CdCl₂, ZnCl₂, MgCl₂, MnSO₄, CaCl₂, NiSO₄, CuSO₄, EDTA, dithiothreitol (DTT), sodium dodecyl sulfate(SDS), nucleotide triphosphates (NTPs), and deoxynucleotide triphosphates (dNTPs) were purchased from Sangon Biotech (Shanghai) Co., Ltd. Amino acids, ATP Assay Kit was purchased from Beyotime Biotechnology (Shanghai, China).

The open reading frame encoding the CaPif1 protein (NCBI number: AOW31431.1) was generated by PCR and cloned into a modified pET-21a vector with an N-terminal SUMO tag to generate His-SUMO-CaPif1. The construct was transformed into E. coli C2566H. The cells were grown in LB medium containing 100 μg/ml ampicillin at 37°C to an OD₆₀₀ of 0.7 and induced with 0.3 mM IPTG at 18°C for an additional 16 h. The cells were then harvested by centrifugation, resuspended in a buffer [20 mM Tris-HCl (pH 7.6), 500 mM NaCl], and sonicated 2–3 times to shear their DNA. All purification steps were performed at 4°C. After centrifugation and filtration using a 0.45 μm membrane, the mix was applied to a HisTrap Ni-Sepharose column (GE Healthcare) and eluted in a gradient increased to 200 mM imidazole over 15 column volumes. The His-SUMO tag was removed by SUMO enzyme. The mixture of proteins was dialyzed overnight against 20 mM Tris-HCl (pH 7.6), and 500 mM NaCl, and applied to the second round of Ni-sepharose chromatography. Finally, the purified protein fractions were concentrated to 10 mg/ml. Glycerol was added at a final concentration of 15%, and the proteins were stored at −80°C.

All DNA substrates used in the experiment were PAGE-purified and purchased from GENEWIZ laboratory (Jiangsu, China). The sequences of unlabeled or fluorescently labeled DNA are shown in Table 1. All of the labeled DNA involved in the experiment was labeled with 3′-fluorescein (3′-FAM). Double-stranded DNA substrates were obtained by mixing equimolar concentrations of complementary single-stranded oligonucleotides in 50 mM Tris-HCl (pH 7.5) and 25 mM NaCl. The mixture was heated to 95°C for 5 min, and annealing was allowed by slow cooling to room temperature. Finally, series of duplex DNA samples were stored at −40°C.

To determine the annealing activity of CaPif1, we used two partially complementary single-stranded DNA (10 nM) samples mixed with different concentrations of CaPif1 in annealing buffer [25 mM Tris-HCl (pH 7.5), 50 mM NaCl, 1.5 mM MgCl₂, 1 mM DTT]. No ATP was used because DNA annealing is usually ATP-independent, while unwinding depends on ATP hydrolysis. If the solution contains ATP, the double-stranded DNA produced by annealing will be unwound, so that annealing cannot be observed.

After incubation at 30°C for 10 min, adding 5 × stop loading buffer (150 mM EDTA, 2% SDS, 30% glycerol, and 0.1% bromophenol blue) to stop the reaction. The samples were electrophoresed on a 12% native polyacrylamide gel and visualized with ChemiDoc MP (Bio-Rad, CA, United States).

Fluorescence polarization assay: The binding of CaPif1 to DNA was monitored by fluorescence polarization assay using the SpectraMax iD3 microplate reader (Molecular Devices, LLC, PA, United States). following previously reported methods (Zhang et al., 2016). Briefly, the fluorescence-labeled DNA substrate (5 nM) was mixed with varying amounts of CaPif1 in binding buffer [25 mM Tris-HCl (pH 7.5), 50 mM NaCl, 1.5 mM MgCl₂, and 1 mM DTT] at 25°C for 10 min. Then, the fluorescence polarization was measured. The anisotropy value of each test point was averaged by three independent tests.

Electrophoretic mobility shift assay (EMSA): The 3′-fluorescein labeled DNA substrates (10 nM) were pre-incubated with varying amounts of CaPif1 in binding buffer at 25°C for 5 min. Before electrophoresis, 5 × loading buffer (60% glycerol and 0.1% bromophenol blue) was added into the mixture. The electrophoresis conditions were as follows: native PAGE (Acr∶Bis = 39∶1), 6% polyacrylamide, 100 V, 40 min. The results were visualized by a ChemiDoc MP Imaging System (Bio-Rad, CA, United States).

CaPif1 was combined with duplex DNA in unwinding buffer [25 mM Tris HCl (pH 7.5), 50 mM NaCl, 1.5 mM MgCl₂, 1 mM DTT, 5 mM ATP], and ATP was added to initiate the reaction at 30°C for 10 min. The reactions were quenched with the addition of 5 × stop loading buffer. Products of DNA unwinding reactions were resolved on native PAGE 12% (Acr∶Bis = 39∶1) gels, 100 V, 80 min. DNA in polyacrylamide gels was visualized using a ChemiDoc MP and quantitated using the Image Lab software (Bio-Rad). The percentage of unwound helicase substrate was calculated using the following Eq. 1, where P is the product (ssDNA), and S is the substrate (dsDNA): % unwinding = 100 % × P / ( S + P ) . (1)

ATPase activity was detected using a commercial ATPase assay kit according to manufacturer’s instructions (Beyotime Biotechnology, Shanghai, China). The wavelength was 540 nm, and the cumulative detection time was 10 s. The experiment was repeated three times and the percentage of ATP consumption was calculated using the following Eq. 2: % consumption = ( P − S ) / P × 100 % , (2) where P is the light value produced by the control group (without CaPif1), S is the light value produced by the experimental group (various CaPif1 concentrations), and P-S is the light value reduced by the ATP consumption of the experimental group. Luminance was detected using the SpectraMax iD3 microplate reader (Molecular Devices, LLC, PA, United States). The fitting of the combined characteristic curve was done using the software ORIGIN.

We successfully expressed and obtained recombinant CaPif1 through HisTrap affinity purification and cutting and subtracting SUMO tags in order to dissect its physiological functions (Figure 1A). Previous studies have shown that helicase will bind to the nucleic acid substrate in advance before the unwinding process (Witkiewicz-Kucharczyk and Bal, 2006). Therefore, in order to test the binding activity of CaPif1 to the nucleic acid substrate, different structures of DNA were selected: ssDNA (S40), blunt end dsDNA (DS32), and 5′- overhanged DNA (OhS22D21). We fixed three substrate concentrations at 10 nM, and CaPif1 at varying concentrations (0–3,000 nM) in binding buffer (Materials and Methods). As the concentration of CaPif1 increased, free DNA (bottom arrow in Figures 1B,C) gradually decreased; in contrast, the protein-DNA complex (up arrow in Figures 1B,C) gradually increased. Through comparison, it was found that CaPif1 with blunt end dsDNA had very weak binding activity, while the other two substrates containing single-stranded tail structures showed high affinity. When CaPif1 reached a certain high concentration, the protein-DNA complexes were found near the spot hole, indicating that the high concentration of CaPif1 may form dimer or polymer (Yu et al., 2021).

In addition, we tested whether CaPif1 had annealing activity (Figure 1D). Two partially complementary single-stranded oligonucleotides (S43 and S21 in Table 1, where S43 has been labeled with 3′-fluorescein) were incubated with different concentrations of CaPif1 (1–300 nM) in annealing buffer. As the concentration of CaPif1 increased, the annealed double-stranded DNA also increased. These results indicated that CaPif1 was competent for promoting DNA annealing.

The helicase of the Pif1 family unwinds a broad range of DNA substrates, a common feature for DNA structures is the presence of a 5′-ss tail. To probe the possible physiological functions of CaPif1 in the cell, different structured DNAs were labeled fluorescent groups (3′-FAM) to determine unwinding activity by EMSA, as shown in the figure (Figures 2A–D). Among the various DNA substrates tested, there was a common feature. CaPif1 could effectively unwind the DNA structures containing the 5′-ss tail, such as 5′-overhang (OhS22D21) and forked (YS22). Even at high concentrations (100 nM) of CaPif1, the blunt end dsDNA (DS32) structure still was not unwound. To note, the property that CaPif1 unwinds more efficiently 5′-ss tail DNA than that of Bubble (BS12) structure. The unwinding ratios are compared in (Figure 2E; Supplementary Table S1), and the sequences of these substrates are summarized in Table 1.

We further compared the effects of the 5′-ss tail and double-strand length on unwinding activity. Two sets of experiments were designed: in one set the double-strands were kept at the same length but the length of the 5′-ss tail gradually increased (OhS4D20, OhS12D20, and OhS20D20) (Figure 3A), while in the other set the 5′-ss tail length was fixed, but the double-strand length gradually increased (OhS12D12, OhS12D20, and OhS12D28) (Figure 3B). The results showed that the unwinding ratio of CaPif1 gradually decreased with the increase of the double-strand length, and the unwinding ratio of CaPif1 increased as the tail chain length increased (Figures 3C,D; Supplementary Table S2).

Helicase can usually use the energy of hydrolyzing NTPs/dNTPs to melt double-stranded DNA, but whether CaPif1 has a preference for adenosine nucleotides as an energy source. Under the same experimental conditions, the unwinding activity of CaPif1 was measured in the presence of different types of NTPs and dNTPs. The results showed that CaPif1 could efficiently unwind the DNA in the presence of ATP, dATP, CTP, and dCTP, and 80% of DNA was unwound at 100 μM (Figures 4A,B). However, in the presence of GTP, dGTP, dTTP, UTP, and dUTP, less DNA was unwound. We found that when the concentration of these low-efficiency adenosine nucleotides was increased to 5 mM, the unwinding activity of CaPif1 was inhibited (Figures 4C–E). We speculate that for translocation helicases, the unwinding is dependent on the hydrolysis of NTPs or dNTPs, which depends on the ratio of NTPs/dNTPs to Mg²⁺ (Harmon and Kowalczykowski, 2001; Choudhary et al., 2004). Increasing the concentration of NTPs/dNTPs will disrupt the ratio and lead to hydrolysis inhibition, just as in NS3 helicase (Frick et al., 2007). These results indicate that CaPif1 can use the energy from the hydrolysis of a wide range of nucleotides to unwind DNA. We calculated the proportion of unwinding DNA at 100 and 5,000 μM NTP for a clearer analysis by scanning the gray values with Image Lab software (Figure 4F) (Supplementary Table S3).

Helicases are usually fully and equally active for DNA unwinding in the presence of metal ions. Mg²⁺ and Mn²⁺ are common. Some metal ions are important for maintaining the protein structure. However, some metal ions destroy the protein structure and cause proteins to lose helicase activity. To address the influence of the metallic cofactor, we further determined the unwinding activity in the presence of various metal ions. Both Mg²⁺ and Mn²⁺ had higher unwinding activities, and the results were relatively similar, while Ca²⁺ supported CaPif1 to partially unwind the DNA duplex. In sharp contrast, in the presence of Zn²⁺, Ni²⁺, Cu²⁺, and Cd²⁺, this activity was completely lost at 1.5 mM (Figure 5A). This showed that CaPif1 was unable to use these four ions to melt double-stranded DNA.

To determine if the unwinding activity of CaPif1 was inhibited by the four ions, we first studied the reactions containing both Mg²⁺ and one of the other metal ions. We found that when CaPif1 was incubated in advance with solutions containing Zn²⁺, Ni²⁺, or Cd²⁺, even with the addition of sufficient Mg²⁺, CaPif1 could not unwind the DNA duplex, while in a reaction containing Cu²⁺, with the addition of Mg²⁺, the unwinding activity of CaPif1 could be clearly observed. The results indicate that although CaPif1 cannot use Cu²⁺ to unwind dsDNA, Cu²⁺ does not inhibit the unwinding activity of CaPif1; however Zn²⁺, Ni²⁺, and Cd²⁺ have significant inhibitory effects on the unwinding activity of CaPif1 (Figure 5B). Ni²⁺ research showed the same result as Cd²⁺, and also demonstrated strong inhibition. However, in the reaction containing Ni²⁺, a brownish-yellow precipitation occurred that made it impossible for us to make an accurate quantitative analysis of the inhibition of the unwinding activity of CaPif1 by Ni²⁺ (Supplementary Table S4.1).

We have also adjusted the different addition order of Zn²⁺/Ni²⁺/Cd²⁺, Mg²⁺, and CaPif1. Furthermore, regardless of whether Zn²⁺, Ni²⁺, or Cd²⁺ was added to reaction system first, or Mg²⁺ was added to reaction system first, or Zn²⁺/Ni²⁺/Cd²⁺ and Mg²⁺ were added at the same time, the inhibitory effects were similar for each heavy metal (Supplementary Table S4.2).

Next, we tested whether EDTA could restore the unwinding activity inhibited by Cd²⁺ and Zn²⁺. After the complete reaction of 1.5 mM Zn²⁺ or Cd²⁺ with 25 nM CaPif1 at 30°C for 5 min, we cleared the metal with 3 mM EDTA (this concentration could completely remove the divalent metals in the mixture but was not sufficient to affect subsequent experiments), and then 1.5 mM Mg²⁺ and 5 mM ATP were added to the mixture to initiate the spin reaction. The Zn²⁺ effect was abolished by adding EDTA (Figure 5C, lane 2 and lane 3), while the Cd²⁺ effect was not counteracted by EDTA (Figure 5C, lane 4 and lane 5). It is possible that Cd²⁺ can damage the active center of CaPif1 more severely at high concentrations (1.5 mM), so that even after removing the ions, the unwinding activity of CaPif1 cannot be restored. Further research showed that lower concentrations of Zn²⁺ and Cd²⁺ significantly inhibited the unwinding activity of CaPif1. When the concentrations of Zn²⁺ and Cd²⁺ were 100 and 10 μM, respectively, almost the same inhibitory effect was observed. Cd²⁺ showed stronger inhibition compared with Zn²⁺, which was at a concentration 10 times lower than that of Zn²⁺ (Figure 5D).

In addition, we sought to examine whether the Zn²⁺ and Cd²⁺ could also inhibit the ATPase and DNA binding activities of CaPif1. Our data showed that whether the binding buffer contained 100 μM of Zn²⁺ (or Cd²⁺) or not, the fluorescence anisotropy values were not significantly different (Figure 5E), indicating that Zn²⁺ and Cd²⁺ had no significant effect on the DNA binding activity of CaPif1. The ATP hydrolysis activity of CaPif1 was a function of ATPase, so the ATP concentration was detected before and after the reaction. The data showed that the ATPase activity containing Zn²⁺ and Cd²⁺ was significantly lower than that without Zn²⁺ and Cd²⁺ (Figures 5F,G). Further analysis revealed that when the concentration of Cd²⁺ reached 100 μM, the ATPase activity of CaPif1 was almost completely inhibited. Thus, we speculate that heavy metals inhibit the unwinding activity of CaPif1 by blocking its energy pathway.

Although it was difficult to recover the effect of high concentration of Cd²⁺ on CaPif1, we found that when Cd²⁺ (10 µM) and Zn²⁺ (70 µM) were at low concentrations (significantly inhibiting the unwinding of CaPif1), Cys and ATP could reduce this damage. Previous studies have reported that free sulfhydryl groups of Cys are good candidates for Cd²⁺ binding (Witkiewicz-Kucharczyk and Bal, 2006).

At the same time, Cys, histidine (His), valine (Val), and leucine (Leu) were also included to check their interaction with Cd²⁺/Zn²⁺. They were incubated with unwinding buffer containing Cd²⁺ at 30°C for 5 min. Then CaPif1 was added and incubated at 30°C for another 5 min, before ATP was added to start the reaction. We found that with increasing Cys concentrations, the inhibition of the unwinding activity of CaPif1 by Cd²⁺ decreased significantly. This may have been due to Cys binding to the Cd²⁺ in the reaction, so the concentration of free Cd²⁺ was low, which reduced the damage to the activity of CaPif1 (Figure 6A). None of the four amino acids significantly alleviated damage in samples containing Zn²⁺ (Figure 6B).

We changed the order that samples were added in. For example, CaPif1 was first pre-incubated with the solution containing 10 µM Cd²⁺, and then 10 mM Cys was added. This process still showed a reduction of the inhibition of CaPif1’s unwinding activity by Cd²⁺ (Supplementary Table S5). This may imply that during the unwinding process of CaPif1, increasing the concentration of Cys to a certain level can prevent the inhibitory effect of Cd²⁺.

Finally, we increased the concentration of ATP to alleviate the inhibition of CaPif1’s unwinding activity by low concentrations of Zn²⁺ and Cd²⁺. The two experiments showed different results. Increasing ATP concentration obviously reduced the inhibition of Zn²⁺, while Cd²⁺ did not resume the unwinding activity of CaPif1 to an obvious degree (Figure 6C). The possible reason may be that the CaPif1 unwinding activity is ATP-dependent. Therefore, increasing the ATP concentration within a certain range can increase the activity of CaPif1, showing an apparent recovery of CaPif1 activity. The damage caused by Cd²⁺ is more serious than that caused by Zn²⁺, so it may not be able to offset this damage through the increase of ATP.

DTT has been previously described having Cd²⁺-induced cell damage and enzyme inhibition (Stacey, 1986; Ahammadsahib et al., 1989). All the results mentioned above were obtained under reducing conditions (in the presence of DTT). We first fixed the concentration of CaPif1 in the reaction without Zn²⁺ and Cd²⁺, and increased the concentration of DTT. The results showed that DTT had no effect on the unwinding activity of CaPif1 (Figure 7A). Then we compared the unwinding activity of CaPif1 in the presence of 1.5 mM Zn²⁺ or 100 μM Cd²⁺ and gradually increased the concentration of DTT. In the reaction without DTT, neither Zn²⁺ nor Cd²⁺ could inhibit the unwinding activity of CaPif1 (Figures 7B,C). This implied that CaPif1 resisted the inhibitory effect of Zn²⁺ and Cd²⁺ in less reducing environments. We speculate that when there is a certain concentration of DTT in the environment, the disulfide bond of the protein is opened, thereby releasing more sulfhydryl groups on the surface of the protein, and Zn²⁺/Cd²⁺ can bind tightly to the sulfhydryl group of cysteine, resulting in more heavy metals destroying the activity of the protein. Interestingly, the reducing status in the body will decrease in the presence of Cd²⁺, which stimulates ROS production in the body (Srinivas et al., 2019). We speculate that CaPif1 has formed a mechanism to respond to heavy metal poisoning by effectively sensing the reducing status changes.