We sought to implement I-PpoI for in crystallo metal ion soaking (Figure 1B), which has been successful in elucidating the catalytic mechanisms of DNA polymerases (Freudenthal et al., 2013; Gao and Yang, 2016; Nakamura et al., 2012; Vyas et al., 2015; Chim et al., 2021; Gregory et al., 2021), nucleases (Samara and Yang, 2018; Wu et al., 2019; Freudenthal et al., 2015; Whitaker et al., 2018), and glycosylase (Demir et al., 2023). First, a complex of I-PpoI and a palindromic DNA was crystalized at pH 6 with 0.2 M sodium malonate. Similar to previous studies, a dimer of I-PpoI was found in the asymmetric unit, with both active sites engaged for catalysis and DNA in the middle bent by 55° (Figure 1—figure supplement 1A). The two I-PpoI molecules were almost identical and thus served as internal controls for evaluating the reaction process (Figure 1—figure supplement 1B). Within each I-PpoI active site, a water molecule (nucleophilic water, WatN) existed 3.6 Å from the scissile phosphate, near the imidazole side chain of His98. On the leaving group side of the scissile phosphate, the metal exhibited an octahedral geometry, being coordinated by three water molecules, two oxygen atoms from the scissile phosphate, and the conserved asparagine (Figure 1C).

Next, we removed malonate in the crystallization buffer, which may chelate metal ions and hinder metal ion diffusion (Deerfield et al., 1991), by equilibrating the crystals in 200 mM NaCl buffer at pH 6, 7, or 8 for 30 min. The diffraction quality of the crystals was not affected during the soaking. The structures of I-PpoI equilibrated at pH 6, 7, or 8 in the presence of 200 mM NaCl showed no signs of product formation and appeared similar to the malonate and previous reported I-PpoI structures (Galburt et al., 1999; Flick et al., 1998; Figure 1—figure supplement 3A). To confirm if a monovalent metal ion binds in the pre-reaction state, we soaked the I-PpoI crystals in buffer containing Tl⁺ and detected anomalous electron density at the metal ion-binding site without product formation (Figure 1—figure supplement 3B). Our results support that the monovalent metal ion can bind within the active site without initiating reaction, in corroboration with our biochemical assays (Figure 1—figure supplement 3C).

To initiate the chemical reaction in crystallo, we transferred the I-PpoI crystals equilibrated in NaCl buffer to a reaction buffer with 500 µM Mg²⁺ (Figure 1B). After 600 s soaking, we saw a significant negative F_o − F_c peak on the leaving group side of the scissile phosphate atom as well as a positive F_o − F_c peak on the other side (Figure 1—figure supplement 3D), indicating that DNA hydrolysis was occurring in crystallo. After DNA cleavage, the newly generated phosphate group shifted 1 Å toward the WatN (Figure 1D). Furthermore, an additional product state was observed after soaking the crystals in 500 µM Mg²⁺ for 600 and 1800 s, at which the newly formed phosphate shifted 3 Å away from the metal ion to form a hydrogen bond with Arg61 (Figure 1E). Our results confirm that the implementation of I-PpoI with in crystallo Me²⁺ soaking is feasible for observing the I-PpoI catalytic process and dissecting the mechanism of His-Me nucleases.

With an established in crystallo reaction system, we next monitored the reaction process by soaking I-PpoI crystals in buffer containing 500 µM Mg²⁺ pH 7, for 10–1200 s. Density between the reactant phosphate and nucleophilic water increased along with longer soaking time, indicating the generation of phosphate products (Figure 2A). During the reaction, the coordination distances of the metal ion ligands decreased 0.1–0.2 Å when Mg²⁺ exchanged with Na⁺, which is consistent with their preferred geometry (Cowan, 2002; Figure 2B, C). At reaction time 160 s, 45% of product had been generated (Figure 2D), which later plateaued to 65% at 300 s. During the reaction process, the sugar ring of the reactant and product DNA that resides around 3 Å away from the scissile phosphate, remained in a C3′-endo conformation. As Mg²⁺ soaking time increased from 10 to 160 s, we observed the WatN approaching the scissile phosphate (Figure 2E). At the same time, the conserved His98 sidechain proposed to deprotonate the WatN was slightly rotated.

Throughout the reaction process with Mg²⁺, we found that the electron density for the Mg²⁺ metal ion strongly correlated (R² = 0.97) with product formation (Figure 3—figure supplement 1A, C), which may suggest that saturation of this metal ion site is required and sufficient for catalysis. However, Mg²⁺’s similar size to Na⁺ makes it suboptimal for quantifying metal ion binding. To thoroughly investigate metal ion dependence, we repeated the in crystallo soaking experiment with Mn²⁺, which is more electron rich and can be unambiguously assigned based on its electron density and anomalous signal. With 500 µM Mn²⁺ in the reaction buffer, we found that the reaction process and the product conformation were similar to that for Mg²⁺. After 160 s Mn²⁺ soaking, clear anomalous signal was present at the metal ion-binding site, confirming the binding of Mn²⁺ (Figure 3—figure supplement 1F). For crystal structures of I-PpoI with partial product formation, Mn²⁺ signal at the metal ion site correlated with product formation with a R² of 0.98 (Figure 3A and Figure 3—figure supplement 1B). To further search if additional and transiently bound divalent ions participate in the reaction, we soaked the I-PpoI crystals in high concentration of Mn²⁺ (200 mM) for 600 s, at which 80% product formed within the active site. However, apart from the single metal ion-binding site, we do not detect anomalous signal for additional Mn²⁺, despite such high concentration of Mn²⁺ (Figure 3B and Figure 3—figure supplement 2). The strong correlation between product phosphate formation with Mn²⁺ binding and the absence of additional anomalous density peaks in heavy Mn²⁺ soaked crystals suggest that one and only one divalent metal ion is involved in I-PpoI DNA hydrolysis.

To examine if additional monovalent metal ions participate during DNA hydrolysis, we titrated Tl⁺, which can replace monovalent Na⁺ or K⁺ and yields anomalous signal (Auffinger et al., 2016; Kiser et al., 2009), at up to 100 mM concentration in the biochemical assay. Apart from precipitation that occurred at 50 and 100 mM Tl⁺, we found that product conversion remained unaffected with increasing Tl⁺ in solution (Figure 3—figure supplement 1G). Similarly, increasing Na⁺ concentration in solution did not increase product conversion (Figure 3—figure supplement 1H). In addition, only one Tl⁺ was detected by its anomalous signal in the pre-reaction state (Figure 1—figure supplement 3B). The biochemical and structural results indicate that additional monovalent metal ion may not be necessary for DNA hydrolysis. However, crystal deterioration limited us from soaking the I-PpoI crystals in high concentration of Tl⁺ for in crystallo reaction.

During DNA hydrolysis, deprotonation of the WatN is required for the nucleophilic attack and phosphodiester bond breakage. This proton transfer has been proposed to be mediated by the highly conserved His98 (Galburt et al., 1999; Flick et al., 1998) that lies within 3 Å from the WatN. We speculated that the ability of His98 to activate the nearby WatN and mediate proton transfer would be affected by pH. The DNA cleavage assay revealed that I-PpoI cleavage activity increased with pH with a pKa of 8.3 (Figure 4A), which is much higher than the pKa of histidine, but the histidine pKa and the proton transfer process may be affected by the active site environment. To explore how pH affects the active site configuration and catalysis, we conducted in crystallo soaking experiments with Mg²⁺ at pH 6 and 8 in addition to the pH 7 data series in Figure 2. Consistent with in solution experiments, higher pH resulted in faster product formation in crystallo (Figure 4B), indicating that metal ion binding may also be affected by pH. To test this, we performed Mg²⁺ titration at different pH in solution. Our results showed that over 100 times higher concentration of Mg²⁺ was needed to yield 50% product in pH 6 versus 8 (Figure 4C), confirming that pH affects metal-ion-dependent I-PpoI catalysis. Likewise, the reactions in crystallo at higher pH reached 50% product formation at shorter soaking times (320 s at pH 6, 160 s at pH 7, and 80 s at pH 8, respectively). Interestingly, the crystal structures that contained 30–35% product at pH 6 and 8 were nearly identical (Figure 4D, E). The positions of the His98 residue and the WatN were practically superimposable (Figure 4F). At all pH, Mg²⁺ binding strongly correlated with product formation (R² > 0.95), suggesting that low pH reduces the overall reaction rate without altering the reaction pathway (Figure 3—figure supplement 1D, E).

We next investigated the deprotonation pathway with mutagenesis. Because His98 has been proposed to primarily activate the nucleophilic water, we first mutated His98 to alanine (Figure 5A). As expected, cleavage activity of H98A I-PpoI drastically dropped. Our assays showed that H98A I-PpoI displayed residual activity but required a reaction time of 1 hr to be comparable to WT I-PpoI (Figure 5B), similar to the H98Q mutant in previous studies (Eastberg et al., 2007). Furthermore, varying the pH resulted in a sigmoidal activity curve of I-PpoI pH dependence, corresponding to a pKa of 7.9 (Figure 5C). The significantly reduced reaction rate and the shift of pH dependence suggest that His98 plays a key role in pH sensing and water deprotonation. On the other hand, the low but existing activity of H98A I-PpoI suggests the presence of alternative general bases for proton transfer. Another histidine (His78) resides on the nucleophile side 3.8 Å from the WatN, with a cluster of water molecules in between (Figure 5A). We hypothesized that His78 may substitute His98 as the proton acceptor. The DNA cleavage assays revealed that the single mutant, H78A I-PpoI, had an activity similar to WT, whereas the double mutant (H78A/H98A I-PpoI) exhibited much lower activity than that of H98A (Figure 5B). Furthermore, varying the pH for H78A/H98A I-PpoI resulted in a sigmoidal activity curve that corresponded to a pKa of 8.7 (Figure 5C). In consistent with previous speculations, our results confirm the possibility of His78 as an alternative general base (Eastberg et al., 2007). The residual activity and pH dependence of H78A/H98A I-PpoI indicated that something else was still activating the nucleophilic water in the absence of any nearby histidine. Furthermore, we found that titrating imidazole in H98A and H78A/H98A I-PpoI partially rescued cleavage activity (Figure 5—figure supplement 1A, B), similar as observed in Cas9 and EndA nuclease (Furuhata and Kato, 2021; Moon et al., 2011). Collectively, our results indicate that His98 is the primary proton acceptor like previous simulation results (Maghsoud et al., 2023) but at the same time, I-PpoI can use alternative pathways to activate the nucleophilic WatN.

We next sought to understand the mechanism of His98-promoted hydrolysis from a structural standpoint. In our in crystallo soaking experiments, we equilibrated H98A I-PpoI crystals in 500 µM Mg²⁺. The structure looked nearly identical to the Na⁺ structure and previous H98A I-PpoI structure with Mg²⁺. As shown earlier, the coordination environment of Na⁺ and Mg²⁺ was quite similar. To confirm divalent metal ion binding, the H98A I-PpoI crystals were soaked for 1800s in 500 µM Mn²⁺, the same concentration used to initiate DNA hydrolysis by WT I-PpoI. However, to our surprise, the metal ion-binding site was devoid of any anomalous signal. Increasing the Mn²⁺ concentration to 1 mM still did not produce anomalous density at the Me²⁺-binding site (Figure 5D and Figure 5—figure supplement 1C). The results indicate that metal ion binding can be altered by perturbing His98 and possibly water deprotonation, even though the metal ion-binding site and His98 exist 7 Å apart without direct interaction. Although we tried soaking the H98A I-PpoI crystals in 1 mM Mg/Mn²⁺ and 100 mM imidazole for 15 h, metal ion binding, imidazole binding, or product formation were not detected (Figure 5—figure supplement 1D), possibly due to the difficulty of imidazole diffusion within the lattice. Our results indicate that perturbing the deprotonation pathway not only affects nucleophilic attack but also metal binding, suggesting that I-PpoI catalyzes DNA catalysis via a concerted mechanism.

WT, His78Ala, His98Ala, and H78A/H98A physarum polycephalum I-PpoI (residues 1–162) were cloned into a modified pET28p vector with a N-terminal 6-histidine tag and a PreScission Protease cleavage site. For protein expression, this I-PpoI plasmid was transformed into BL21 DE3 E. coli cells, which were grown in a buffer that contained (10 g/l glucose, 40 g/l α-lactose, 10% glycerol) for 24 hr (20°C). The cell paste was collected via centrifugation and re-suspended in a buffer that contained 50 mM Tris (pH 7.5), 1 M NaCl, 1 mM MgCl₂, 10 mM imidazole, 2 mM β-mercaptoethanol (BME), and 5% glycerol. After sonification, I-PpoI was loaded onto a HisTrap HP column (GE Healthcare), which was pre-equilibrated with a buffer that contained 50 mM Tris (pH 7.5), 1 M NaCl, 1 mM MgCl₂, 10 mM imidazole, 2 mM BME, and 5% glycerol. The column was washed with 300 ml of buffer to remove non-specific bound proteins and was eluted with buffer that contained 50 mM Tris (pH 7.5), 1 M NaCl, 1 mM MgCl₂, 300 mM imidazole, and 2 mM BME. The eluted I-PpoI was incubated with PreScission Protease to cleave the N-terminal 6-histidine-tag. Afterwards, I-PpoI was desalted to 50 mM Tris (pH 7.5), 167 mM NaCl, 1 mM MgCl₂, 2 mM BME, and 5% glycerol and was loaded onto a Heparin column (GE Healthcare) equilibrated with 50 mM Tris (pH 7.5) and 167 mM NaCl. The protein was eluted with an increasing salt (NaCl) gradient, concentrated, and stored at 40% glycerol at −80°C.

DNA hydrolysis activity of varying time was assayed by the following: The reaction mixture contained 100 nM WT I-PpoI, 50 mM NaCl, 100 mM Tris (pH7.5), 1.5 mM dithiothreitol (DTT), 0.05 mg/ml bovine serum albumin (BSA), 50 nM DNA, 10 µM ethylenediaminetetraacetic acid (EDTA), and 4% glycerol. The hydrolysis assays were executed using a palindromic 5′-fluorescein-labeled DNA duplex (5′-TTG ACT CTC TTA AGA GAG TCA-3′). Reactions were initiated by adding 10 mM MgCl₂ to the reaction mixture for 0–1 hr at 37°C and were stopped by mixing with equal volume of a quench buffer, which contain 80% formamide, 100 mM EDTA (pH 8.0), 0.2 mg/ml xylene cyanol, and 0.2 mg/ml bromophenol.

The DNA hydrolysis activity at different pH was assayed by the following: The reaction mixture contained 100–3000 nM WT, H98A, and H78A/H98A I-PpoI, 50 mM NaCl, 1.5 mM DTT, 0.05 mg/ml BSA, 50 nM DNA, 10 µM EDTA, and 4% glycerol. Reactions were initiated by adding 50 mM 2-(N-morpholino)ethanesulfonic acid (MES) (pH 5.5–6.5), 50 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) (pH 6.5–7.5), or 50 mM Tris (pH 7.5–9.5) together with 10 mM MgCl₂ to the reaction mixture for 30 min at 37°C and were stopped by adding equal volume of quench buffer.

The DNA hydrolysis activity with different mutants was assayed by the following: The reaction mixture contained 50 mM NaCl, 100 mM Tris (pH7.5), 1.5 mM DTT, 0.05 mg/ml BSA, 50 nM DNA, 10 mM MgCl₂, 10 µM EDTA, and 4% glycerol. Reactions were initiated by adding 0–3000 nM of WT, H78A, H98A, H78A/H98A I-PpoI to the reaction mixture for 15 s or 1 hr at 37°C and stopped by adding equal volume of quench buffer.

The DNA hydrolysis activity with different metal ions was assayed by the following: The reaction mixture contained 100 nM WT I-PpoI, 50 mM NaCl, 100 mM Tris (pH 7.5), 1.5 mM DTT, 0.05 mg/ml BSA, 50 nM DNA, 10 µM EDTA, and 4% glycerol. Reactions were initiated by adding 10 mM of MgCl₂, MnCl₂, CaCl₂, NiCl₂, and ZnCl₂ to the reaction mixture for 15 s at 37°C and were stopped by adding equal volume of quench buffer.

The DNA hydrolysis activity with Tl⁺ or additional Na⁺ was assayed by the following: The reaction mixture contained 100–3000 nM WT I-PpoI, 100 mM Tris (pH7.5), 1.5 mM DTT, 0.05 mg/ml BSA, 50 nM DNA, 10 µM EDTA, and 4% glycerol. Reactions were conducted at 37°C for 30 min by adding 0–350 mM TlCl or NaCl together with 10 mM MgCl₂ to the reaction mixture and were stopped by adding equal volume of quench buffer.

The DNA hydrolysis activity with imidazole was assayed by the following: The reaction mixture contained 100–3000 nM H98A and H78A/H98A I-PpoI, 50 mM NaCl, 100 mM Tris (pH7.5), 1.5 mM DTT, 0.05 mg/ml BSA, 50 nM DNA, 10 µM EDTA, and 4% glycerol. Reactions were conducted at 37°C for 30 min by adding 0–100 mM imidazole together with 10 mM MgCl₂ to the reaction mixture and were stopped by adding equal volume of quench buffer.

For all reactions, after heating the quenched reaction mix to 97°C for 5 min and immediately placing on ice, reaction products were resolved on 22.5% polyacrylamide urea gels. The gels were visualized by a Sapphire Biomolecular Imager and quantified using the built-in software. Quantification of percentage cleaved and graphic representation were executed by Graph Prism.

WT or H98A I-PpoI in a buffer containing 20 mM Tris 7.5, 300 mM NaCl, 3 mM DTT, and 0.1 mM EDTA was added with (5′-TTG ACT CTC TTA AGA GAG TCA-3′) DNA at a molar molar of 1:1.5 for I-PpoI and DNA and added with threefolds volume of buffer that contained 20 mM Tris 7.5, 3 mM DTT, and 0.1 mM EDTA. This I-PpoI–DNA complex was then cleaned with a Superdex 200 10/300 GL column (GE Healthcare) with a buffer that contained 20 mM Tris 7.5, 150 mM NaCl, 3 mM DTT, and 0.1 mM EDTA. The I-PpoI–DNA complex was concentrated to 2.8 mg/ml I-PpoI (confirmed by Bradford assay). All crystals were obtained using the hanging-drop vapor-diffusion method against a reservoir solution containing 0.1 M MES (pH 6.0), 0.2 M sodium malonate, and 20% (wt/vol) PEG3350 at room temperature within 4 days.

To identify the monovalent Me⁺ species that binds during the pre-reaction state, WT I-PpoI crystals were transferred and incubated in a buffer containing 0.1 M MES (pH 6.0), 70 mM thallium acetate and 20% (wt/vol) PEG3350 for 30 min. Afterwards, the crystals were quickly dipped in a cryo-solution supplemented with 20% (wt/vol) glycerol and flash-cooled in liquid nitrogen.

The WT I-PpoI crystals were first transferred and incubated in a pre-reaction buffer containing 0.1 M MES (pH 6.0 or 7.0) or 0.1 M Tris (pH 8.0), 0.2 M NaCl, and 20% (wt/vol) PEG3350 for 30 min. The chemical reaction was initiated by transferring the crystals into a reaction buffer containing 0.1 M MES (pH 6.0 or 7.0) or 0.1 M Tris (pH 8.0), 0.2 M NaCl, and 20% (wt/vol) PEG3350, and 500 µM MgCl₂ or MnCl₂. After incubation for a desired time period, the crystals were quickly dipped in a cryo-solution supplemented with 20% (wt/vol) glycerol and flash-cooled in liquid nitrogen.

To observe any additional Me²⁺ binding sites during DNA hydrolysis, WT I-PpoI crystals were first transferred and incubated in a pre-reaction buffer containing 0.1 M MES (pH 6.0), 0.2 M NaCl, and 20% (wt/vol) PEG3350 for 30 min. The chemical reaction was initiated by transferring the crystals into a reaction buffer containing 0.1 M MES (pH 6.0), 0.2 M NaCl, and 20% (wt/vol) PEG3350, and 200 mM MnCl₂. After incubation for 600 s, the crystals were quickly dipped in a cryo-solution supplemented with 20% (wt/vol) glycerol and flash-cooled in liquid nitrogen.

The metal ion soaking experiments with His98Ala I-PpoI were performed following the similar protocol as that of WT I-PpoI. His98Ala I-PpoI crystals were first incubated in a pre-reaction buffer containing 0.1 M MES 7.0, 0.2 M NaCl, and 20% (wt/vol) PEG3350 for 30 min, followed by 1800s incubation in a reaction buffer containing 0.1 M MES 7.0, 0.2 M NaCl, and 20% (wt/vol) PEG3350, and 1 mM MnCl₂. To observe whether soaking in imidazole can initiate the reaction in the absence of His98, His98Ala I-PpoI crystals were first transferred and incubated in a buffer containing 0.1 M MES (pH 6.0), 0.2 M NaCl, 1 mM MnCl₂, and 20% (wt/vol) PEG3350 for 30 min. The crystals were then transferred into a reaction buffer containing 0.1 M MES (pH 6.0), 0.2 M NaCl, 1 mM MgCl₂ or MnCl₂, and 100 mM imidazole, and 20% (wt/vol) PEG3350. After incubation for a desired time period, the crystals were quickly dipped in a cryo-solution supplemented with 20% (wt/vol) glycerol and flash-cooled in liquid nitrogen.

Diffraction data were collected at 100 K on LS-CAT beam lines 21-D-D, 21-ID-F, and 21-ID-G at 1.1 or 0.97 Å at the Advanced Photon Source (Argonne National Laboratory) or beamlines 5.0.3 at 0.97 Å at Advanced Light Source (ALS). Data were indexed in space group P3₁21, scaled with XSCALE and reduced using X-ray Detector Software (XDS) (Kabsch, 2010). Isomorphous I-PpoI structures with Na⁺ PDB ID 1CZ0 were used as initial models for refinement using PHENIX (Adams et al., 2010) and COOT (Emsley et al., 2010).

Occupancies were assigned for the reaction product until there were no significant F_o − F_c peaks. Occupancies were assigned for the metal ions, following the previous protocol (Gao and Yang, 2016) until (1) there were no significant F_o − F_c peaks, (2) the B value had roughly similar values to its ligand (3) it matched the occupancy of the reaction product. For the structures in which some F_o − F_c peaks were present around the Me²⁺ binding sites or reaction product, no change in the assigned occupancy was executed when a 10% change in occupancy (e.g. 100–90%) failed to significantly change the intensity of the F_o − F_c peaks. Source data of the electron densities in r.m.s. density are provided as a Source Data file. Each structure was refined to the highest resolution data collected, which ranged between 1.42 and 2.2 Å. Software applications used in this project were compiled and configured by SBGrid (Morin et al., 2013). Source data of data collection and refinement statistics are summarized in Supplementary file 1. All structural figures were drawn using PyMOL (https://www.pymol.org/).

Electron density (r.m.s.d.) from the F_o − F_c map of the product phosphate and Me²⁺ were calculated by running a round of B-factor refinement in PHENIX after omitting the reaction product phosphate and Me²⁺ atoms from phase calculation in COOT. All structures from the same experiment (Mg²⁺-pH 6.0, Mg²⁺-pH 7.0, Mg²⁺-pH 8.0, Mn²⁺-pH 6.0) were refined to the lowest resolution in the same experiment group (1.80 Å for Mg²⁺-pH 6.0, 1.59 Å for Mg²⁺-pH 7.0, 1.70 Å for Mg²⁺-pH 8.0, and 1.80 Å for Mn²⁺-pH 6.0).