Structural mechanism underpinning Thermus oshimai Pif1‐mediated G‐quadruplex unfolding

Abstract

G‐quadruplexes (G4s) are unusual stable DNA structures that cause genomic instability. To overcome the potential barriers formed by G4s, cells have evolved different families of proteins that unfold G4s. Pif1 is a DNA helicase from superfamily 1 (SF1) conserved from bacteria to humans with high G4‐unwinding activity. Here, we present the first X‐ray crystal structure of the Thermus oshimai Pif1 (ToPif1) complexed with a G4. Our structure reveals that ToPif1 recognizes the entire native G4 via a cluster of amino acids at domains 1B/2B which constitute a G4‐Recognizing Surface (GRS). The overall structure of the G4 maintains its three‐layered propeller‐type G4 topology, without significant reorganization of G‐tetrads upon protein binding. The three G‐tetrads in G4 are recognized by GRS residues mainly through electrostatic, ionic interactions, and hydrogen bonds formed between the GRS residues and the ribose‐phosphate backbone. Compared with previously solved structures of SF2 helicases in complex with G4, our structure reveals how helicases from distinct superfamilies adopt different strategies for recognizing and unfolding G4s.

The crystal structure of ToPif1 complexed with a G4 shows that ToPif1 recognizes the entire native G4 via a G4‐Recognizing Surface. It reveals how helicases from distinct superfamilies use different strategies for recognizing and unfolding G4s.

Introduction

G‐quadruplexes (G4s) are higher order DNA/RNA structures formed from nucleic acid sequences that contain four stretches of three or more guanines interspaced by at least one random nucleotide (Burge et al, ²⁰⁰⁶; Castillo Bosch et al, ²⁰¹⁴). The G‐tetrad module, composed of four guanines, forms a planar structure stabilized by non‐canonical Hoogsteen hydrogen bonds in the presence of monovalent cations, such as sodium or potassium (Castillo Bosch et al, ²⁰¹⁴; Griffin & Bass, ²⁰¹⁸). Genome‐wide bioinformatics analyses have revealed the prevalence of G4 motifs in key regulatory regions of the human genome, such as promoters, gene bodies, and untranslated regions (Huppert & Balasubramanian, 2005; Todd et al, ²⁰⁰⁵; Eddy & Maizels, ^{2006, 2008}; Varshney et al, ²⁰²⁰). Following this prediction, a tremendous increase in cell experimental data has demonstrated that G4 structures indeed exist under physiological conditions and are involved in a variety of biological processes, including telomere maintenance, gene expression, epigenetic regulation, and DNA replication (Schierer & Henderson, 1994; Maizels & Gray, ²⁰¹³). The high thermodynamic stability of G4s in cells can erect significant barriers to replication fork progression. Therefore, failing to resolve the folded G4s during DNA transactions can cause replication to stall and thus may trigger the rampant genomic instability characteristic of certain cancer cells (De & Michor, 2011; Edwards et al, ²⁰¹⁴; Eddy et al, ²⁰¹⁵; Fontana & Gahlon, ²⁰²⁰). Cells have evolved the capability of resolving stable G4s with a group of helicases that play important roles in the maintenance of genomic stability in many organisms (Mendoza et al, ²⁰¹⁶). Pif1 family helicases, belonging to the superfamily 1 (SF1), are conserved across evolution from bacteria to humans and have been shown to unfold G4, being critical for DNA replication, telomere maintenance, and genome stability (Li et al, ²⁰¹⁶; Dahan et al, ²⁰¹⁸; Wang et al, ²⁰¹⁸). For instance, a recent study has demonstrated that Pif1 helicase promotes break‐induced replication in mammalian cells (Li et al, ²⁰²¹). DNA G4‐resolving helicases are also found in superfamily 2 (SF2) in groups such as the XPD family (FANCJ and RTEL1) and the RecQ family of helicases (BLM, WRN, and RecQ). FANCJ bearing G4‐resolving activity is necessary for DNA replication and has shown a unique substrate specificity for G4 architecture (Bharti et al, ²⁰¹³; Lee et al, ²⁰²¹). RecQ helicases have been found to efficiently unfold a variety of G4 DNA structures and play essential roles in DNA repair and telomere maintenance (Singh et al, ²⁰¹²). Furthermore, mutations in FANCJ, BLM, or WRN are associated with human diseases (Levitus et al, ²⁰⁰⁵; Budhathoki et al, ²⁰¹⁴; Zimmer et al, ²⁰²⁰). Though G4 resolvases have attracted much attention, few solved co‐crystal structures in which helicase complexed with G4 are available. The X‐ray crystal structure of DHX36 in complex with a Myc‐promoter‐derived G4 was firstly reported and shows that G4 is captured by a specific DHX36‐specific motif (DSM) which bears a signature peptide AKKQ found in FANCJ (Wu & Spies, 2016). Based on DHX36‐G4 crystal structures, it was suggested that DSM, in coordination with the helicase core, unfolds the G4 one base at a time in an ATP‐hydrolysis‐independent manner. Recently, a crystal structure (Voter et al, ²⁰¹⁸) solved from a helicase core of a RecQ family helicase (CsRecQ) complexed with oligonucleotides harboring a human telomere sequence revealed that a guanine base within the putative G‐tetrad is located at the 5′‐gate of the electropositive channel, where the G‐base is flipped and sequestrated in a guanine‐specific pocket (GSP) (Voter et al, ²⁰¹⁸). This structure is interpreted to represent a post‐catalytic complex of CsRecQ bound to unfolded G4. Therefore, SF2 helicases may use a DSM/AKKQ motif or GSP site for G4 binding/unfolding, but these motifs are not found in Pif1.

Though crystal structures of Pif1 helicases from Homo sapiens, Saccharomyces cerevisiae, Bacteroides sp., and Thermus oshimai have been recently solved with different DNA substrates including a ssDNA or a DNA substrate that mimics a stalled replication fork, little is known about how Pif1 helicases interact with G4 (Zhou et al, ²⁰¹⁶; Lu et al, ²⁰¹⁸; Dehghani‐Tafti et al, ²⁰¹⁹; Dai et al, ²⁰²¹). Elucidating the molecular mechanism by which Pif1 helicase recognizes and unfolds G4 molecules largely depends on the knowledge of crystal structures. Access to structures that represent a “substrate” complex, intermediates, and a “product” complex provides very valuable mechanistic information. Although the helicase‐G4 DNA complex structures of DHX36 (Chen et al, ²⁰¹⁸) and CsRecQ (Voter et al, ²⁰¹⁸) can be considered as representations of an intermediate substrate complex (Chen et al, ²⁰¹⁸) and a final product complex of helicases (Voter et al, ²⁰¹⁸), respectively, the remaining challenge is to obtain the structure in which the helicases interact with an intact G4 (initial substrate complex) or to capture the structure of the very first steps leading to the transition state.

In this study, we report the X‐ray crystal structures solved from the Pif1‐family helicase from T. oshimai (ToPif1) bound to a prefolded G4 bearing 6–8 nt on both sides of the G‐tetrads, thereby mimicking the physiological G4 state formed during DNA replication. This is the first structural snapshot of a helicase complexed with an integral G4. In combination with bulk and single‐molecular fluorescent resonance energy transfer assays (smFRET), our studies provide new mechanistic insight into how a Pif1‐family helicase recognizes, binds, and unfolds G4s.

Results

ToPif1 unfolds G‐quadruplexes in an ATP‐dependent manner without topological preference

Parallel and antiparallel intramolecular G4s, 12 nt‐ssDNA, and 24 bp blunt‐end DNA were titrated with increasing concentrations of ToPif1 under equilibrium conditions. The apparent dissociation constant (K _d,app) determined from the titration curves (Fig 1A and Table 1) demonstrated that ToPif1 binds both antiparallel (G4^Tel) and parallel G4s (G4^CEB) (Fig EV1A) with similar affinity (K _d,app ^antipara ≈ 29.63 nM (G4^Tel); K _d,app ^para ≈ 23.27 nM (G4^CEB)) without any topological preference. Furthermore, although both the G4 and ssDNA (T₁₂) were bound with essentially the same affinities (K _d,app ^antipara ≈ 29.63 nM versus K _d,app ^ssDNA ≈ 29.12 nM) (Fig 1A (insert) and Table 1), blunt‐end DNA bounds ToPif1 very weakly (K _d,app > 250.00 nM; Fig 1A (insert) and Table 1). The fact that the apparent dissociation constant of G4 and ssDNA is nearly identical suggests that ToPif1 possesses two independent binding sites for ssDNA and G4, respectively. Indeed, competitive binding assays between the labeled G4 and unlabeled ssDNA demonstrated that labeled G4 binding was not influenced by increasing unlabeled ssDNA and vice versa (Fig EV1B). We then investigated ToPif1‐mediated G4‐unfolding activities by stopped‐flow assay according to the previous report (Liu et al, ²⁰²⁰) with both parallel and antiparallel G4 substrates (Cheng et al, ²⁰¹⁸) in the absence or the presence of ATP. The kinetic data were fit to equation (2) to determine the unfolding amplitude (A _m) and unfolding rate (k _cat) measured with different G4 DNAs (Fig 1B and Table 1). Although unfolding activity is undetectable in the absence of ATP or the presence of the non‐hydrolyzable ATP analog (Fig EV1C), the values of unfolding amplitudes (A _m) and unwinding rates (k_cat (s⁻¹)) determined with antiparallel G4s were 1.31‐fold and 1.39‐fold higher than those determined with parallel G4s, respectively (Fig 1B and Table 1). The calculated P‐value (0.001) from the above data is less than 0.01, indicating that the observed differences between parallel and anti‐parallel are statistically significant.

Figure 1. Binding and unwinding activities of ToPif1 for different configurations of antiparallel and parallel G4s.

• A
  Comparison of the binding activity of ToPif1 for different configurations of antiparallel (G4^Tel, black points) and parallel (G4^CEB, red points) G4s. (Insert: Comparison of the binding activity of ToPif1 to single‐stranded DNA (T₁₂, black points) with G4^Tel, red points), the blue points represent the double‐strand DNA (D₂₄)). The binding activity was measured using steady‐state fluorescence anisotropy assays; 5 nM of fluorescein‐labeled DNA substrate and 0.5 mM ADP·AlF₄ were titrated with increasing protein concentrations at 37 °C. The different binding curves represent an average of three to four independent experiments for each substrate with standard deviations. The solid line represents the fit of data according to equation (1).
• B
  Stopped‐flow DNA unwinding kinetics of ToPif1 for different configurations of G4 DNA (G4^AP‐424, 4G4^AP‐333, G4^P‐214) under multiple turnover conditions. All curves represent the average of at least 10 individual traces and the plots are representative of three independent experiments. 4 nM G4 DNA and 100 nM ToPif1 were used under experimental conditions as described in ‘Materials and Method’.
• C, D
  The typical smFRET trajectories of antiparallel (smG4^AP‐424, sm4G4^AP‐333) and parallel (smG4^P‐111) G4 unwinding catalyzed by 50 nM ToPif1 and 150 μM ATP. Black arrows in the solid style indicate the time of adding proteins. Dashed arrows in black represent the unwinding time of ToPif1 with different G4 DNA substrates.

Table 1.

Binding and unwinding activities of ToPif1 for different DNA substrates ¹ .

Substrate	Binding 2	Unwinding 3
	K d,app (nM)	A m	kcat (s−1)
Parallel G4
G4CEB	23.27 ± 3.12	ND 4	ND
G4P‐214	ND	0.55 ± 0.02	1.57 ± 0.07
G4P‐241	ND	0.35 ± 0.02	1.23 ± 0.05
G4P‐124	ND	0.22 ± 0.01	1.17 ± 0.04
Antiparallel G4
G4Tel	29.63 ± 2.64	ND	ND
G4AP‐424	ND	0.72 ± 0.06	2.18 ± 0.02
4G4AP‐333	ND	0.70 ± 0.03	0.57 ± 0.01
ssDNA
T12	29.12 ± 4.19	－	－
dsDNA
S26D17	ND	0.59 ± 0.02	3.53 ± 0.02
Blunt‐end DNA
D24	250~280	－	－

¹ Data were determined from 2–3 independent experiments under the experimental conditions as described in Materials and Methods.

² Data obtained from Fig 1A.

³ Parameters were fitted from Figs 1B and 6B. The original curve of ToPif1 unwinding substrate S₂₆D₁₇ were not shown.

⁴ ND, not determined.

To gain mechanistic insight into G4 unfolding and understand the molecular basis underlying the differences in G4 unfolding between the parallel and antiparallel G4s, smFRET assays were performed with G4 DNAs in which the bases near the 5′ and 3′ sides of G4 were fluorescently labeled. Therefore, its unfolding activity could be sensitively recorded by smFRET time traces. The smFRET time trajectories recorded with antiparallel G4s comprising three or four G‐tetrads (smG4^AP‐424 and sm4G4^AP‐333, respectively) were characterized by a stepwise unfolding procedure, in which the smFRET level successively decreased from E ≈ 0.90 to E ≈ 0.12 and then finally disappeared due to the escape of the unwound DNA from the coverslip (Fig 1C). This phenomenon has been observed previously in studies of ScPif1 by two different laboratories (Zhou et al, ²⁰¹⁴; Hou et al, ²⁰¹⁵) and was interpreted as that Pif1 helicase first unfolds G4 from the 5′‐lateral G‐column, resulting in a G‐triplex (G3), then further transforms G3 into a G‐hairpin structure and finally completely unwinds the duplex. Interestingly, for parallel G4 and under the same experimental conditions, although 30–35% time traces (type I) recorded were essentially the same as those observed with antiparallel G4, ~65% smFRET time traces (type II) for parallel G4 oscillated between E ≈ 0.90 and E ≈ 0.60 without further decrease before the fluorescence signal bleaching (Fig 1D). Thus, the smFRET assays confirmed that ToPif1 unfolds both parallel and antiparallel G4 using the same mechanism but with different G4 unfolding efficiencies. The reduced unfolding efficiency in parallel G4 may simply result from the higher thermal stability in the parallel G4 compared with the antiparallel G4 (Fig EV1D and E). We then experimentally address this possibility by determining T _m values of antiparallel (G4^AP‐424) and parallel (G4^P‐214) G4 DNAs using a CD melting assay. The results (Fig EV1D and E) show that T _m values determined with antiparallel G4 (G4^AP‐424) and parallel G4 (G4^P‐214) are 40.1 °C and 43.8 °C, respectively, indicating that the reduced unfolding efficiency in parallel G4 may simply result from the higher thermal stability of parallel G4.

The overall structure of ToPif1 in complex with G4

To probe the structural basis of ToPif1‐mediated G4 unfolding, ToPif1 was complexed with parallel and antiparallel G4 DNAs flanked at their 5′‐ and 3′‐ ends with 6 and 8 nt poly(dT), respectively, mimicking G4s formed at a replication fork (Lemmens et al, ²⁰¹⁵; Lee et al, ²⁰²¹). Only parallel G4 (G4^T6/T8) in complex with ToPif1 in a 1:2 ratio in the presence of ADP·AIF₄ was crystallized and diffracted to 2.58 Å. The asymmetric unit of the crystal structure contained one G4 that was sandwiched between two ToPif1 molecules (named as molecules a and b) bound to ssDNAs with 5′‐3′ polarity. The two protein molecules in the asymmetric unit were linked through the DNA substrate and related by a rotation of 102.8°, forming a dumbbell‐shaped structure (Fig 2A and B). Molecule a anchored close to the 5′‐most G‐tetrad and molecule b was bound downstream of the 3′‐most G‐tetrad. The distance between the binding/unwinding surface constituted by domains 1B/2B and the nearest G‐tetrad in the molecule a is ≈ 3.00 Å with the 5′‐most G‐tetrad, whereas in molecule b, the distance with the last G‐tetrad at the 3′ end is greater than 12.0 Å (Fig 2A and B). Therefore, molecule a in 5′ is in an active unwinding state during G4 processing, establishing many interactions with G4, whereas molecule b in 3′ is not positioned for G4 unwinding and establishes few contacts with G4 (Fig 2A and B). The domain folding and the spatial arrangement of the modules in ToPif1 adopt a similar architecture as those observed in the solved crystal structures of ToPif1 (Dai et al, ²⁰²¹) and other Pif1 family helicases from yeast (Lu et al, ²⁰¹⁸) and Bacteroides species (Zhou et al, ²⁰¹⁶). Domains 1A and 2A constitute a deep cleft where ADP·AIF₄ is bound, and domain 2B is composed of an SH3‐like domain and a prominent β‐hairpin (loop3) (Figs 2A and B and Fig EV1, EV2, EV3, EV4, EV5F–I). Each ToPif1 molecule binds a 6 nt poly(dT) stretch. Molecule a binds T1–T6, and molecule b binds T22–T27. The conformation of T2–T5 and T24–T27 are conserved, and they superpose with a root mean square deviation (RMSD) of 1.20 Å, as already shown for ToPif1 complexed with several ligands, ssDNA, and ss/dsDNA (Dai et al, ²⁰²¹); thus, domain 2B exhibits high conformational flexibility. Structural superposition between molecule a and the previously determined ToPif1 complexed with ss/dsDNA (Dai et al, ²⁰²¹) on domain 1A demonstrated that the center of mass of domain 2B moves upward by 5.40 Å and rotates 25.70° upon G4 binding. The domains 2A/1B, the configurations of bound ssDNA, and the amino acid residues involved in ssDNA binding overlap substantially (RMSD = 0.28 Å over 103 residues) (Fig EV2, EV3, EV4, EV5A–F), indicating that domain 2B assumes different conformations to accommodate G4 and double‐stranded DNA (dsDNA) binding, respectively. Similarly, superposing molecules a and b on domain 1A shows that G4 binding induced a stable β‐hairpin (loop3) formation, and the center of mass of domain 2B in the molecule a moved upward 5.70 Å, accompanied by 22.90° rotation (Fig 2C), suggesting that domain 2B undergoes a significant conformational adjustment before and after G4 unfolding. G4 is complexed with three K⁺ ions: two are located inside G4 between the tetrads and an extra K⁺ is on top of the 3′ side of G4. This K⁺ ion is stabilized by the last tetrad (tetrad III: G9·G13·G17·G21), T22, and R448 of ToPif1 molecule b. The overall topology of the G4 structure solved in this work is consistent with G4s previously reported in complexed with other proteins. They are characterized by a stack of three G‐tetrads, and the three short loops containing single thymine cross the grooves of the parallel G4 helix, joining the top and bottom G‐tetrads in the stack (Fig 2D). ToPif1‐G4^T6/T8 crystal structure exhibits a good fit with SAXS data (Figs EV3, EV4, EV5A–E and Table EV3) with a χ ² of 1.91, confirming that the conformation observed in the crystal state is also the one found in the solution.

Figure 2. Structure analysis of ToPif1 complexed with G‐quadruplex (G4) DNA and ADP·AlF₄ .

1. Overall structure of ToPif1‐G4^T6/T8‐ADP·AlF₄ ternary complex in cartoon mode with the conserved domains 1A (green), 1B (orange), 2A (pink), 2B (blue), and C‐terminal (CTD, brown). G4 DNA (G4^T6/T8) is colored in red and ADP·AlF₄ is shown as a black stick. ToPif1 molecules binding at the 5′‐ and 3′‐ends of G4 DNA are labeled as molecule a and molecule b, respectively.
2. Surface view after cutting the crystal structure ToPif1‐G4^T6/T8‐ADP·AlF₄ ternary complex and the domains are colored as in (A).
3. Structural superposition for the structures of molecules a and b with the mass centers of domain 2B shown as a sphere (mass centers of molecules a colored in blue and that of molecules b colored in cyan).
4. Comparison of G4^T6/T8 (red) in the ToPif1 ternary complex, G4^Myc (PDB: 5VHE; Chen et al, ²⁰¹⁸) colored in cyan and G4 ‘T‐loops’ (PDB: 6LDM; Traczyk et al, ²⁰²⁰) colored in yellow.
5. Residues of the molecule a that interact with G4 layers in the ToPif1‐G4^T6/T8‐ADP·AlF₄ ternary complex; G4 DNA (G4^T6/T8) is shown in red and the interacting residues are represented as sticks.

Figure EV1. ToPif1 recognition for G4 and ssDNA.

1. CD experiments of G4 DNA.
2. Left panel: DNA binding of ToPif1 to G4^CEB with 1 μM T₁₂ (black line) or 1 μM G4^CEB (blue line) or none of them (red line). The binding of fitting of the data to the equation (1) yields a K _d,app value for ToPif1‐G4^CEB of 30.30 ± 1.17 nM, a K _d,app value for ToPif1‐G4^CEB (added with 1 μM T₁₂) of 30.91 ± 2.32 nM. Right panel: DNA binding of ToPif1 to T₁₂ with 1 μM G4^CEB (red line) or 1 μM T₁₂ (blue line) or none of them (black line). Fitting of the data to the equation (1) yields a K _d,app value for ToPif1‐T₁₂ of 29.10 ± 0.53 nM, a K _d,app value for ToPif1‐T₁₂ (added with 1 μM G4^CEB) of 26.50 ± 0.71 nM. DNA binding assays were carried out in buffer A (25 mM Tris–HCl (pH 7.5), 50 mM NaCl, 2 mM MgCl₂, and 2 mM DTT) with 0.5 mM ADP·AlF₄.
3. Stopped‐flow unwinding kinetic curves of ToPif1 unwinding G4 DNA (G4^AP‐424, G4^P‐214, 4G4^AP‐333) in buffer A with 1 mM ADP·AlF₄.
4. CD analysis of G4^AP‐424 and G4^P‐214.
5. Normalized melting curves from CD melting assays for G4^AP‐424 and G4^P‐214 at different temperatures (20–75 °C).
6. Cartoon representation of ToPif1 (molecule a) with G4^T6/T8 in the presence of ADP·AlF₄ shown as a black stick and loop3 highlighted in magenta.
7. Cartoon representation of ToPif1‐dT₁₅‐ADP·AlF₄ (PDB: 6S3 M; Dai et al, ²⁰²¹) with ADP·AlF₄ shown as a black stick and loop3 highlighted in magenta.
8. Cartoon representation of BaPif1‐dT₁₀‐ ADP·AlF₄ (PDB: 5FHE; Zhou et al, ²⁰¹⁶) with ADP·AlF₄ shown as a black stick and loop3 highlighted in magenta.
9. Cartoon representation of ScPif1p‐poly (G3T5)‐ADP·AlF₄ (PDB: 5O6B; Lu et al, ²⁰¹⁸) with ADP·AlF₄ shown as a black stick and loop3 highlighted in magenta.

Figure EV2. Comparison of domain 2B of the molecule a of the ToPif1‐G4^T6/T8‐ADP·AlF₄ ternary complex and the previously determined ToPif1‐ss/dsDNA structure (PDB: 6S3H; Dai et al, ²⁰²¹).

1. Cartoon representation of molecule a in the ToPif1‐G4^T6/T8‐ADP·AlF₄ structure with domain 2B and the center mass (sphere) of domain 2B is highlighted in blue. G‐quadruplex (G4) DNA is shown in red.
2. Cartoon representation of the ToPif1‐S₇D₁₁‐ADP·AlF₄ structure (PDB: 6S3H; Dai et al, ²⁰²¹) with domain 2B shown in magenta. The center mass of domain 2B is shown as a red sphere.
3. Superposition on domain 1A of the structures in (A) and (B).
4. Domain 2B and G4 DNA of molecule a in the ToPif1‐G4^T6/T8‐ADP·AlF₄ structure with the close‐up view of interactions between ToPif1 and the G4 tetrad I.
5. Domain 2B and modeled duplex DNA in the ToPif1‐S₇D₁₁‐ADP·AlF₄ structure (PDB: 6S3H; Dai et al, ²⁰²¹).
6. Superposition on domain 1A of the structure in (C) with only domain 2B and G4 DNA being shown.

Figure EV3. SAXS results of ToPif1‐G4^T6/T8‐ADP·AlF₄ .

1. Fit curve of the SAXS data of ToPif1‐ G4^T6/T8‐ADP·AlF₄ with two modeled ToPif1 molecules bound to the substrate G4^T6/T8 calculated with Crysol.
2. The model of ToPif1‐G4^T6/T8‐ADP·AlF₄ superimposed on the ab initio envelope calculated with DAMMIF.
3. The SAXS Model of ToPif1‐G4^T6/T8 ‐ADP·AlF₄.
4. Crystal structure of ToPif1‐G4^T6/T8 ‐ADP·AlF₄.
5. Superposition for SAXS model and crystal structure of ToPif1‐G4^T6/T8‐ADP·AlF₄.

Figure EV4. Functional validation of ToPif1 mutants and ligand interaction diagram of T6/G7 in the ToPif1 complex compared with the guanine‐specific pocket (GSP) in the CsRecQ structure.

1. CD results of ToPif1 and its mutants.
2. ATP hydrolysis of ToPif1 and its mutants with or without DNA effector (S₁₈H₁₁).
3. Ligand interaction between T6 and ToPif1 in the ToPif1‐G4^T6/T8‐ ADP·AlF₄ structure.
4. Ligand interaction between G7 and ToPif1 in the ToPif1‐GR₁₇‐ ADP·AlF₄ structure (PDB: 7BIL; Dai et al, ²⁰²¹).
5. Ligand interaction between G21 and CsRecQ in CsRecQ‐ssDNA structure (PDB: 6CRM; Voter et al, ²⁰¹⁸).

Figure EV5. Sequence alignment of Pif1 helicase from bacteria, humans, and yeast.

Thermus bacteria Pif1 (1–8) are from Thermus oshimai (ToPif1), Thermodesulfovibrio yellowstonii (TyPif1), Sulfurihydrogenibium sp (SuPif1), Anaerobaculum hydrogeniformans (AhPif1), Thermodesulfobacterium commune (TcPif1), Deferribacter desulfuricans (DePif1), Mycolicibacterium thermoresistibile (MtPif1), and Synergistales bacterium (SbPif1). Bacteria at room temperature are from Bacteroides sp (BsPif1). Homo sapiens and Saccharomyces cerevisiae are labeled as HsPif1 and ScPif1. Residues involved in G4‐Recognizing Surface (GRS) are shown with green stars.

Structural basis of recognitions and interactions of the entire G‐quadruplex and the presumed G‐base from the unwound G4

ToPif1 recognizes the entire G4 through a cluster of amino acids at domains 1B/2B, constituting G4‐Recognizing Surfaces (GRS; Fig 2E). Although R419 and R392 interact with G11 and G19 in the 5′‐most G‐tetrad (G7·G11·G15·G19) through cation‐π interactions, respectively, the side chain amino group of Q327 interacts with the deoxyribose group of G11 through nonbonded contacts (Figs 3A and EV2D). Intriguingly, the negatively charged carboxyl group of E397 against the negatively charged oxygen atom on the ribose‐phosphate backbone near G7 will exert a repulsion force, which may influence the global conformation of the G4 (Fig 3B). Furthermore, the phosphate group of G20 in the middle tetrad (G8·G12·G16·G20) is stabilized by R135 (Fig 3C). Two positively charged residues, K329 and R150, interact with G9 and G21 in 3′‐G‐tetrad (G9·G13·G17·G21) through ionic and cation‐π interactions (Fig 3D). Thus, the residues of GRS form structural clusters which differentially recognize the guanine bases from three different tetrads in the G4. It is, therefore, plausible that the spatial conformation of G4 is destabilized by these interactions, and G4 is efficiently unfolded in coordination with translocation forces derived from ATP hydrolysis.

Figure 3. Atomic interactions between the integral G‐quadruplex (G4) DNA and ToPif1.

• A
  Interactions of ToPif1 and the G4 layer I.
• B
  The negative–negative repulsion interaction between E397 and the G7 of G4^T6/T8 in the ToPif1 ternary complex structure. The electrostatic potential at ± 5 kTe ⁻¹ was colored in blue (basic/positive), white (neutral), and red (acidic/negative).
• C, D
  Interaction of ToPif1 and G4 layers II and III, respectively. Molecular electrostatic potential map calculated with G4 DNA omitted from the co‐crystal structure (blue to red, ± 5 kTe ⁻¹).

In our ToPif1‐G4 complex structure (ToPif1‐G4^T6/T8), the T6 just upstream of the 5′‐tetrad is interacting with a cluster of amino acids (Fig 4A). T6 also stacks against G19 in the flat 5′‐G‐tetrad (G7·G11·G15·G19) through a π–π interaction. R392 and V395 are specifically involved in thymine base recognition through hydrogen bonds (R392‐T6 = 2.50 Å and V395‐T6 = 2.80 Å). Furthermore, the base configuration is further stabilized by π–π stacking between F394 and T6. The residues R355 and R135 interact with the ribose and the phosphodiester of T6 but not its thymine base. To determine whether the binding of a guanine base at the T6 position can remodel the conformation of GRS to accommodate G‐base‐specific binding, an ssDNA bearing a G4‐forming sequence (GR₁₇) was directly complexed with ToPif1 without prefolding (Dai et al, ²⁰²¹). The crystal structure at a resolution of 2.21 Å demonstrated that the global conformation of ToPif1‐GR₁₇ (Dai et al, ²⁰²¹) is similar to the complex ToPif1‐G4^T6/T8 (Fig 4B), except that the G4 structure is absent. Structural superposition between the two structures revealed three remarkable features: (i) the G7 in binary complex (ToPif1‐GR₁₇) occupies the equivalent spatial position of T6 in ToPif1‐G4^T6/T8 (Fig 4C); (ii) both G7 and T6 flip to the same degree (Fig 4D, ≈113.0°); (iii) both the bases and ribose groups of G7 and T6 are bound by the same residues (R392, V395, and R335) with essentially the same conformation, except that residue N356 establishes an additional hydrogen bond with the phosphate group of G7 upon its binding (Fig 4D). Therefore, if base G7 is considered as the first G base from an unfolded G‐tetrad, there is no striking structural reorganization to accommodate the binding of G rather than T. Although G7 in GR₁₇ described previously provides a conformation for G4 being unwound in the first base, other mechanisms may also exist.

Figure 4. Structural superposition of ToPif1 with the G4^T6/T8 and the GR₁₇ .

1. Interaction between ToPif1 and the T6 of G4^T6/T8 in the ToPif1‐G4^T6/T8‐ADP·AlF₄ ternary complex.
2. Superposition on domain 1A of structures in ToPif1‐G4^T6/T8 and ToPif1‐GR₁₇.
3. Interaction between ToPif1 and the G7 of GR₁₇ in the ToPif1‐GR₁₇‐ADP·AlF₄ ternary complex.
4. Structural superposition of interacting residues for ToPif1 with the T6 of G4^T6/T8 and the G7 of GR₁₇.

Mutational analysis of the residues involved in G4 binding and unfolding

According to the potential functions of certain residues in G4 binding or/and unfolding, structurally guided single alanine substitution variants of ToPif1 were purified to homogeneity, including (i) the mutants R392A, R135A, and R355A, which interact with the T6/G7 base and/or the ribose/phosphodiester moieties; (ii) four variants (Q327A, E397A, R419A, and K329A) involved in the 5′‐most G‐tetrad; and (iii) R135 and R150 involved in the interactions with middle G‐tetrad and the 3′‐most G‐tetrad, respectively (Table 2). To make sure that the each given residue mutation does not result in a global conformational change, which will affect ascertain the distinct role of the given residue in G4 binding/unfolding, we first performed circular dichroism (CD) spectroscopy and dynamic light scattering (DLS) assays, which will provide the secondary structure information and hydrodynamic behavior (monomer or aggregation) of the variants. The results show that the variants displayed the same CD spectra as ToPif1, and the determined DLS parameters are essentially the same between ToPif1 and variants (Fig EV4A and Table EV4). The DNA‐dependent ATPase activities of the ToPif1 variants were measured using a coupled spectrophotometric ATPase assay. Fig EV4B and Table EV4 demonstrated that the determined K _m and k _cat values with the variants are not significantly different from those determined with the wild‐type ToPif1.

Table 2.

Parameters of binding and unwinding activities of ToPif1 and its variants ¹ .

Variants	G4 (G4Tel) binding	G‐quadruplex (AP‐S16‐TelG4) unwinding			ssDNA (T12) binding	dsDNA (S26D17) unwinding
	K d,app (nM)	A m	k cat (s−1)	k cat/K d,app 2	K d,app (nM)	A m	k cat (s−1)	k cat/K d,app 3
ToPif1	29.63 ± 2.64	0.53 ± 0.01	0.48 ± 0.01	0.0162 ± 0.0012	30.20 ± 1.41	0.59 ± 0.02	3.53 ± 0.02	0.1169 ± 0.0031
R392A	28.50 ± 2.20	0.40 ± 0.01	0.27 ± 0.02	0.0095 ± 0.0001	28.30 ± 1.80	0.32 ± 0.01	3.50 ± 0.02	0.1237 ± 0.0067
R355A	133.00 ± 18.04	0.38 ± 0.02	0.07 ± 0.01	0.0005 ± 0.0001	38.10 ± 1.15	0.34 ± 0.01	1.27 ± 0.01	0.0333 ± 0.0005
R135A	119.00 ± 10.67	0.34 ± 0.02	0.05 ± 0.01	0.0004 ± 0.0001	30.90 ± 0.47	0.47 ± 0.01	1.42 ± 0.02	0.0460 ± 0.0010
Q327A	51.50 ± 3.23	0.48 ± 0.01	0.41 ± 0.02	0.0080 ± 0.0002	63.40 ± 9.42	0.43 ± 0.03	6.14 ± 0.03	0.0968 ± 0.0121
R419A	38.50 ± 2.25	0.24 ± 0.02	0.21 ± 0.01	0.0055 ± 0.0001	38.20 ± 5.47	0.32 ± 0.02	2.46 ± 0.01	0.0644 ± 0.0016
K329A	48.70 ± 2.65	0.37 ± 0.01	0.29 ± 0.02	0.0060 ± 0.0001	30.00 ± 1.38	0.39 ± 0.02	3.38 ± 0.01	0.1127 ± 0.0040
R150A	152.00 ± 12.97	0.28 ± 0.01	0.06 ± 0.01	0.0004 ± 0.0001	31.30 ± 1.76	0.43 ± 0.03	2.50 ± 0.01	0.0799 ± 0.0039
E397A	28.30 ± 2.21	1.00 ± 0.01	1.19 ± 0.01	0.0420 ± 0.0027	29.40 ± 1.36	0.49 ± 0.02	3.79 ± 0.02	0.1289 ± 0.0050
E397L	33.00 ± 3.24	0.88 ± 0.01	1.18 ± 0.01	0.0358 ± 0.0020	36.00 ± 2.84	0.48 ± 0.01	2.36 ± 0.02	0.0656 ± 0.0043
E397H	44.60 ± 1.68	0.78 ± 0.03	1.32 ± 0.01	0.0296 ± 0.0009	50.90 ± 3.45	0.31 ± 0.02	2.64 ± 0.02	0.0519 ± 0.0030
E397D	38.80 ± 8.07	0.45 ± 0.01	0.35 ± 0.02	0.0090 ± 0.0011	28.00 ± 3.18	0.49 ± 0.01	3.27 ± 0.01	0.1168 ± 0.0116

¹ The reported data are the average values determined from 3‐4 independent assays as performed under experimental conditions as described in ‘Materials and Methods’.

² The values are the equivalent ratios of k _cat (G‐quadruplex unwinding) and K _d,app (G4 binding) comes from the data of Figs 1A, 5A and 6A.

³ The values are the equivalent ratios of k _cat (dsDNA S₂₆D₁₇ unwinding) and K _d,app (ssDNA T₁₂ binding).

To ascertain qualitatively how the modified residues affect the unfolding activities, the K _d,app values, and the steady‐state kinetic parameters, including the unfolding magnitude (A _m) and the unfolding rate (k _cat), were determined with fluorescent‐labeled ssDNA (T₁₂), partial duplex DNA (S₂₆D₁₇), and G4 DNA substrates as described above. Several representative G4 unfolding kinetic curves are shown in Fig 5A, and the kinetic parameters of all variants are summarized in Table 2. The parameters (K _d,app, A _m, and k _cat) determined with G4 and ssDNA substrates showed a range of variability. To best understand the mutation results, the ratio of k _cat/K _d,app, a useful index for comparing the catalytic efficiency (Barlow et al, ²⁰⁰⁹), was determined (Table 2). Arranging the G4 k _cat/K _d,app values in ascending numerical order shows that the relative increase in G4 k _cat/K _d,app values can be roughly classified into three categories (Fig 5B): (i) group I (R355A, R135A, and R150A) is characterized by zero values in k _cat/K _d,app for G4 unfolding, while their k _cat/K _d,app values determined from dsDNA unwinding are reduced by 32% (R150A) and 72% (R355A); (ii) group II (R392A, Q327A, K329A, and R419A) is marked by about 41–66% reduction in k _cat/K _d,app values for G4 unfolding and the corresponding values determined with dsDNA are essentially the same as these determined with ToPif1, except that k _cat/K _d,app value determined with R419A for dsDNA is reduced to 45%; (iii) group III (E397A, E397L, and E397H) displays surprising results in which k _cat/K _d,app values determined with G4 are increased to 2–2.6 folds, whereas the k _cat/K _d,app values for dsDNA are inversely reduced to 50% compared with that of ToPif1. Taken together, these results indicate that alteration of the residues in groups I–II, particularly the residues in group I, significantly impair G4 unfolding but only moderately reduces dsDNA unwinding. Consistently, smFRET time trajectories recorded with the mutants in group III revealed that in sharp contrast to ToPif1, the curves only oscillate between 0.90 and 0.60 but scarcely attain a completely unfolded level, indicating that the variants can just release the 5′‐most lateral G column from G4 DNA but is unable to unfold the integral G4 DNA completely (Fig 5C–F).

Figure 5. G‐quadruplex (G4) unwinding activity of wild‐type ToPif1 (ToPif1) and its mutants.

• A
  Stopped‐flow kinetics of the G4 (AP‐S₁₆‐TelG4) unwinding activity of ToPif1 and the various modified proteins. The different unwinding curves represent an average of three to four independent experiments for each substrate. The solid line represents the fit of data according to equation (2).
• B
  Bar plot of k _cat/K _d,app values in Table 2 (n = 3 biological replicates). The k _cat/K _d,app values of G4 in descending numerical order (front panel), and these are determined with the dsDNA (S₂₆D₁₇, back panel). Data are presented as means ± SD.
• C–F
  The typical smFRET trajectories of antiparallel G4 (smG4^AP‐424) unwinding catalyzed by different ToPif1 mutants based on single‐molecular FRET assays. The dashed magenta, black, and red lines represent the states of G4: fully folded, incompletely folded, and fully unfolded, respectively. The concentrations of the R355A (C), R135A (D), R150A (E), and R419A (F) mutants were 50 nM with 150‐μM ATP. The experimental conditions are described in Methods. Black arrows in the solid style indicate the time of adding proteins.

In group III, the kinetic parameters determined with E397A were significantly higher in terms of unfolding amplitude (A _m) and rate (k _cat), respectively (Figs 5B and 6A and Table 2). Compared with ToPif1, the k _cat/K _d,app value of E397A determined for G4 unfolding increased by 2.5‐fold, although that for dsDNA remained unchanged (Table 2). This stimulating effect was also observed with parallel G4s (Fig 6B). To confirm the observation, residue E397 was, therefore, replaced with a hydrophobic (E397L), a positively charged (E397H), and a negatively charged (E397D) residue. As expected, G4 unfolding activities determined with the first two variants were systematically higher than those of ToPif1. E397D bearing a negative charge displayed activity comparable to the wild type (Fig 6A). To gain in‐depth mechanistic insight into how residue E397 affects G4‐unfolding activity, all of the variants were further studied with smFRET techniques. Analysis of the smFRET time traces demonstrated that the unfolding times (t(s)) determined with the three typical variants were significantly reduced, except those for E397D, whose unfolding time remained essentially the same as that of ToPif1 (Fig 6C–F). The mutation‐stimulating G4‐resolution activity is surprising and raises the question of whether this may be a general autoregulatory mechanism by which the activity of G4 unfolding is regulated both in vitro and in vivo.

Figure 6. Stimulating effect upon mutation of E397.

• A
  Stopped‐flow kinetics of the antiparallel G4 (AP‐S₁₆‐TelG4) unwinding activity of ToPif1 and its mutants. The different curves represent an average of three to four independent experiments for each substrate. The solid line represents the fit of data according to equation (2).
• B
  E397A mutant‐stimulating effect determined with parallel G4 DNA.
• C–E
  The typical FRET trajectories of antiparallel G4 (smG4^AP‐424) unwinding catalyzed by different point mutations on E397 using single‐molecular FRET assays. The dashed magenta and red lines represent the fully folded and unfolded states of G4, respectively. The different unwinding curves represent an average of three to four independent experiments for each substrate. Black arrows in the solid style indicate the time of adding proteins.
• F
  Error bar plot of unwinding time of E397A, E397L, and E397H determined from unfolding curves of (C), (D), and (E), respectively (n = 3 biological replicates). The experimental conditions are described in Materials and Methods and data are presented as means ± SD.

Discussion

While increasing evidence has demonstrated that helicase‐mediated G‐quadruplex unfolding plays an essential role in preserving genome stability, the lack of structural information on G4‐processing helicases still hampers our mechanistic understanding of G4 resolution. With a wide array of approaches ranging from bulk/smFRET assays to structural biology analyses, our results reported here provide in‐depth insight into the currently largely unknown, but mechanistically important issue of how the Pif1 helicase family recognizes the integral G4 structure.

To our best knowledge, the structure reported here is the first co‐crystal structure of SF1 helicase bound to G4, whereas crystal structures of SF2 helicase‐G4 complexes available are DHX36 and CsRecQ proteins. Thus, a structural comparison between SF1 and SF2 helicase‐mediated G4 binding/unfolding may provide the common and distinct mechanism underlying G4 unfolding.

DHX36 protein, one of the most extensively studied G4‐unfolding helicases, is characterized by a DHX36‐specific motif (DSM), which specifically binds parallel G4s (Hossain et al, ²⁰²¹). Interestingly, the X‐ray crystal structure of DHX36 (Chen et al, ²⁰¹⁸) in complex with a Myc promoter‐derived G4 demonstrates a rearrangement of the 5′‐G‐tetrad with a one‐base translocation upon DSM binding (Chen et al, ²⁰¹⁸). Because an NMR study shows that G4 keeps its integral state upon the binding of an isolated DSM (Heddi et al, ²⁰¹⁵), it was, therefore, suggested that DSM, in coordination with the helicase core, unfolds the G4 one base at a time in an ATP‐hydrolysis‐independent manner (Chen et al, ²⁰¹⁸). Recently, a crystal structure solved from a helicase core of a RecQ family helicase (CsRecQ) complexed with oligonucleotides harboring a human telomere sequence (Voter et al, ²⁰¹⁸) revealed that a guanine base within the putative G‐tetrad is located at the 5′‐gate of the electropositive channel where the G‐base is flipped and sequestrated (Voter et al, ²⁰¹⁸). This structure is interpreted to represent “a product complex of CsRecQ‐unwound G4 rather than a folded quadruplex” (Voter et al, ²⁰¹⁸), although the complexed ssDNA was not prefolded to form a stable G4 before the preparation of the G4‐CsRecQ complex.

DHX36 uses DSM to bind and unfold G4s, and some G4‐resolving helicases as FANCJ also possess a motif equivalent to DSM (Chen et al, ²⁰¹⁸; Lee et al, ²⁰²¹). However, not all the SF2 helicases bear DSM. An intriguing and unresolved question is how the G4‐resolving helicases generally recognize and unfold G4s. We demonstrated that the 1B/2B domains in ToPif1 constitute a G4‐Recognizing Surface (GRS) in which the residues recognize the whole G4 structure, not just a tetrad, and differentially interact with the ribose and phosphate units in different tetrads. Superposition of the structural conformation of G4 complexed with ToPif1 with that of the DHX36‐Myc complex (Chen et al, ²⁰¹⁸) demonstrated that, when the G4s well superposed, loop3 in ToPif1 occupies the spatial position of the DSM motif in DHX36, binding on the 5′ side of the G4. Given that Pif1 and DHX36 have opposite polarity for ssDNA translocation, it appears that helicase binding on the 5′ side of G4 is a critical feature for G4 unwinding. Furthermore, it appears that the previously identified GSP in the RecQ helicase family is not conserved in ToPif1 because the residues involved in the first base downstream of 5′‐G‐tetrad are not selective for guanine and do not form the equivalent hydrogen bonds that stabilize the guanines within G4s as observed in GSPs (Figs 1, 2, EV1, EV2, EV3, EV4, EV5, 3, 4, 5, 6A–D and EV4C–E). If base G7 is considered as the first G base from an unfolded G‐tetrad, there is no striking structural reorganization to accommodate the binding of G rather than T. Compared with the previously identified GSP, the cluster of residues surrounding the base (T6/G7) upstream of the G‐tetrad in ToPif1 is not bound with the equivalent residues identified in the GSP in CsRecQ (Fig EV4, EV5C–E). Of note, the G‐rich ssDNA including GR₁₇ used by Dai et al (2021) and that by Voter et al (2018) were not prefolded to form stable G4 DNA in the ToPif1 and CsRecQ structures; it, therefore, remains to be determined whether these structures (ToPif1‐GR₁₇; Dai et al, ²⁰²¹ and CsRecQ‐ssDNA; Voter et al, ²⁰¹⁸) represent an intermediate or/and a product complex of helicases bound to unwound G4 DNA or just ssDNA bound to the proteins.

Our binding studies demonstrated that ToPif1 possesses strong G4‐binding activity without any topological preference. This is not surprising because ToPif1 does not bear a DSM motif which was reported exclusively bind parallel G4 (Hossain et al, ²⁰²¹). The smFRET time trajectory feature of ToPif1‐mediated G4 unfolding is mainly characterized by a stepwise unfolding procedure, without a long oscillation period, which is consistent with the previously recorded smFRET time trajectory with ScPif1 (Zhou et al, ²⁰¹⁴; Hou et al, ²⁰¹⁵). These results imply that ToPif1 translocates along the DNA lattice and sequentially unfolds G4 into G‐triplexes and then into G‐hairpins. ToPif1‐mediated G4 processing is radically different from previous reports on DHX36 (Chen et al, ²⁰¹⁸) and the RecQ (Voter et al, ²⁰¹⁸) helicase family, which were proposed to use a common mechanism of ATP‐independent repetitive unfolding to resolve G4s: binding of these helicases triggers the reorganization and destabilization of G4 with a one‐at‐a‐time type base translocation, and ATP is only needed for releasing the unfolded G4 (Chen et al, ²⁰¹⁸; Voter et al, ²⁰¹⁸). Therefore, it appears that SF1 helicases unfold G4 in a stepwise manner, whereas SF2 helicases unfold G4 repetitively; however, establishing these observations as a general rule needs further studies. Another interesting finding is that two negative charges between E397 and the oxygen atom on the DNA backbone near G7 repel one another. A series of mutations at E397 demonstrated that replacement of the negative residue by any non‐negative residue stimulates G4‐unfolding activity, raising the question that whether an autoinhibitory mechanism applies to other ToPif1 helicases. A sequence alignment of Pif1 family helicases revealed that a negative residue (E/D) is highly conserved in thermophilic gram‐negative bacteria (Fig EV5). However, there is a tendency toward replacing negative residue (E/D) with no‐negative residues (S/G) in evolutionarily advanced species (Fig EV5). Therefore, it appears that ToPif1 has not evolved with an optimum efficiency, but this could be due to growth conditions, G4 frequency in the genome, or a yet unknown regulatory mechanism stimulating ToPif1 activity.

The relation between growth temperature, G4 frequency, and Pif1 efficiency is not straightforward. Analysis of the occurrence of putative G4 sequences (PQS) in bacteria has recently been conducted. Though thermophile bacteria have genomes with higher GC content, the distribution of PQS is higher than expected by random chance in bacteria groups Deinococcales and Thermales while using either quadparser (Ding et al, ²⁰¹⁸) or G4Hunter (Bartas et al, ²⁰¹⁹) algorithms. In a later study, it was shown that T. oshimai had the highest frequency of PQS among all bacterial genomes explored (Table EV5). Surprisingly, the same study showed that Thermotogae, which are also extremely thermophilic gram‐negative bacteria, have the lowest frequency of PQS (Table EV5). Therefore, G4 occurrence does not seem to be linked to thermal conditions of growth. Bacteria from the Thermotoga group possess also a Pif1 helicase which exhibits G4 activity (Andis et al, ²⁰¹⁸). The biochemical activities of Pif1 from Thermotoga elfi (TePif1) have been measured, and TePif1 exhibits a C‐terminal WYL domain with ssDNA‐binding function (Andis et al, ²⁰¹⁸) necessary for substrate binding and helicase activity of TePif1. WYL domain is shared by Pif1 from other groups, such as Deferribacteres, but ToPif1 does not have one. Since Pif1 from other thermophilic bacteria such as Thermotoga elfii possessed an extra WYL domain (Andis et al, ²⁰¹⁸), modulation of Pif1 activity in cis (WYL domain) and a regulatory effect on Pif1 in trans by unidentified co‐factor proteins can be hypothesized.

It was recently reported that the proteins Mgs1 and Mms1 in S. cerevisiae binding to G4 motifs in vivo partially depends on the helicase ScPif1 (Schwindt & Paeschke, 2018; Paeschke & Burkovics, ²⁰²¹). Using the sequence similarity search tool, we identified the Mgs1‐like protein in the Thermus oshimai genome (ToMgs1 like, GeneBank: WP_018462088.1) which shares 29% identity with ScMgs1 across their lengths. ToMgs1 could be a co‐factor regulating ToPif1 in trans.

Though Pif1 is conserved from bacteria to humans, Pif1 has very diverse functions in bacteria, yeast, and human. Pif1 sequence exhibits variations with the insertion domain 2C in budding yeast (Lu et al, ^{2018, 2021}) and accessory domains as WYL (Andis et al, ²⁰¹⁸). Furthermore, sequence alignments show that some of the key residues identified in GRS are shared with other bacteria but not with Pif1 from eukaryotes (for an alignment, see fig S1 in Dai et al, ²⁰²¹). We show here that ToPif1 can rely only on its helicase core domain to recognize and unfold G4 DNA, whereas the activity of TePif1 needs a WYL accessory domain (Andis et al, ²⁰¹⁸). Therefore, it appears that Pif1 from different species adopts different solutions for enhancing its activity. Thus, co‐factors acting in cis or trans may enhance Pif1 activity for certain species, and the situation is far from being universal even in the bacteria domain. Further studies, instigated by the work reported here, are needed to complete our understanding of the general mechanism underlying how G4‐resolving helicases recognize and unfold G4s.

Materials and Methods

Reagents and buffers

All chemicals were reagent grade and all buffers were prepared in high‐quality deionized water from a Milli‐Q ultrapure water purification system (Millipore) with a resistivity of > 18.2 MΩ.cm and were filtered again on a 0.20‐μm filter before use. All DNA unwinding and binding assays were performed in buffer A (20 mM Tris–HCl (pH 7.5 at 37 °C), 50 mM NaCl, 2 mM MgCl₂, and 2 mM DTT) which was optimized previously by our group (Dai et al, ²⁰²¹).

DNA substrate preparation

All the DNA substrates used in this study were chemically synthesized and HPLC‐purified by Sangon Biotech (Shanghai) as listed in Table EV1. The oligonucleotides used in binding and unwinding assays were prepared at a 2 μM working concentration. The duplex or G4 DNA used in the stopped‐flow and smFRET assays was heated to 95 °C for 5 min in stocking buffer (20 mM Tris–HCl, pH 7.5, 100 mM NaCl) and annealed by slow cooling to room temperature. The DNA substrates in crystallization were dissolved in 20 mM Tris–HCl (pH 7.5) with 100 mM KCl, heated to 95 °C, and allowed to cool slowly to room temperature in a water bath. After purification on a Mono Q column, the formation of G4 structures was checked using circular dichroism (CD) spectropolarimetry.

Protein expression and purification

ToPif1 (residues 64‐507) and its mutants (R355A, R135A, R150A, R419A, R392A, Q327A, K329A, E397A, E397L, E397H, and E397D) were all cloned into pET15b‐SUMO and then transformed into the C2566H E. coli strain (New England Biolabs), respectively. When the culture reached the early stationary phase (OD₆₀₀ = 0.55–0.6) at 37 °C, 0.3 mM IPTG was added and the protein expression was induced at 18 °C over 16 h. Cells were harvested by centrifugation (4,500 g, 4 °C, 15 min) and pellets were suspended in lysis buffer (20 mM Tris–HCl pH 7.5, 500 mM NaCl, 10 mM imidazole, and 5% glycerol (v/v)). Cells were broken with a French press and then further sonicated two to three times to shear DNA. After centrifugation at 12,000 rpm for 40 min, the supernatants were filtered through a 0.45 μm filter and loaded onto a Ni²⁺ charged IMAC column (GE Healthcare). After washing twice, the SUMO‐ToPif1 was then eluted from the Ni²⁺ affinity column with elution buffer (20 mM Tris–HCl, pH 7.5, 500 mM NaCl, 300 mM imidazole, and 5% glycerol (v/v)) at 4 °C. The eluted protein was treated with SUMO protease (Invitrogen, Beijing). Then, the SUMO digested protein was further purified by a HiTrap Heparin column (GE Healthcare) to remove the SUMO‐tag and other protein impurities. The eluted fraction containing ToPif1 was collected and concentrated. The final purified protein was dialyzed against the storage buffer (20 mM Tris–HCl, pH 7.5, 500 mM NaCl, 1 mM DTT) and concentrated to approximately 10 mg/ml for crystallization and was about 95% pure as determined by SDS–PAGE. Mutations and truncations were engineered by PCR overlapping‐PCR protocol.

Equilibrium binding assays

The isothermal binding curves were determined using a fluorescence polarization assay on an Infinite F200 plate reader (Tecan). FAM‐labeled DNA substrates were used in this study. Varying amounts of protein were added to a 150 μl aliquot of buffer A containing 5 nM FAM‐labeled DNA. Each sample was allowed to equilibrate in solution for 5 min at 37 °C, and then fluorescence polarization was measured. Less than 5% change was observed between the 5 and 10 min measurements, indicating that equilibrium was reached in 5 min. The equilibrium dissociation constants were determined by fitting the binding curves using equation (1):

$$\mathrm{\Delta }r=\mathrm{\Delta }{r}_{max}\times P/(K_{d,app}+P)$$ (1)

where ∆r _max is the maximal amplitude of the anisotropy (i.e., r _max – r _free,DNA), P is the helicase concentration, and K _d,app is the midpoint of the curve corresponding to the apparent dissociation constant.

Crystallization of ToPif1‐nucleic acid complexes

For crystallization, purified ToPif1 was incubated with the G4^T6/T8 after dissolving and folding in 20 mM Tris–HCl (pH 7.5) with 100 mM KCl at a molar ratio of 2:1 in the presence of ATP analog ADP·AlF₄. The resulting ToPif1‐G4^T6/T8‐ADP·AlF₄ complex was purified using size‐exclusion chromatography and then concentrated at approximately 10 mg/ml. Crystallization trials on ToPif1 and its complexes with DNA (Table EV1) and ADP·AlF₄ were performed at 20 °C using the sitting‐drop vapor diffusion method. Crystallization screening was carried out at 20 °C using commercial screening kits (Hampton Research, Molecular Dimensions, and Rigaku Reagents), where the ToPif1‐DNA complex was mixed at a 1:1 ratio with the reservoir solution. Crystals of ToPif1‐G4^T6/T8‐ADP·AlF₄ were obtained using 0.1 M Bis‐Tris (pH = 6.5), 10% PEG20000. All these conditions were optimized with a grid search using 48 well Linbro plates at 20 °C, where 1 μl of protein sample and 1 μl of precipitant were mixed and equilibrated with 60 μl of precipitant.

X‐ray data collection, phasing, and refinement

All X‐ray diffraction data were collected on beamline BL19U1 (Zhang et al, ²⁰¹⁹) at the Shanghai Synchrotron Radiation Facility (China) using a Pilatus 6 M detector (Dectris) and were processed using XDS (Kabsch, 2010). ToPif1‐G4^T6/T8‐ADP·AlF₄ structures were solved by molecular replacement, performed in the PHENIX software suite (Adams et al, ²⁰¹⁰) with Phaser (McCoy et al, ²⁰⁰⁷), using the ToPif1 apo structure (PDB: 6S3E; Dai et al, ²⁰²¹) as the search model (Lovelace et al, ²⁰⁰⁵). Manual rebuilding was done using the Coot (Emsley et al, ²⁰¹⁰) and further refinement was performed in PHENIX. Cell parameters and data collection statistics are reported in Table EV2.

Small‐angle X‐ray scattering assay

SEC‐SAXS experiment was carried out at beamline SWING (SOLEIL Synchrotron, Saint‐Aubin, France) with SEC‐HPLC coupled to SAXS. The sample of ToPif1‐G4 complex was injected at a concentration of 10 mg/ml on a Superdex 200 5/150 Increase column (Cytiva) at a flow rate of 0.2 ml/min equilibrated in buffer 25 mM Hepes (pH = 7.5), 150 mM KCl, and 5% glycerol. Scattering data were collected at 20 °C using a Pilatus 1 M detector (Dectris), and the data reduction and processing of images were done with Foxtrot (David & Pérez, 2009). Further analysis was done with ATSAS 2.8 suite (Petoukhov et al, ²⁰¹²). Experimental R _g, I(0), and D _max values were calculated with PRIMUS and GNOM4 programs, respectively. Ab initio envelopes for isolated complexes were determined using DAMMIF with the pair distance distribution function calculated with GNOM4. Full atomic models derived from crystal structure were modeled with MODELLER and adjusted to SAXS data with DADIMODO. In this modeling procedure, only the missing parts from the X‐ray structure were kept flexible. Profiles of atomic models were calculated and fitted to the experimental data using CRYSOL and aligned on ab initio bead models with SUPCOMB. All the SAXS parameters are summarized in Table EV3.

Stopped‐flow unwinding assay

Briefly, unwinding kinetics were measured in a two‐syringe mode, where ToPif1 and fluorescently labeled DNA substrate were preincubated at 37 °C in one syringe for 5 min and the unwinding reaction was initiated by rapidly mixing with ATP from another syringe. Each syringe contained unwinding reaction buffer A (25 mM Tris–HCl (pH 7.5), 50 mM NaCl, 2 mM MgCl₂, and 2 mM DTT). All concentrations listed are after mixing, unless otherwise noted. For converting the output data from volts to percentage unwinding, a calibration experiment was performed in a four‐syringe mode, where the helicase, the hexachlorofluorescein‐labeled single‐stranded oligonucleotides, the fluorescein‐labeled single‐stranded oligonucleotides, and ATP were in four syringes, respectively. The fluorescent signal of the mixed solution from the four syringes corresponded to 100% unwinding. The standard reaction was performed with 4 nM DNA substrates, 1 mM ATP, and 100 nM ToPif1 in buffer A.

All stopped‐flow kinetic traces were averages of ≥ 10 individual traces. The kinetic traces were analyzed using Biokine (version 4.26, BioLogic, France) using equation (2):

$$A(t)=A_1\left(1‐e^{‐k_1\left(t‐t_0\right)}\right)+A_2\left(1‐e^{‐k_2(t‐t_0)}\right),$$ (2)

where A(t) represents the fraction of DNA unwound at time t, A ₁ and A ₂ are the unwinding amplitudes, k ₁ and k ₂ are the unwinding rate constants of the two phases, and t ₀ is the time at which the fraction of DNA unwound starts to rise. From the four parameters obtained through the fitting, we can determine the total unwinding amplitude A _m = A ₁ + A ₂ and the initial unwinding rate (i.e., the slope of the kinetic unwinding curve at early times) k _cat = k ₁ A ₁ + k ₂ A ₂.

Single‐molecule fluorescence data acquisition

Fifty picomolar fluorescently labeled DNA was added to the chamber containing imaging buffer composed of 20 mM Tris–HCl (pH 7.5), 1 mM MgCl₂, 50 mM NaCl, 1 mM DTT, and an oxygen scavenging system (0.8% D‐glucose, 1 mg/mL glucose oxidase, 0.4 mg/mL catalase, and 1 mM Trolox). After immobilization for 10 min, free DNA molecules were removed by washing with the imagining buffer. We used an exposure time of 100 ms for all recordings at a constant temperature of 22 °C. FRET efficiency was calculated using I _A/(I _D + I _A), where I _D and I _A represent the intensities of donor and acceptor, respectively. Basic data analysis was carried out by scripts written in MATLAB, and all data fitting were generated using Origin 9.0. Histograms were fitted to Gaussian distributions, with the peak positions unrestrained.

Circular dichroism spectropolarimetry

Circular dichroism (CD) experiments were performed with a BioLogic MOS450/AF‐CD optical system (BioLogic Science Instruments, France) equipped with a temperature‐controlled cell holder, using a quartz cell with a 1 mm path length. A 2.5 μM solution of G4 DNA was prepared in 25 mM Tris–HCl (pH 7.5) and 100 mM NaCl. CD spectra were recorded in the UV (220–320 nm) regions in 0.75 nm increments with an averaging time of 2 s at 25 °C.

Dynamic light scattering (DLS) assay

Dynamic light scattering measurements were performed at 37 °C using a DynaPro NanoStar instrument (Wyatt Technology Europe GmbH, Germany) equipped with a thermostated cell holder and disposable cuvettes (UVette, Eppendorf). All solutions were filtered using 0.22 μm filters. Purified ToPif1 or its mutants (10 μM) were dialyzed in buffer A (final volume, 30 μl). The scattered light was collected at an angle of 90°. Recording times were typically between 3 and 5 min (20−30 cycles on average, 10 s in each cycle). The analysis was performed with the Dynamics 7.0 software using the regularization method (Wyatt Technology). The molecular weight was calculated from the hydrodynamic radius using the following empirical equation (3),

$$M\mathrm{w}={\left(1.68\times R_{\mathrm{H}}\right)}^{2.34}$$ (3)

where Mw and R _H represent the molecular weight (in kDa) and the hydrodynamic radius (in nm), respectively.

ATP hydrolysis measurement

The malachite green assay (MAK113; Sigma‐Aldrich) was used to test the ATP hydrolysis activity of ToPif1 and its mutants by measuring the release of inorganic phosphate as previously described (Liu et al, ²⁰²¹). Briefly, 100 nM purified protein and ATP (concentrations ranging from 0 μM to 400 μM) were incubated with or without a 1 mM DNA effector (S₁₈H₁₁, sequence were shown in Table EV1) using a VICTOR Nivo Multimode Microplate Reader (PerkinElmer, Waltham, MA, USA). All reactions were performed in triplicates of 150 μl each at 25 °C in buffer A. The parameters of K _m and k _cat were calculated by fitting the experimental data to the Michaelis–Menten equation using Origin 9.0 (OriginLab Corporation, Northampton, MA, USA).

Author contributions

Conceived and designed study: X‐GX, X‐MH, and SR. Experiments: Y‐XD, H‐LG, N‐NL, W‐FC, XA, and H‐HL. Data processing: Y‐XD, N‐NL, W‐FC, and H‐LG. Discussion and interpretation of data: X‐GX, X‐MH, W‐FC, and SR. Manuscript writing and editing: X‐GX, Y‐XD, N‐NL, BS, and SR, with input from all authors.

Disclosure and competing interests statement

The authors declare that they have no conflict of interest.

Supporting information

Expanded View Figures PDF

Table EV1

Table EV2

Table EV3

Table EV4

Table EV5

EMBO reports (2022) 23: e53874.

Data availability

Accession codes: The atomic coordinate and structure factors have been deposited in the Protein Data Bank (https://www.rcsb.org/) with the following accession numbers: 7OAR (ToPif1‐G4^T6/T8‐ADP·AlF₄).