_HRDC/ATP (n = 147) (magenta). (B) Same experiment as in Fig. 3A, except with forked DNA FK-50(D¹⁸A) in which Cy3 fluorophore was labeled at 18th nucleotide on racking strand from junction and Cy5 fluorophore is labeled on complementary strand. Representative time traces display the fluorescence intensities of Cy3 and Cy5, along with the corresponding E_FRET for the following conditions: FK-50(D¹⁸A) alone (i), FK-50(D¹⁸A) in the presence of HIM-6^Hlc_HRDC and ATP (ii), HIM-6^Hlc_WH and ATP (iii), and HIM-6^Hlc_ZB and ATP (iv). Data were collected at 100 ms frame rate. (v) smFRET histograms compare FK-50(D¹⁸A) alone (n = 116 traces) (gray), FK-50(D18) in the presence of both HIM-6^Hlc_ZB and ATP (n = 166) (green), HIM-6^Hlc_WH and ATP (n = 122) (cyan), and HIM-6^Hlc_HRDC and ATP (n = 118) (magenta).

Since the HIM-6 fragments repetitively unwound the FK-50(D⁰A), we assessed the extent of repetitive unwinding by each HIM-6 fragment using two forked DNA substrates in which the Cy3/Cy5 pair was placed at different sites within the duplex region. In the FK-50(D¹⁸A) substrate, Cy3 was placed on the tracking strand 18 bp (D¹⁸) within the duplex region and Cy5 on the displaced strand 15 bp within the duplex region (Fig. 3B-i and Supplementary Table S1). In the FK-50(D²⁵A) substrate, Cy3 was placed on the tracking strand 25 bp (D²⁵) within the duplex region and Cy5 on the displace strand 22 bp within the duplex region (Supplementary Fig. S6-i and Supplementary Table S1) Repeated FRET level transitions were observed for the FK-50(D¹⁸A) with HIM-6^Hlc_HRDC and HIM-6^Hlc_WH (Fig. 3B-ii and iii), but not for HIM-6^Hlc_ZB (Fig. 3B-iv). FRET distributions again revealed higher repetitive unwinding activity of HIM-6^Hlc_HRDC and HIM-6^Hlc_WH (Fig. 3B-v). Contrarily, neither HIM-6^Hlc_HRDC nor HIM-6^Hlc_WH showed FRET transitions on the FK-50(D²⁵A) (Supplementary Fig. S6-i–iii). Accordingly, the FRET histogram for HIM-6^Hlc_HRDC and HIM-6^Hlc_WH showed a distribution similar to that of FK-50(D²⁵A) alone (Supplementary Fig. S6-vi). The sharp contrast between the FK-50(D¹⁸A) and FK-50(D²⁵A) indicates that both HIM-6^Hlc_HRDC and HIM-6^Hlc_WH unwind a narrowly defined length of duplex between 18 and 25 bp. The differing unwinding capacities of the three HIM-6 fragments suggest that the helicase domain along with ZB element is sufficient for repetitive DNA unwinding, while the WH domain enhances DNA unwinding probably by facilitating translocation and destabilizing base pairs.

Shuttling of HIM-6 along DNA during the repetitive unwinding process

Repetitive DNA unwinding by RecQ helicases, inferred from the separation and rejoining of a fluorophore pair on a DNA substrate indirectly suggests reversible movement of RecQ helicases, prompting investigation into the direct visualization of this movement. We monitored the shuttling of HIM-6^Hlc_HRDC along DNA using ⁵⁴⁹HIM-6^Hlc_HRDC and FK-50 labeled with Cy5 on the tracking strand 25 bp within the duplex region (FK-50^A25, Fig. 4A). The substrate equilibrated with the ⁵⁴⁹HIM-6^Hlc_HRDC in the absence of ATP-Mg showed a stable E_FRET ~0.1, reflecting the HIM-6 fragment bound at the fork junction (Fig. 4B). At varying ATP concentrations (5, 20, and 100 μM ATP), repeated fluctuations of E_FRET reaching 0.6 were observed, indicating that the shuttling of ⁵⁴⁹HIM-6^Hlc_HRDC along DNA is ATP-driven process (Fig. 4C). FRET traces were segmented into t₁ (the duration of unwinding and rewinding) and t₂ (the waiting time before the next unwinding). The histograms of t₁ were fitted with a gamma distribution, while the histograms of t₂ were fitted with a single-exponential decay (Fig. 4D). Both t₁ and t₂ decreased with increasing ATP concentrations. For ⁵⁴⁹HIM-6^Hlc_HRDC, t₁ was 3.2, 1.61, and 0.76 s for 5, 20, and 100 μM ATP, respectively, indicating that faster shuttling of the HIM-6 fragment requires greater ATP consumption. These data reveal that the repetitive DNA unwinding by HIM-6 is accompanied by its ATP-driven shuttling along DNA.

Figure 4.

HIM-6 shuttles along DNA while performing repetitive unwinding. (A) Schematic of smFRET experiment to monitor ATP-dependent translocation of ⁵⁴⁹HIM-6 construct on FK-50^A25 substrate. Cy5 (A) labeled at 25th nucleotide on tracking strand from junction. Fluorescently labeled protein ⁵⁴⁹HIM-6^Hlc_HRDC (1 nM) was introduced to chamber and incubated for 2 min. Unbound protein was washed away with imaging buffer. Reaction began by addition of imaging buffer (without ATP or with ATP). smFRET recording began immediately thereafter (low FRET). (B) Representative time traces showing fluorescence intensities of ⁵⁴⁹HIM-6 (green) and Cy5 (magenta), along with corresponding E_FRET: ⁵⁴⁹HIM-6^Hlc_HRDC in the absence of ATP. Time traces were acquired at a resolution of 100 ms. (C) Representative E_FRET time traces in varying ATP concentrations. Upper panel: In the presence of 100 μM ATP. Middle panel: In the presence of 20 μM ATP. Lower panel: In the presence of 5 μM ATP. Time t₁ indicates translocation time and time t₂ indicates waiting time. FRET traces were manually segmented for analysis. Traces acquired at100 ms resolution. (D) smFRET histograms of translocation time (t₁) and waiting time (t₂) for ⁵⁴⁹HIM-6^Hlc_HRDC on FK-50^A25 substrate: (i) in the presence of 100 μM ATP (n = 31 traces), (ii) in the presence of 20 μM ATP (n = 69), and (iii) in the presence of 5 μM ATP (n = 69). Histograms of t₁, fitted with Gamma distribution, are shown, while histograms of t₂ with single-exponential decay.

Intrinsic turning point derives HIM-6 transition between unwinding and pulling-in

To assess whether HIM-6 adopts different conformations in response to different DNA structures, we analyzed its behavior on a 5′-overhang partial duplex (5′-OH). Based on the 3′→5′ directionality of HIM-6, we predicted that HIM-6 could translocate along the 5′ tail from the ss/dsDNA junction toward its 5′ end. ⁵⁴⁹HIM-6^Hlc_HRDC was tested with two 5′-OH substrates, which commonly consists of a 24-bp duplex and a 40-nt 5′ tail. In one substrate, Cy5 was placed at the 3′ end of the short strand (5′-OH^A0, Fig. 5A). In the other, Cy5 was placed at the 5′ end of the 5′-overhang strand (5′-OH^A40, Fig. 5B). Upon initiating the reaction with each substrate and ⁵⁴⁹HIM-6^Hlc_HRDC, no FRET change was observed for the 5′-OH^A0 substrate. In contrast, periodic FRET transitions were observed for the 5′-OH^A40 substrate, indicating that HIM-6 fragment remained at the junction, while the 5′ end of the 5′ tail moved toward the fragment (Fig. 5A and B). These findings suggest that the HIM-6 fragment progressively pulls in the 5′ tail, forming an ssDNA loop behind it, and the loop subsequently released (Fig. 5B).

Figure 5.

HIM-6 pulls in 5′ tail of 5′-overhang partial duplex DNA. (A) Schematic of smFRET-based pulling in assay using fluorescently labeled HIM-6 fragment. Upper panel: 5′-OH^A0 DNA substrate, consisting of 24-bp duplex with 3′ end of short strand labeled with Cy5. 5′ end of short strand was immobilized on slide surface. ⁵⁴⁹HIM-6^Hlc_HRDC protein (1 nM) was added and incubated for 2 min. Unbound protein was washed away with imaging buffer. Reaction began by addition of imaging buffer (without ATP or with 100 μM ATP) and smFRET recording began immediately thereafter. Lower-left panel: Representative time traces show FRET efficiencies without ATP (no ATP) and with ATP. Lower-right panel: smFRET histograms, generated by averaging E_FRET values over 1-s interval (n > 5000 molecules), are shown and red lines indicates Gaussian fits to the FRET histograms. (B) Schematic of smFRET-based pulling in assay using fluorescently labeled HIM-6 fragment. Upper panel: 5′-OH^A40 DNA substrate, composed of 24-bp duplex in which 5′ end of 5′ overhang strand was labeled with Cy5 (A). 5′ end of short strand was immobilized on slide surface. ⁵⁴⁹HIM-6^Hlc_HRDC protein (1 nM) is added and incubated for 2 min. Unbound protein waswashed away with imaging buffer. Reaction began by addition of imaging buffer (without ATP or with 100 μM ATP) and smFRET recording began immediately thereafter. Lower-left panel: Representative time traces showing FRET efficiencies without ATP (no ATP) and with ATP (ATP). Lower-right panel: smFRET histograms, generated by averaging E_FRET values over 1-s internals (n > 5000 molecules) are shown and red lines indicate Gaussian fits to the FRET histograms. (C) Schematic of smFRET-based pulling in assay using a substrate labeled with two fluorophores. (i) A 5′-overhang partial duplex, 5′-OH(D40A), composed of 24-bp DNA duplex with 5′ oligo-dT₄₀ tail. Cy3 (D) was placed at 5′ end of 5′ overhang strand, while Cy5 (A) was placed at 3′ end of short strand. D40A indicates 40-nucleotide separation between donor and acceptor fluorophores. (ii) Representative time traces show fluorescence intensities of Cy3 (green) and Cy5 (magenta), along with corresponding E_FRET for 5′-OH(D40A) alone. (iii) Full-length HIM-6 protein (1 nM) was added and incubated for 2 min. Unbound protein waswashed away with imaging buffer. Reaction begins upon the addition of ATP and smFRET recording began immediately thereafter at 30 ms frame rate. Representative time trace showing fluorescence intensities of Cy3 (green) and Cy5 (magenta), along with corresponding E_FRET revealing repetitive pulling-in activity. (D) Effects of ATP concentrations on pulling-in activity of HIM-6^Hlc_HRDC. Left panel: FRET efficiencies of 5′-OH(D40A) in the presence of HIM-6^Hlc_HRDC and varying concentrations of ATP. Time interval (Δt) between successive two peaks was manually determined by visually picking. Right panels: Distributions of Δt were fitted to Gamma distribution. Distribution peaks at 0.63 s for 100 μM ATP (n = 93 traces), 1.16 s for 20 μM ATP (n = 81), and 1.39 s for 5 μM ATP (n = 93). (E) Effect of substrate tail length of substrate on pulling-in activity of HIM-6^Hlc_HRDC. (i) 5′-OH(D60A) with 5′ oligo-dT₆₀ tail (right panel). Representative time traces showing fluorescence intensity of 5′-OH(D60A) alone and corresponding E_FRET (left panel). (ii) HIM-6^Hlc_HRDC and 100 μM ATP. Time interval (Δt) between successive two peaks was manually determined by visually picking and distributions of Δt were fitted to Gamma distribution (right panel). Distribution peaks at 1.24 s for 100 μM ATP (n = 100 traces). (F, G) Comparison of pulling-in activity between HIM-6^Hlc_HRDC and HIM-6^Hlc_WH. (F) Post-synchronized FRET density plot showing distribution of E_FRET over time. Single-molecule traces were synchronized at onset of repetitive FRET fluctuations (0 s), which indicate pulling-in ssDNA strand. Upper panel: Post-synchronized FRET density plots of 5′-OH(D40A) substrate in the presence of HIM-6^Hlc_HRDC/ATP (n = 212). Lower panel: HIM-6^Hlc_WH/ATP (n = 90). FRET histograms, generated from the E_FRET values between 0 and 4 s, are shown and were fitted to multipeak Gaussian functions. Gaussian fits are shown as red, green, and dark blue lines corresponding stages (I), (II), and (III), respectively. Percentage of each stage were calculated based on area under curve of each Gaussian fit. (G) Same analysis as in panel (F), except with 5′-OH(D60A) substrate. Upper panel: Post-synchronized FRET density plots and histograms of 5′-OH(D60A) in the presence of HIM-6^Hlc_HRDC/ATP (n = 174 traces). Lower panel: HIM-6^Hlc_WH/ATP (n = 95).

We further investigated the ssDNA pulling activity using a 5′-OH substrate labeled with Cy3 at the 5′ end of the 5′ tail and Cy5 at the 3′ end of the short strand, resulting in a donor–acceptor spacer of 40 nt [5′-OH(D40A); Fig. 5C-i]. The substrate alone exhibited a stable E_FRET ~0.2 (Fig. 5C-ii). Prior to analyzing the pulling activity of HIM-6 fragments, we assessed the activity of the full-length HIM-6 protein. The full-length protein exhibited repeated pulling in ssDNA (Fig. 5C-iii), indicating that this activity represents an intrinsic feature of HIM-6. Upon initiating the reaction with the substrate and HIM-6^Hlc_HRDC at varying ATP concentrations (5, 20, and 100 μM ATP), a cycle of FRET levels between 0.8 and 0.2 was repeatedly observed, indicating that the 5′ tail was pulled toward the junction, and subsequently released (Fig. 5D, left panels). The period of each cycle was measured as the time interval (∆t) between two successive peaks and it showed a gamma distribution, whose average shifted to shorter time with increasing ATP concentration: 1.39, 1.16, and 0.63 s for 5, 20, and 100 μM ATP, respectively (Fig. 5D, right graphs). HIM-6^Hlc_ZB showed no ssDNA pulling activity. These results further support that HIM-6 bound at the junction actively extrudes the 5′ tail, forming an ssDNA loop.

The same measurement was performed on a 5′-OH substrate with a 60-nt tail to assess the effect of the tail length on the pulling time [5′-OH(D60A); Fig. 5E-i]. The substrate alone exhibited a stable E_FRET ~0.2 (Fig. 5E-i, left panel). At 100 μM ATP, the average time interval was 1.24 s, a two-fold increase compared to that with a 40-nt tail (Fig. 5E-ii). This supports that the pulling by HIM-6^Hlc_HRDC is progressive and length-dependent. Density plots of post-synchronized smFRET traces were used to compare the pulling activities of HIM-6^Hlc_WH and HIM-6^Hlc_HRDC fragments. With HIM-6^Hlc_HRDC, the pulled-in population at E_FRET ~0.8 population was 28% and 11%, for the 5′-OH(D40A) and 5′-OH(D60A), respectively (Fig. 5F and G). HIM-6^Hlc_WH also pulled in the 5′ tail, but exhibited a lower fraction (7%) on the 5′-OH(D40A) substrate and none on the 5′-OH(D60A) substrate. This suggests that HIM-6^Hlc_WH cannot fully pull the 5′ tail to the junction and likely releases it midway through the process, showing only mid-FRET (E_FRET ~0.5) (Supplementary Fig. S7). Collectively, these results suggest that the WH domain contributes to ssDNA pulling, as HIM-6^Hlc_ZB (lacking the WH domain) showed no detectable pulling activity, while the HRDC domain stabilizes the ssDNA loop until the 5′ tail is fully pulled to the junction.

Given that HIM-6^Hlc_HRDC repetitively unwinds forked DNA substrates, it is important to explore whether it pulls in the growing displaced strand before returning. We designed two forked DNA substrates, commonly consisting of a 24-bp duplex and a 3′ tail of oligo-dT₁₇ (Fig. 6A). FK-24^A40 contains a 5′ oligo-dT₄₀ tail and FK-24^A60 a 5′ oligo-dT₆₀ tail. Cy5 (A) was placed at the 5′ end of its 5′ tail (Fig. 6A). ⁵⁴⁹HIM-6^Hlc_HRDC was added to initiate the reaction. In the absence of ATP-Mg, a stable E_FRET ~0.1 was observed for both substrates (Fig. 6B and C), reflected in histograms of E_FRET. Upon the addition of ATP-Mg, repeated fluctuations of E_FRET were observed for both substrates; increase to ~0.8 followed by a subsequent drop (Fig. 6B and C), reflected in the histograms of E_FRET.

Figure 6.

HIM-6 pulls in the displaced ssDNA during DNA unwinding. (A) Schematic of smFRET-based pulling in assay with forked DNA. FK-24^A40 or FK-24^A60 DNA substrate, consisting of 24-bp duplex with 3′ oiligo-dT₁₇ tail and 5′ ssDNA tail labeled with Cy5 (A) at its 5′ end. A40 and A60 indicate 5′ oligo-dT tail lengths. 5′ end of complementary strand was immobilized on slide surface via biotin. Fluorescently labeled protein ⁵⁴⁹HIM-6^Hlc_HRDC (1 nM) were added and incubated for 2 min. Unbound protein was washed away with imaging buffer. Following the addition of imaging buffer containing 100 μM ATP, smFRET recording was initiated immediately at 30 ms resolution. (B) Representative time traces show FRET efficiencies of FK-24^A40 in the absence of ATP (upper trace) and in the presence of ATP (lower trace). Right panels: smFRET histograms, generated by averaging E_FRET values over 1-s intervals (n > 5000 molecules), are shown and were fitted to Gaussian distribution. (C) Representative time traces show FRET efficiencies of FK-24^A60 in the absence of ATP (upper trace) and in the presence of ATP (lower trace). Right panels: smFRET histograms generated by averaging E_FRET values over 1-s intervals (n > 5000 molecules) and fitted to Gaussian distribution.

Taken together, these findings suggest that HIM-6 exhibits intrinsic turning point, a transition between DNA unwinding and ssDNA pulling while remaining anchored at the leading fork junction, forming of an ssDNA loop behind it. During this process, the 5′ end can be released at the end and eventually, HIM-6 returns to its starting position and begins another cycle of unwinding and pulling.

HIM-6 returns without switching strands

Repetitive unwinding of HIM-6 has been characterized using duplex-containing DNA substrates; however, its activity on RNA-containing substrates has not yet been tested. We first examined the helicase activity of HIM-6^Hlc_HRDC on a DNA/DNA forked DNA [FK-18(D/D)], an RNA/RNA forked RNA [FK-18(R/R)], and a partial RNA/DNA duplex [3′-OH(R/D)] composed of a 3′-overhang RNA strand and a complementary DNA strand (Fig. 7A). All three substrates consisted a common 18-bp duplex region and were labeled with Cy3/Cy5 dyes. The number of Cy3 spots per imaging area was counted and the fractions of fully unwound DNA molecules were plotted as a function of time (Fig. 7B, left graph). No unwinding was observed for the FK-18(R/R) and FK-18(R/D) substrates, whereas HIM-6^Hlc_HRDC efficiently unwound FK-18(D/D), as evidenced by the decreases in Cy3 spot counts (Fig. 7B, right graph), indicating that HIM-6 cannot act on RNA-containing tracking strands.

Figure 7.

Effects of RNA strand-containing substrates on HIM-6 activities. (A) Nucleic acid substrates (left panel) for unwinding by HIM-6. FK-18(D/D), composed of 18-bp DNA duplex with 3′ oligo-dT₃₀ tail and 5′ oligo-dT₂₄ tail. Tracking strand (3′-tailed) was labeled with Cy3 (D). Complementary strand was labeled with Cy5 (A) and its 5′ end was immobilized on slide surface via biotin. FK-18(R/R), composed of 18-bp RNA duplex with both 3′ and 5′ oligo-rU₁₅ tails. Tracking strand was labeled with Cy5. Complementary strandwas labeled with Cy3 and its 5′ end was immobilized on slide surface via biotin. 3′-OH(R/D) composed of 3′ oligo-rU₁₅ tail with 18-bp RNA-DNA hybrid region. RNA strand was labeled with Cy5 and its 5′ end was immobilized on slide surface via biotin. The DNA complementary strand was labeled with Cy3. (B) Unwinding activity of HIM-6^Hlc_HRDC on substrates in (A). Left panel: Time course of DNA unwinding by HIM-6 fragments. HIM-6^Hlc_HRDC (10 nM) was incubated with substrates for 5 min and unbound protein was washed away before introducing imaging buffer containing 1 mM ATP. Fully unwound fractions of [FK-18(R/R) and 3′-OH(R/D)] substrates were plotted over time. Remaining Cy3 spot counts collected from 50–70 movies; data derived from two independent experiments. Right panel: Percentage of fully unwound substrates was calculated based fraction of disappeared Cy3 spots. Error bars represent standard error of mean (SEM). Data were derived from three independent experiments. (C) smFRET-based repetitive unwinding of 3′-OH(D/D) by HIM-6 fragments. (i) Schematic of 3′-OH(D/D) substrate (left panel), composed of 44-bp DNA duplex region with 3′ oligo-dT₃₀ tail. Cy3was labeled at 7th nucleotide from ss/dsDNA junction toward 3′ end and Cy5 is at junction. Indicated proteins (10 nM) were incubated for 5 min, unbound protein was washed away before introducing imaging buffer containing 1 mM ATP. Representative time traces show E_FRET of 3′-OH(D/D) alone (right panel). (ii) In the presence of HIM-6^Hlc_HRDC (left panel). smFRET recording began immediately thereafter. Representative time traces show fluorescence intensities of Cy3 and Cy5, along with corresponding E_FRET (right panel). (iii) Same experiment as (ii) except with HIM-6^Hlc_WH. Representative time traces show fluorescence intensities of Cy3 and Cy5, along with corresponding E_FRET (right panel). Data collected at 100 ms frame rate. (vi) smFRET histograms were generated from >5000 traces under conditions (i–iii): 3′-OH(D/D) alone (gray), HIM-6^Hlc_WH and ATP (cyan), and HIM-6^Hlc_HRDC and ATP (magenta). (D) Same experiment as in panel (B), except for substrate. (i) 3′-OH(D/R-D) chimeric substrate (left panel), composed of 3′ overhang DNA strand annealed to complementary chimeric strand containing 21-nt RNA and 23-nt DNA. Cy3 was labeled at 7th nucleotide from ss/dsDNA junction toward 3′ end and Cy5 is at 5′ end of chimeric strand. Representative time traces show E_FRET of 3′-OH(D/R-D) alone (right panel). (ii) In the presence of HIM-6^Hlc_HRDC (left panel). smFRET recording began immediately thereafter. Representative time traces show fluorescence intensities of Cy3 and Cy5, along with corresponding E_FRET (right panel). (iii) Same experiment as (ii) except with HIM-6^Hlc_WH. Representative time traces show fluorescence intensities of Cy3 and Cy5, along with corresponding E_FRET (right panel). Data collected at 100 ms frame rate. (vi) smFRET histograms generated from >5000 traces under conditions (i–iii): 3′-OH(D/R-D) alone (gray), HIM-6^Hlc_WH and ATP (blue), and HIM-6^Hlc_HRDC and ATP (magenta). (E) Pulling-in activity. Left panel: Schematic of 5′-OH(D30A) substrate, containing 24-bp DNA duplex region and 5′ 30-nt RNA tail. Cy3 (green) and Cy5 (magenta) dyes are placed as indicated. Right panel: Representative FRET time trace of 5′-OH(D30A) in the presence of HIM-6 WT and ATP showing fluorescence intensities of Cy3 (green) and Cy5 (magenta) and corresponding E_FRET.

Since HIM-6 does not efficiently substrates with an RNA-containing tracking strand, RNA in the displaced strand would block its backward movement if HIM-6 switches and translocating along it during rewinding. Second, the behavior was assessed using two 3′-overhang partial duplex substrates: one with both strands composed of DNA [3′-OH(D/D)] and the other with the chimeric short strand consisting of a 21-nt RNA segment and a 23-nt DNA segment [3′-OH(D/R-D)] (Fig. 7C-i and D-i). The pair of Cy3 and Cy5 was labeled at the indicated positon and both substrates yielded a stable E_FRET ~0.9. Contrary to the expectation, both HIM-6^Hlc_HRDC and HIM-6^Hlc_WH clearly exhibited DNA unwinding and rewinding on both substrates in the same manner (Fig. 7B-ii–iii and C-ii–iii), as evidenced by repeated FRET transitions and the corresponding FRET populations (Fig. 7B-iv and C-iv). These data suggest that the HIM-6 fragment moves back along the tracking strand without switching strands using a built-in capacity for reversible action.

Finally, we assessed ssRNA pulling of full-length HIM-6 on a 5′-OH(D30A) substrate containing a 24-bp DNA duplex region and a 30-nt 5′ RNA tail (Fig. 7E). No FRET change was observed for the substrate, indicating that HIM-6 cannot pull in a 5′ ssRNA tail.

A single ribonucleotide impedes DNA unwinding by HIM-6 and serve as an activity switch

Given that HIM-6 was unable to unwind substrates with RNA tracking strands, we investigated whether the presence of a single ribonucleotide on the strand affects its enzymatic activities. On a forked substrate with an 18-bp duplex region, a uridine was introduced at the 8th position from the junction on the tracking strand and Cy5 was placed at the 5′ end of the same strand [FK·(rU)18^A18; Fig. 8A]. The substrate yielded a stable low FRET level in the presence of ⁵⁴⁹HIM-6^Hlc_HRDC only (Fig. 8B). Upon the addition of ATP-Mg, repeated FRET transitions were interestingly observed (Fig. 8B) and ⁵⁴⁹HIM-6^Hlc_HRDC poorly unwound FK·(rU)18^A18 compared to FK-18 ^A18 (Fig. 8C). These data suggest that ⁵⁴⁹HIM-6^Hlc_HRDC is unable to fully unwind the substrate as a single ribonucleotide impedes HIM-6′s translocation, but instead ⁵⁴⁹HIM-6^Hlc_HRDC shuttle along the DNA. It is likely that a single ribonucleotide mimics a turning point intrinsic to the HIM-6, which exhibits transition between DNA unwinding and ssDNA pulling.

Figure 8.

Single ribonucleotide induces a shift of ssDNA pulling. (A) Schematic of FK·(rU)18^A18 unwinding by ⁵⁴⁹HIM-6^Hlc_HRDC. FK·(rU)18^A18, composed of 18-bp duplex with 3′ oligo-dT₃₀ tail and 5′ oligo-dT₂₄ tail. Cy5 (A) was attached to 5′ end of 3′-tailed strand, while 3′ end of complementary strand was immobilized on slide surface. Single uridine (rU) was incorporated at 8th nucleotide from junction on tracking strand. Cy5 disappearance was used to quantify DNA unwinding. FK-18^A18(2), composed of 18-bp duplex with 3′ oligo-dT₃₀ tail and 5′ oligo-dT₂₄ tail. Cy5 dye was attached to 5′ end of tracking strand, while 3′ end of complementary strand was immobilized on slide surface. (B) Representative time traces showing fluorescence intensities of ⁵⁴⁹HIM-6^Hlc_HRDC and Cy5, along with corresponding E_FRET. ⁵⁴⁹HIM-6^Hlc_HRDC (10 nM) was incubated for 5 min, unbound protein was washed away before introducing imaging buffer containing 1 mM ATP. (C) Percentage of fully unwound substrates over time based on disappearance of Cy5 spots. FK-18^A18(2) by ⁵⁴⁹HIM-6^Hlc_HRDC/ATP (magenta); FK·(rU)18^A18 by ⁵⁴⁹HIM-6^Hlc_HRDC/ATP (cyan); FK·(rU)18^A18 by ⁵⁴⁹HIM-6^Hlc_HRDC alone. Fraction of disappeared Cy5 spots was quantified from over 50 movies; data derived from three independent experiments. (D) Schematic of FK·(rU)24^A40: 24-bp duplex with 3′ oligo-dT₁₇ tail and 5′ oligo-dT₄₀ tail. Cy5 was attached to 5′ end of 5′-tailed strand, while 3′ end of complementary strand was immobilized on slide surface. Single uridine (rU) incorporated at 2nd nucleotide from junction on tracking strand (left diagram). (E) Representative time traces showing fluorescence intensities of ⁵⁴⁹HIM-6^Hlc_HRDC and Cy5, along with corresponding E_FRET. (F) Post-synchronized FRET density plots of FK·(rU)24^A40 substrate in the presence of ⁵⁴⁹HIM-6^Hlc_HRDC/ATP (n = 151). Single-molecule traces were synchronized at onset of repetitive FRET fluctuations (0 s), which indicate pulling-in ssDNA strand. FRET histograms generated from E_FRET values between 0 and 4 s and fitted to multipeak Gaussian functions.

We investigated whether HIM-6 is capable of pulling in the growing displaced strand when its unwinding is impeded by a single ribonucleotide. For this purpose, we employed a forked substrate with a 24-bp duplex containing a uridine at the 2nd base from the junction on the tracking strand, a tail of 3′ oligo-dT₁₇, and a 5′ tail of oligo-dT₄₀. Cy5 was labeled at the 5′ tail end of the complementary strand [FK·(rU)24^A40; Fig. 8D]. A40 denotes a 40-nucleotide spacer between the junction and the Cy5. The reaction with ⁵⁴⁹HIM-6^Hlc_HRDC in the presence of ATP-Mg exhibited repeated FRET transitions, indicating repeatedly pulling in the 5′ tail (Fig. 8E). FRET histograms and overlaid density plots of the FRET traces revealed populations with E_FRET greater than 0.4 (Fig. 8F). These results indicate that HIM-6 paused at an intrinsic block can repeatedly pull in the 5′ tail.

Discussion

Our study using various HIM-6 fragments identified that the helicase domain and ZB element together function as the minimal functional unit that drives repetitive, partial DNA unwinding. The larger HIM-6 fragment containing the WH domain displayed enhanced DNA unwinding activity and conferred a new ssDNA pulling activity. We found that HIM-6 preferentially bind to the ss/dsDNA junction in a monomeric form to initiate unwinding. Once bound to the junction, HIM-6 appears to unwind only short stretches of DNA and to pull in the 5′-displaced strand at the leading fork, forming ssDNA loop behind the protein before rewinding the DNA. Upon releasing the ssDNA loop, HIM-6 slides back along the tracking strand without switching strands, permitting DNA rewinding. It then resumes unwinding at the fork junction. Using fluorescently labeled HIM-6 fragments, we directly visualized HIM-6 shuttling along DNA accompanied by its repetitive unwinding. Notably, we identified, for the first time, that when HIM-6 encounters a single ribonucleotide on the tracking strand, HIM-6 appears to switch to the pulling mode.

A minimal functional unit required for repetitive unwinding

Our findings suggest that HIM-6^Hlc_ZB is the minimal functional unit sufficient for DNA strand separation. HIM-6^Hlc_ZB was also capable of repetitively unwinding short stretches of DNA. This fragment may transiently lose its grip on the DNA during unwinding, possibly due to reduced stability, resulting in a shorter unwinding length compared to the larger fragment and subsequent sliding back along the DNA. Still, the fragment can grab the tracking strand during backsliding and reinitiate unwinding the fork. Consistent with this observation, truncated constructs of human WRN and C. elegans WRN-1, which comprises both the Hlc domain and ZB element, exhibited similar repetitive unwinding behaviors [10]. Furthermore, AtRecQ3 from plant A. thaliana, a plant RecQ orthologue that naturally lacks the WH and HRDC domains but retains the Hlc domain and ZB element, demonstrated repetitive unwinding [11].

The repetitive unwinding of fork substrate by HIM-6 implies a turning point where protein makes a structural transition and reverses its motion. Such behavior may reflect an intrinsic conformational cycle of the protein. Crystal structures of several RecQ helicases have shown that the helicase subdomains are involved in both ATP binding and ssDNA binding, while the ZB element helps position the DNA correctly for the Hlc domain to work on [30–37]. Within the Hlc domain, conformational changes and power stroke motions of RecA-like domain 1 and RecA-like domain 2 drive a unidirectional movement of the helicase along an ssDNA [32, 38]. Although the detailed crystal structure of HIM-6 is not yet available, we generated a predicted structure using AlphaFold3, which were allowed us to align its Hlc domain and ZB element with those of structurally characterized RecQ helicases (Supplementary Fig. S8). This analysis revealed that the Hlc domain and ZB element adopt a similar configuration to those observed in other RecQ helicases. Conformational changes within the Hlc domain in HIM-6^Hlc_ZB likely direct alternating modes of unwinding and backsliding.

HIM-6^Hlc_ZB unwinds only short DNA segments (<18 base pairs) due to a low efficiency of unwinding, whereas inclusion of the WH domain in larger fragments enhances the activity, potentially by stabilizing the separated DNA strands. In the absence of this domain, HIM-6^Hlc_ZB weakly binds to DNA and may be prone to slipping off the DNA, thereby limiting its processivity. By comparison, previous studies have shown that an isolated construct containing the helicase domain along with a Zn-binding element from WRN and BLM are capable of unwinding DNA [39, 40]. However, comparable measurements of these constructs have not yet been reported.

Functional role of the WH domain in DNA unwinding and ssDNA pulling

The larger fragment containing the WH domain enhanced repetitive DNA unwinding. Both HIM-6^Hlc_WH and HIM-6^Hlc_HRDC containing the WH domain were able to repeatedly unwind DNA longer than 18 bp, but shorter than 25 bp, whereas HIM-6^Hic_ZB reversed before unwinding 18 bp. The WH domain likely destabilizes duplex DNA by direct interacting with DNA near the fork junction and inducing local structural distortion. This mechanism is supported by crystal structures of other RecQ helicases, which demonstrated the interaction between the WH domain and either a blunt-ended DNA duplex or the duplex region of tailed DNA substrate [33–35, 41, 42]. Within this domain, a specific motif was found to contribute to DNA strand separation. Structural and biochemical studies on RecQ1, BLM and WRN have demonstrated that a prominent “β-hairpin” present in the WH domain is essential for DNA unwinding (Supplementary Fig. S9A and B) [43]. Notably, the β-hairpin protrudes from the WH domain and is located at the ss/dsDNA junction, effectively destabilizing the terminal base pair(s). This helps begin strand separation by functioning as a wedge [41, 44–46]. While the exact DNA-binding interfaces in HIM-6 remain unknown, AlphaFold3 prediction of HIM-6 structure indicates the presence of a β-hairpin structure (amino acids 738–744, VPNQAAA) within the WH domain that occupies a position similar to the well-characterized β-hairpin found in other RecQ helicases (Supplementary Fig. S9C). To clarify the functional role of the β-hairpin in strand separation, mutational studies of the key residues within the β-hairpin are necessary.

Additionally, we found that HIM-6 fragments containing the WH and HRDC domains repetitively pull in the 5′ tail of a 5′-OH substrate in an ATP-dependent manner. This activity may represent an intrinsic feature of HIM-6, as the full-length protein exhibits a similar ability to repeatedly pull in ssDNA. During this process, the ssDNA looped out behind the protein as 5′ tail end is closely located to HIM-6. This repetitive pulling-in activity depends on stable anchoring at the duplex junction, likely mediated by a subdomain within the helicase. The WH domain appears to be critical for this function as HIM-6^Hlc_ZB lacking the WH domain, failed to pull in the ssDNA tail to the end. Since ssDNA pulling-in was initially reported for PcrA helicase [47], several members of the RecQ family, including human BLM, G. gallus WRN, and C. elegans WRN-1, have also shown the same pulling-in activity [8, 10, 48]. However, the domains involved in pulling ssDNA have not been reported for these RecQ helicases.

We found that both HIM-6^Hlc_WH and HIM-6^Hlc_HRDC preferentially bind to the ss/dsDNA junction in a monomeric form to initiate unwinding and perform unwinding and pulling in a monomeric form. A similar oligomeric state of BLM has been reported. A previous EM and AFM analyses performed at markedly lower protein concentrations of BLM^642–1290 demonstrated that BLM is predominantly monomeric in both the absence and presence of various DNA structures, including ssDNA, a Holliday junction and immobile D-loops. Consistent with this observation, active translocation along ssDNA was demonstrated to be mediated by a monomeric form [49]. More recently, photobleaching step measurements revealed that the majority of the mutant of GFP-BLM^641-1417 bound to DNA ends exists as monomers [50]. Notably, however, these studies also reported a propensity for multiple BLM monomers to interact with one another when bound to specific DNA substrates, such as DNA ends or a DNA-loop with a 5′ invading end.

RecQ shuttling along DNA during repetitive DNA unwinding

We observed, for the first time, the shuttling motion of RecQ helicase during repetitive DNA unwinding using fluorescently labeled HIM-6 fragments tagged with SNAP. The SNAP-tag is widely used in imaging applications under a wide range of experimental conditions [51]. Although the SNAP tag is located at the N-terminus of the fragment, the tag may still introduce artifacts due to its size. However, fluorescent labeling and SNAP tagging did not noticeably interfere with the helicase activity of HIM-6 fragments in this study. HIM-6 WT (full-length HIM-6) and HIM-6(205–788; a construct spanning the helicase domain to the WH domain), without the SNAP-tag, showed unwinding extents similar to those of the SNAP-tag containing proteins HIM-6^Hlc_HRDC and HIM-6^Hlc_WH (Supplementary Fig. S10). Fluorescent labeling did not affect the unwinding activity of HIM-6^Hlc_HRDC, HIM-6^Hlc_WH, and HIM-6^Hlc_ZB shown in Fig. 1D and G. This is consistent with previous findings show that SNAP-tagged proteins such as the replication helicase Mcm2–7 and the primosomal protein PriA retain their native activity following labeling [26, 28].

The movement of HIM-6 was visualized using two different approaches. The first approach tracks the forward movement from the fork junction to the duplex end. A similar forward translocation pattern has been reported for fluorescently labeled UvrD, a member of SF1 helicase family, during the unwinding forked DNA substrates [52]. The second one tracks the shuttling from the fork junction to an internal point (Fig. 3). A single HIM-6 fragment was capable of traveling approximately 209.4 Å (a round trip, Supplementary Fig. S4C and D) within 0.76 s in the presence of 100 μM ATP (Fig. 4C). This shuttling was accompanied by repetitive DNA unwinding. We confirmed that HIM-6 actually passes through the internal nucleotides during unwinding. Repeated FRET transitions were observed for a forked DNA substrate where Cy5 was placed at base pair 8 (Supplementary Fig. S11A and B).

Interestingly, while previous smFRET studies demonstrated that both WRN and BLM are capable of repetitive unwinding, direct imaging using GFP-tagged versions of these proteins revealed only forward movement [17, 53]. Their shuttling behavior was not apparent in those studies. The repeated shuttling motions occurred on a short length scale of only several dozens of base pairs, making them difficult to detect with GFP-tagged proteins, which are better suited for long-range tracking. The SNAP-tagged design of HIM-6 fragments used in this study provides a complementary approach for the study of RecQ helicases.

A turning point in HIM-6 function: from DNA unwinding to ssDNA pulling

Our data from the fluorescently labeled HIM-6 fragments suggest that when HIM-6 reaches its unwinding limit, it stays anchored at the leading fork junction and starts pulling in the displaced strand, forming an ssDNA loop. Once that loop is released, HIM-6 returns to its initial position while the unwound strands re-anneal. A comparable ssDNA pulling-in activity on a forked DNA substrate was reported for human BLM, in which the 5′ end of the displaced strand was immobilized on the surface, making it uncertain whether the 5′ end could reach BLM [48]. As the 5′-OH substrate supports pulling activity rather than unwinding activity, the time interval between successive pulling-in events would be shorter than that observed with a forked DNA substrate. The average time interval was 0.63 s on the 5′-OH substrate and 2.55 s on a forked DNA substrate (Supplementary Fig. S12). This prolonged interval likely reflects the additional time required for backsliding and unwinding. Taken together, HIM-6 fragment pulls in the displaced strand following the unwinding of duplex DNA.

It is possible that there are in vivo factors that impede DNA unwinding. Ribonucleotides are frequently incorporated into DNA during by DNA polymerases [54, 55]. While previous studies of other RecQ helicases have demonstrated the effect of RNA strand on their unwinding activity, the impact of RNA strands on pulling activity or the effect of a single ribonucleotide on helicase activity has remained largely unexplored. We newly discovered that HIM-6 was unable to pull in RNA-containing nontracking strand and that a single ribonucleotide significantly impeded HIM-6-mediated DNA unwinding. This suggests that ribonucleotides can act as a physical barrier for HIM-6 helicase activity. Interestingly, when HIM-6 was paused at the ribonucleotide, it continued to engage the DNA by pulling in the displaced strand rather than dissociating from the substrate, suggesting that a ribonucleotide can trigger a turning point for HIM-6, which may promote the recruitment of ribonucleotide-processing enzymes (e.g. RNase H) that help resolve RNA-DNA structures. Through this transition, HIM-6 maintains a stable interaction with DNA even under mechanical hindrance and undergoes a conformation change to enable an altered mode of activity. Once pulled ssDNA, HIM-6 slides back on the tracking strand without switching strands. This backsliding mode closely resembles that observed in helicases such as human WRN, S. cerevisiae Pif1, and S. pombe Pfh1, which exhibit repetitive unwinding activity on DNA/RNA hybrid duplexes [17, 18].

Role of iterative activities of HIM-6

HIM-6 helicase on its own displays relatively low intrinsic processivity—it typically unwinds only short stretches of DNA before dissociating. While this limited unwinding activity may act as a regulatory mechanism to avoid excessive or unnecessary strand separation, it would generally render the process inefficient due to frequent dissociation events. However, HIM-6 seems to overcome this limitation by switching its DNA-binding mode. Upon stalling of DNA unwinding, HIM-6 undergoes conformational changes and pulls in the ssDNA at the leading fork junction, helping stabilize its association with the DNA and stabilize the unwound region by restricting the diffusion of the ssDNA. As a result, an ssDNA loop is formed behind the helicase. The looped ssDNA near the ss/ds junction likely increases the local concentration of complementary sequences, thereby enhancing the chances of rezipping between two ssDNA strands. While HIM-6 slips and backslides to the initial fork, DNA is rewound. HIM-6 again re-engages in DNA unwinding. Taken together, our results suggest that iterative activities of HIM-6; cycling through DNA unwinding, ssDNA pulling, and backsliding may serve as a dynamic safeguard, helping HIM-6 preserve DNA accessibility and prevent genome instability by remaining stably bound to the DNA.

Supplementary data

Supplementary data is available at NAR online.

Conflict of interest

None declared.

Funding

This work was supported by National Research Foundation of Korea [RS-2022-NR0757641 to B.A.]; National Research Foundation of Korea [RS-2024-00342990 to H.K.]; 2025 Research Fund of UNIST [1.250006.01 to H.K.]. Funding to pay the Open Access publication charges for this article was partly provided by 2025 Research Fund of UNIST.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author