Structural insights into RNase H catalytic mechanism from room-temperature X-ray and neutron crystallography of apo- and RNA/DNA hybrid-bound enzyme

TL;DR

This study uses advanced imaging techniques to reveal how RNase H enzymes cut RNA strands, which is important for developing new drugs and gene therapies.

Contribution

The study provides structural insights into the RNase H catalytic mechanism using room-temperature neutron and X-ray crystallography.

Findings

Neutron structures show E109 in the DEDD motif changes protonation states, suggesting its role in RNA cleavage.

RNA/DNA hybrids bind to BhRNase H1 to mimic product and Michaelis complexes depending on metal site occupancy.

Protonation of the leaving O3’ group of RNA is proposed to occur via the side chain of E109.

Abstract

RNase H enzymes are sequence-nonspecific endonucleases that cleave RNA strands in RNA/DNA hybrid duplexes, an enzymatic process essential in DNA replication and repair in both prokaryotes and eukaryotes. Also, RNase H activity of the reverse transcriptase in human immunodeficiency viruses (HIV-1 and HIV-2) is indispensable for the viral replication cycle. RNase H enzymes play an central role in the development of gene therapies and are targets for novel antivirals. It is therefore of great importance to gain a detailed understanding of the RNase H catalytic mechanism to improve drug design. We utilized Bacillus halodurans RNase H1 (BhRNase H1) to shed light on its function and catalytic mechanism. Room-temperature neutron crystallography of the wild-type and inactive D132N mutant enzymes revealed that E109, belonging to the catalytic DEDD motif, can change its protonation state, allowing us to propose its role in the protonation of the leaving O3′ hydroxyl group of RNA. X-ray crystallography has demonstrated the ability of the RNA/DNA duplex to slide along the protein surface upon metal ion binding at site MA, transforming a product mimic into a Michaelis-like complex, which confirms an essential role of the MA metal ion in catalysis.

Image 1

•Neutron structures of apo-wild-type and apo-D132N mutant BhRNase H1 map protonation states in the enzyme active site.•E109 from the catalytic DEDD motif changes its protonation state between apo-wild-type and apo-D132N structures.•RNA/DNA hybrids bind to BhRNase H1 to mimic product and Michaelis complexes depending on the metal site occupancy.•Protonation of the leaving O3’ group of cleaved RNA is proposed to occur from O2’ through the side chain of E109.

Neutron structures of apo-wild-type and apo-D132N mutant BhRNase H1 map protonation states in the enzyme active site.

E109 from the catalytic DEDD motif changes its protonation state between apo-wild-type and apo-D132N structures.

RNA/DNA hybrids bind to BhRNase H1 to mimic product and Michaelis complexes depending on the metal site occupancy.

Protonation of the leaving O3’ group of cleaved RNA is proposed to occur from O2’ through the side chain of E109.

X-ray and neutron structures of BhRNase H1 in the apo-form and in complex with RNA/DNA hybrids are described.

Figures

Proposed proton transfer pathway to protonate the O3’ leaving group by O2′-OH assisted by E109 and the subsequent re-protonation of O2’ through a Grotthuss mechanism involving a chain of H-bonded water molecules and enzyme residues. Proton hopping is depicted by blue curved arrows. All distances are given in Angstrom.Fig. 5

Full text

Introduction

1

Ribonuclease H (RNase H) enzymes belong to the class of sequence-nonspecific endonucleases that cleave RNA strands in RNA/DNA hybrid duplexes. RNases H are found in all domains of life from bacteria to humans, and in some viruses. The cellular function of RNases H is to remove RNA primers from Okazaki fragments and to process R-loops required in DNA replication and repair processes (Kogoma and Foster; Broccoli et al., 2004). In bacteria, Archaea and yeast, the RNase H function is redundant allowing the RNase H knockout organisms to survive (Arudchandran et al., 2000). In sharp contrast, RNase H knockout mice die in embryo, whereas defective RNase H function in humans leads to autoimmune diseases such as lupus, genetic disorders such as Aicardi-Goutieres syndrome, and cancer (Cerritelli et al., 2003; Crow et al., 2006; Tadokoro and Kanaya, 2009). In addition, RNase H activity encoded as part of the reverse transcriptase in human immunodeficiency viruses (HIV-1 and HIV-2), converting the viral single-stranded genomic RNA into the double-stranded proviral DNA, is indispensable for the HIV replication cycle (Tisdale et al., 1991; Beilhartz and Götte, 2010; Menendez-Arias et al., 2017). A detailed understanding of the cellular functions and catalytic mechanism of RNase H, playing a key role in antisense oligonucleotide-mediated degradation of RNA, is of great importance to improve the design of short interfering RNAs (siRNAs) in gene therapy (Bennett and Swayze, 2010; Plashkevych et al., 2017; Mutisya et al., 2017; Hu et al., 2020; Pallan et al., 2023) and of novel antiviral drugs against HIV-1, hepatitis B, and other viruses (Ponzar et al., 2022; Tramontano et al., 2019).

RNase H is a metalloenzyme, utilizing two divalent metal ions for the RNA phosphodiester bond hydrolysis (Fig. 1A). The metal ions are coordinated by four conserved carboxylates in the active site arranged in a motif called DEDD and occupy the metal binding sites denoted as M_A_ and M_B_. The DEDD motif is prevalent in the two main RNase H classes – RNase H1 and H2 types. Moreover, some bacteria and Archaea have a third type – RNase H3 – that uses a slightly different DEDE motif for the metal binding (Hyjek et al., 2019). A general two-metal catalysis mechanism was first proposed by Steitz and Steitz (1993), although a one-metal catalysis may be possible in some enzymes, such as in E. coli RNase H (Oda et al., 1993; Keck et al., 1998). Normally, Mg^2+^ is the physiological metal, with Mn^2+^ being as good as Mg^2+^ (Keck et al., 1998; Goedkin and Marqusee, 2001; Nowotny et al., 2005; Nowotny and Yang, 2006). Other transition metal ions, such as Co^2+^, Zn^2+^, Ni^2+^ and Cd^2+^, can also support RNA cleavage but at a lower level (Uchiyama et al., 1994a, Uchiyama et al., 1994b; Keck et al., 1998). Conversely, the alkaline-earth Ca^2+^ essentially inhibits the RNase H activity (Rosta et al., 2014). Interestingly, recent time-lapse X-ray crystallographic studies of Bacillus halodurans RNase H1 (BhRNase H1) (Samara and Yang, 2018) have demonstrated that catalysis may be assisted by transient binding of a third Mg^2+^ and a monovalent metal cations in the active site at different stages of the phosphodiester hydrolysis reaction, even though earlier theoretical calculations interpreted the third metal binding as inhibitory (Palermo et al., 2015). In BhRNase H1 enzyme, the amino acid residues making up the DEDD motif are D71, E109, D132, and D192 (Fig. 1B). M_A_ is located on the protein surface and is coordinated by D71, E188, D192 and two water molecules. In a Michaelis-like complex between the inactive D132N mutant and an RNA/DNA hybrid (Nowotny et al., 2005), M_A_ is also bound to the non-bridging pro-S_P_ oxygen of the scissile phosphate completing the proper octahedral coordination sphere. M_B_ is positioned deeper in the protein interior and is coordinated by D71, E109, D132, pro-S_P_ oxygen, and O3′ of the scissile bond in an irregular geometry. Mutagenesis studies confirmed the essential roles of the DEDD motif residues in the catalytic activity of RNase H enzymes as their substitutions inactivate the enzyme (Kanaya et al., 1990; Haruki et al., 1994; Wu et al., 2001; Nowotny et al., 2005; Nowotny and Yang, 2006).Fig. 1Three-dimensional structure of BhRNase H1 bound to an RNA/DNA hybrid. (A) Cartoon representation of BhRNase H1 complexes with a substrate nucleic acid duplex representing a Michaelis-like complex (PDB ID 1ZBI). (B) A close-up view of the active site. (C) Schematic representation of the active site. The proposed nucleophilic water is colored red. The dashed red arrow depicts the nucleophilic attack on the phosphorus atom. Arrows colored magenta indicate possible proton transfer pathways for the nucleophile deprotonation and the leaving group's O3′ protonation.Fig. 1

The RNA phosphodiester bond hydrolysis is initiated by a nucleophilic attack of a hydroxide anion, generated either prior to the reaction or through a catalytic water deprotonation, on the nucleic acid backbone phosphorus. The reaction then proceeds with the O3′–P scissile bond cleavage to give the final products – the 3′-nonphosphorylated and 5′-phosphorylated RNA chains. In principle, this reaction can proceed according to dissociative, associative, and concerted mechanisms (Hengge, 2005). In the dissociative mechanism designated D_N_ + A_N_ or S_N_1 by IUPAC, the O3′–P scissile bond breaks first to produce a stable intermediate metaphosphate-like ion that is subsequently attacked by an incoming nucleophile. In the associative mechanism (A_N_ + D_N_), a metastable pentacoordinated phosphorane intermediate is formed after the nucleophile attacks the phosphorus atom. In the concerted mechanism, A_N_D_N_ or S_N_2, the reaction proceeds through a trigonal bipyramidal pentacoordinated transition state, in which fission of O3′–P bond to the leaving group and formation of a P–O covalent bond to the nucleophile are concurrent. Importantly, in the S_N_2 mechanism no stable or metastable intermediates are formed.

RNase H-catalyzed chemistry is governed by the general acid-base catalysis. There is consensus that the nucleophile at physiological pH is a metal-activated catalytic water molecule that is deprotonated by a general base located nearby, whereas a hydroxide anion is unlikely to be present in the active site at pH of ∼7.4. However, neither the general base accepting a proton from the catalytic water nor the general acid donating a proton to the leaving group's O3′ hydroxyl have been identified. Theoretical studies using quantum mechanics/molecular mechanics calculations and molecular dynamics simulations disagree on the specific mechanistic details of these proton transfer pathways, although the calculated energy barriers of the overall phosphodiester hydrolysis reaction agree reasonably well with the measured reaction rate (De Vivo et al., 2008; Rosta et al., 2011; Palermo et al., 2015). Several proton transfer pathways have been considered for catalytic water deprotonation and O3′ protonation (Fig. 1C). The catalytic water can be deprotonated through the direct proton transfer to a nonbridging phosphate oxygen in the downstream nucleotide or to a neighboring E188 carboxylate that is not bound to a metal, or through a proton transfer via the Grotthus mechanism involving another water molecule to the pro-R_P_ oxygen attached to the phosphorus of the scissile bond. The latter oxygen could then transfer the proton directly to the leaving group O3′. Conversely, O3′ can abstract a proton from O2′ located on the same ribose ring or accept a proton from the protonated D132, but protonation of D132 bound to a metal center would be highly unusual because the metal's 2+ charge would significantly lower the pKa of its carboxylic group. Therefore, to fully map the reaction mechanism of RNA hydrolysis catalyzed by RNase H, it is crucial to experimentally determine the protonation states in the enzyme active site and to follow proton transfers along the catalytic reaction pathway. Protonation states are defined by the locations of hydrogen (H) atoms. A powerful experimental technique capable of determining H atom positions is neutron macromolecular crystallography (Niimura and Podjarny, 2011; Kono et al., 2022; Hjorth-Jensen and Budayova-Spano, 2024). Neutron scattering of H, and its isotope deuterium (D), is similar to those of C, N, and O, because neutrons are scattered by the atomic nuclei rather than by electron clouds as do X-rays, and the neutron scattering is independent of the atomic number, i.e. the number of electrons present in a specific atom. Hence, H and D positions can be accurately resolved in neutron crystal structures to reveal new insights for enzyme catalytic mechanisms and to guide drug design (Kneller et al., 2021; Kneller et al., 2022; Drago et al., 2022; Murakawa et al., 2023; Yano et al., 2024; Drago et al., 2024). Importantly, the cold neutrons with wavelengths in the range of 2-5 Å used at neutron macromolecular diffraction beamlines do not induce radiation damage in protein crystals, allowing data to be collected at near-physiological (room) temperature.

We present here a combined X-ray and neutron crystallography study of BhRNase H1 that sheds new light on the enzyme catalytic mechanism. We determined room-temperature (RT) neutron structures of wild-type (WT) and inactive D132N mutant BhRNase H1 in the apo-form. In addition, we obtained three X-ray structures of D132N BhRNase H1 complexed with two different RNA/DNA hybrids. Based on our neutron and X-ray crystallographic results we suggest a proton transfer pathway for protonating the leaving O3′ group of RNA that is similar, but not identical, to that proposed previously (Rosta et al., 2011) and confirm the essential role of metal binding site M_A_ in RNase H catalysis.

Results

2

The RT neutron structures of apo-WT and apo-D132N BhRNase H1 were determined at 2.48 Å and 2.40 Å resolutions, respectively (Fig. 2A and B). We mapped the protonation states and compared their differences in the active site between the wild-type and mutant enzymes before it is bound to an RNA/DNA hybrid substrate. Importantly, we found E109 carboxylic group protonated in D132N neutron structure, but not in the wild-type enzyme, with all the other aspartate and glutamate active site residues in both structures observed as deprotonated and negatively charged. In addition, we obtained three X-ray structures of D132N BhRNase H1 complexed with two different RNA/DNA hybrids – 1) 5′-UCGACA-3′/5′-ATGTCGp-3′ (Hybrid1) designed with two 1-nucleotide sticky ends forming a duplex with a single nucleotide overhang on each end, where ‘p’ denotes phosphorylation of the DNA's 3′-end, and 2) 5′-CGACAU-3′/5′-ATGTCG-3′ (Hybrid2) designed to produce a complementary duplex. At Mg^2+^ concentration of 5 mM, we observed a single metal ion at site M_B_ in the RT X-ray structure of D132N/Mg/Hybrid1 complex at 1.80 Å resolution (Fig. 2C), but both metal-binding sites were occupied in the RT structure of D132N/2Mg/Hybrid2 complex at 1.90 Å resolution (Fig. 2D). D132N/Mg/Hybrid1 structure represents a product complex, where the 3′ nick is located in the active site and the O3′ hydroxyl coordinates M_B_. Conversely, D132N/2Mg/Hybrid2 complex mimics the Michaelis complex, in which the scissile phosphate connecting nucleotides A5 and U6 occupies the active site and bridges M_A_ and M_B_. Remarkably, when the crystals of D132N/Mg/Hybrid1 were soaked in 15 mM Mn^2+^, the metal binding at site M_A_ resulted in the hybrid duplex sliding by one nucleotide to position the scissile phosphate connecting nucleotides C5 and A6 in the active site, transforming the product mimic into a Michaelis-like D132N/MnMg/Hybrid1 mixed metal complex (structure obtained at 100K at 2.00 Å resolution, Fig. 2E).Fig. 2Three-dimensional structures of BhRNase H1 in the apo-form and in complexes with Hybrid1 and Hybrid2 RNA/DNA duplexes. (A, B) Neutron structures of the BhRNase H1 wild-type (A) and D132N mutant (B) enzymes in the apo-forms. (C-E) X-ray structures of D132N/Mg/Hybrid1 (C), D132N/2Mg/Hybrid2 (D) and D132N/MnMg/Hybrid1 (E) complexes.Fig. 2

Protonation states in wild-type and D132N BhRNase H1

2.1

The C-terminal domain of BhRNase H1 used in this study consists of residues 59-196 (Fig. S1) from the full-length enzyme. BhRNase H1 has the αβα Rossmann-like fold, with a β-sheet containing three antiparallel and two parallel strands and three major α-helices. A fourth short α-helix is located close to the C-terminus preceding residues E188 and D192 placed on an extended loop (Fig. 2). The chosen BhRNase H1 construct crystallizes readily in the apo-form for both the wild-type and D132N proteins (Fig. S2). To visualize the active site of BhRNase H1 at the atomic level, we obtained room-temperature neutron crystallographic structures of nucleic acid-free wild-type and D132N BhRNase H1 (referred to henceforth as apo-WT and apo-D132N) at the resolutions of 2.48 and 2.40 Å, respectively. The neutron diffraction data were jointly refined with 1.90 and 2.20 Å room-temperature X-ray data for apo-WT and apo-D132N, respectively (Table S1).

Fig. 3 depicts the 2F_O_-F_C_ electron density and 2F_O_-F_C_ neutron scattering length density maps for the active sites in apo-WT and apo-D132N structures for comparison of the structural information available from X-ray and neutron data. Neutron data provide clear visualization of protonation states, water molecules orientations and H bonding networks that X-ray data lack. In apo-WT, the carboxylic side chains of the catalytic residues D71, E109, D132, and D192, and the nearby E188 are all oriented towards each other, interacting through a network of water-mediated H bonds (Fig. 3A and B). Four water molecules, W1-W4, hydrate the carboxylate groups. None of the carboxylate side chains are protonated according to the neutron scattering density map. Although W1 is equidistant from both D71 and D132 (O···O distances are of 2.8 Å), it makes weak H bonds with both residues, with the D···O distances of 2.7 Å to D132 and 2.4 Å to D71. W2 appears to interact with E109, D132, main-chain carbonyl of V70 and main chain amide ND of D132 with the four distances indicative of H bonds. However, this water is oriented such that it makes shorter H bonds with the main-chain groups of V70 and D132 (D···O distances of 1.7 and 2.1 Å, respectively), longer H bonds with both E109 carboxylate oxygens (D···O distances of 2.6 Å), and no H bond with D132 side chain. W3 and W4 water molecules each bridges two carboxylate groups and are properly oriented to donate their Ds to form H bonds with D71 and D192 for W3 and D132 and E188 for W4. W3 is also anchored through accepting a D from the main-chain amide ND of Y193 and W4 accepts a D in an H bond with the side chain of T183.Fig. 3Electron density and neutron scattering length density maps for BhRNase H1 active site residues. The 2F_O_-F_C_ electron density maps (dark green mesh) are contoured at 1.5σ level and the 2F_O_-F_C_ neutron scattering length density maps (sand mesh) are contoured at 1σ level. (A and B) The active site of apo-WT. (C and D) The active site of apo-D132N. The omit F_O_-F_C_ neutron scattering length density map (red mesh) in panel D showing a D atom on the carboxylic group of E109 is contoured at 3.5σ level. D atoms are shown in lime green, and H atoms are omitted for clarity. All distances are given in Angstrom.Fig. 3

The D132N substitution does not dramatically alter the active site geometry (Fig. S3). The apo-WT and apo-D132N neutron structures align with the RMSD on the main-chain atoms of 0.3 Å. The only notable difference compared to the apo-WT structure observed in the electron density is the lack of W1 between D71 and N132 (Fig. 3C). But, a detailed analysis of the neutron scattering density map reveals reorientation of W3 and W4 that form different H bonds compared to those observed in apo-WT, and E109 carboxylic group becomes protonated, thus has a neutral charge (Fig. 3D). Based on the distances between the heavy atoms W2, W3 and W4 can each form several H bonds, yet according to the neutron scattering length density map each water molecule makes three H bonds. W2 donates its D atoms in H bonds with the main-chain carbonyl of V70 and side-chain carboxylate of E109 and accepts a main-chain amide ND in an H bond with N132. Thus, this water molecule is oriented identically in apo-D132N as in apo-WT. Conversely, W3 and W4 are reoriented in the mutant structure. Unexpectedly, W3 trades an H bond with D192 observed in the apo-WT for an H bond with the main-chain carbonyl of V72, but it maintains the H bonds with D71 and the main-chain amide ND of Y193, even though positions of the residues surrounding W3 do not change due to mutation. Reorientation of W4 is not surprising. W4 is forced to rotate to accommodate the conformation adopted by the N132 side chain, accepting its D atom in an H bond. W4 then donates its D atoms in H bonds with E188 and T183, and the W4 reorientation cascades to an obligatory 180° flip of the T183 side-chain hydroxyl relative to its orientation in apo-WT. Remarkably, we observed the protonation of E109 carboxyl side chain on the oxygen atom not involved in the H bond with W2. There is a clear difference density peak for a D atom in the omit F_O_-F_C_ neutron scattering length density map (Fig. 3D). The Oε2-D bond is rotated away from W2 and faces the S131 side chain. The O···O distance between the E109 Oε2 and S131 Oγ atoms is 3.7 Å, permitting the collocation of the D atoms.

Binding of different oligonucleotides facilitates trapping reaction intermediate mimics

2.2

To obtain crystals of D132N BhRNase H1 in complex with its substrate we examined two oligonucleotide RNA/DNA hybrids. Hybrid1 (6-mer: 5′-UCGACA-3′/5′-ATGTCGp-3′) produces a duplex with a single nucleotide overhang on each end including an additional phosphate on the 3′ end that does not interfere with the RNA/DNA duplex formation. Hybrid1 has identical sequence for both RNA and DNA compared to that used in our earlier study where both DNA's G nucleobases were chemically modified to replace the carbonyl's oxygen with selenium (Abdur et al., 2014). Hybrid2 (6-mer: 5′-CGACAU-3′/5′-ATGTCG-3′) forms a complementary duplex. Within both hybrids, the RNA and DNA oligonucleotides in these complexes adopt mixed A and B conformations, respectively. The RNA strand A conformation is characterized by the 3′-endo sugar puckers, and the DNA strand B conformation is recognized through the 2′-endo and/or 1′-exo puckered sugars. In the crystals, the RNA/DNA hybrids are arranged into infinite fibrils irrespective of the presence or absence of the nucleotide sticky ends, with one BhRNase H1 per the 6-mer nucleotide (Fig. S4), similar to previously reported structures (Nowotny et al., 2005; Nowotny and Yang, 2006).

D132N/Mg/Hybrid1 room-temperature X-ray structure was obtained at 1.80 Å resolution, unlike the previously reported 2.7 Å cryo-structure of a similar complex having the RNA/DNA duplex of the same sequence but lacking the 3′-phosphate in DNA (PDB ID 2G8U, Nowotny and Yang, 2006). Due to the poor electron density and unmodeled difference density peaks observed in the active site of 2G8U, we refrain from comparing it with the current structure. Unlike in 2G8U, the electron density in the D132N/Mg/Hybrid1 complex active site clearly shows a single Mg^2+^ ion positioned in the M_B_ site (Fig. 4A). The structure lacks electron density for metal in M_A_ site, with two water molecules (W2 and W3) located nearby. Hybrid1 duplex binds to the enzyme with the 3′-nick located in the active site, placing the 3′-OH of A6 nucleotide to interact with Mg_B_. As a result, D132N/Mg/Hybrid1 structure is resembling a product complex. The nonphosphorylated 5′-OH of U1∗ corresponding to the symmetry-related RNA oligonucleotide is positioned next to the 3′-end of A6 and engages the former OH in a short 2.5 Å H bond (Fig. 4B). Mg_B_ coordination sphere can be described as a distorted trigonal bipyramid, having short 2.2-2.3 Å equatorial Mg-O distances with the side chains of D71, N132 and E109 and much longer axial distances of 2.8 and 3.1 Å to W1_B_ and RNA's 3′-OH, respectively. Unexpectedly, when Hybrid2 containing complementary RNA and DNA sequences was used to obtain the D132N/2Mg/Hybrid2 complex, the oligonucleotide duplex bound to the enzyme placing a scissile phosphate connecting A5 and U6 at the active site where both catalytic metal sites M_A_ and M_B_ are occupied by Mg^2+^ ions (Fig. 4C). Consequently, D132N/2Mg/Hybrid2 can be defined as a Michaelis-like complex. Notably, both current complexes and 2G8U were made using 5 mM Mg^2+^ in the starting samples before mixing them with the crystallization well solutions. Therefore, the absence of Mg_A_ in D132N/Mg/Hybrid1 may be a consequence of the product-like nick present in the active site rather than low Mg^2+^ concentration. Indeed, Mg_B_ has virtually identical distorted trigonal pyramidal coordination sphere in D132N/2Mg/Hybrid2 (Fig. 4D) as it does in D132N/Mg/Hybrid1. Mg_A_ is bound having 6 ligands in its coordination sphere arranged in almost perfect octahedral geometry. It binds to carboxylates of D71 and D192, three water molecules and pro-S_P_ oxygen of the RNA main-chain phosphate. A similar Mg_A_ coordination has been observed previously and is considered essential for catalysis (Nowotny et al., 2005; Nowotny and Yang, 2006; Samara and Yang, 2018) as it places the catalytic water (W3_A_) less than 3.0 Å away from the scissile phosphorus atom. In a previous cryo-X-ray structure of D132N BhRNase H1 in complex with a 12-mer RNA/DNA hybrid (PDB ID 1ZBI; Nowotny et al., 2005), E188 was found directly coordinated to Mg_A_. Conversely, in D132N/2Mg/Hybrid2, E188 is not bound to Mg_A_ but rather is H-bonded to W2_A_ that mediates its interaction with the metal and N132 (Fig. 4D and S5). It appears that E188 coordination to Mg_A_ seen in 1ZBI pushes W2A away from the metal, but it is captured through H bonding by N132 and E188 (Fig. S5). Coordination of E188 to the metal in 1ZBI and detection of Mg_A_ in 2G8U complexes could be due to temperature effects as the earlier structures were obtained at low (cryo) temperature or due to crystal packing effects as the current and previous complexes were crystallized in different unit cells and E188 is located on the protein surface. Cryo temperature is known to distort the conformational ensembles in protein molecules and ligands (Keedy et al., 2014; Otten et al., 2018; Fischer, 2021). In addition, it has been demonstrated that crystal flash-freezing can result in a metal ion moving in to coordinate a main-chain phosphate in a DNA oligonucleotide replacing a proton on the oxygen atom (Vandavasi et al., 2018).Fig. 4Electron density maps, metal binding and active site interactions in BhRNase H1 complexes with RNA/DNA duplexes. (A) 2F_O_-F_C_ electron density map (orange mesh) in D132N/Mg/Hybrid1 is contoured at 2.0σ level (1.5σ level for E188 and D192) showing presence of a single Mg_B_ and depicting RNA binding. (B) Mg_B_ coordination and active site interactions in D132N/Mg/Hybrid1. (C) 2F_O_-F_C_ electron density map (orange mesh) in D132N/2Mg/Hybrid2 is contoured at 2.0σ level (1.0σ level for E188, W2_A_ and W4_A_) showing presence of two Mg^2+^ ions and depicting RNA binding. (D) Mg_A_ and Mg_B_ coordination and active site interactions in D132N/2Mg/Hybrid2. (E) 2F_O_-F_C_ electron density map (orange mesh) in D132N/MnMg/Hybrid1 is contoured at 1.5σ level showing presence of Mn_A_ and Mg_B_ metal ions and depicting RNA binding. (F) Mn_A_ and Mg_B_ coordination and active site interactions in D132N/MnMg/Hybrid1. Residues and RNA are colored by atom type. The symmetry related RNA oligonucleotides are colored by atom type with gray carbon atoms, labeled with asterisks. All distances are given in Angstrom.Fig. 4

Second metal binding changes oligonucleotide position in the active site

2.3

We next asked a question: Could a second metal ion bind at the vacant M_A_ site in D132N/Mg/Hybrid1 complex upon crystal soaking in a solution with metal concentration >5 mM? For the soaking experiments, we decided to use Mn^2+^ at a concentration of 15 mM for two reasons. Firstly, Mn^2+^ has almost twice as many electrons as does Mg^2+^, making its detection in the D132N BhRNase H1 active site straightforward because Mn^2+^ binding should result in a strong electron density peak at the metal site M_A_. Secondly, although Mn^2+^ is not a physiological metal, it supports catalysis and, similar to Mg^2+^, forms the octahedral coordination sphere (Samara and Yang, 2018; Hyjek et al., 2019; Pang et al., 2022). Following soaking a crystal of D132N/Mg/Hybrid1 complex with Mn^2^, we collected a low-temperature X-ray diffraction dataset to 2.0 Å resolution. Inspection of the electron density in the active site suggested the presence of metal ions at both metal sites. The peak heights in the difference F_O_-F_C_ electron density map measured 18 and 10 e/Å^3^ for sites M_A_ and M_B_, respectively (Fig. S6). We concluded that Mn^2+^ occupied site M_A_, whereas Mg^2+^ remained in site M_B_, and the resulting D132N/MnMg/Hybrid1 structure represents a mixed metal complex obtained without disturbing Mg^2+^ ion located at site M_B_.

The 2F_O_-F_C_ electron density map for the D132N/MnMg/Hybrid1 active site is shown in Fig. 4E. The electron density unequivocally demonstrates that the RNA nucleotide is positioned such that the phosphate connecting nucleotides C5 and A6 is placed between the two metal ions. Accordingly, this structure represents a Michaelis-like complex similar to that observed for D132N/2Mg/Hybrid2. The D132N/MnMg/Hybrid1 complex was generated from the starting D132N/Mg/Hybrid1 complex evidently through a one-nucleotide sliding of Hybrid1 on the protein surface after adding the metal ion to M_A_ binding site. Metal coordination spheres and H-bonding interactions in D132N/MnMg/Hybrid1 (Fig. 4F) remain identical to those observed in D132N/2Mg/Hybrid2. Importantly, addition of Mn^2+^ did not introduce crystal lattice rearrangements as the unit cell of D132N/MnMg/Hybrid1 remained identical that for D132N/Mg/Hybrid1 (Table S1).

Discussion

3

General acid-base catalysis governs many biochemical reactions through proton transfer events between enzyme residues, substrates, intermediates, products and water species. Proton transfers alter protonation states, and consequently electrical charges, shaping the electrostatics of the enzyme active site to promote a catalytic reaction or to bind a ligand (e.g., substrate, drug molecule, etc.). It is therefore of critical importance to know the locations of H atoms and to track their movements for deeper understanding of enzyme mechanisms and for improved structure-assisted drug design. In this context, BhRNase H1 catalyzes a seemingly uncomplicated hydrolysis reaction in which a water molecule attacks the phosphorus atom of the scissile phosphate within an RNA molecule to break the nucleic acid polymer apart (Fig. 1C). The BhRNase H1 catalysis must involve a series of proton transfers, including deprotonation and activation of the catalytic water molecule and protonation of the product's O3′ hydroxyl after the scissile P-O bond cleavage. However, the proton transfer pathways occurring during BhRNase H1 catalysis have not been experimentally elucidated. Several mechanisms for the BhRNase H1 catalytic reaction have been proposed based on the theoretical calculations and experimental evidence that, predictably, differ largely in the proton transfer pathways and also in the possible involvement of additional metal ions transiently moving into the active site during RNA hydrolysis to assist catalysis (De Vivo et al., 2008; Rosta et al., 2011; Palermo et al., 2015; Samara and Yang, 2018).

In this study we succeeded in obtaining room-temperature neutron structures of wild-type and inactive D132N mutant BhRNase H1 enzyme in the apo-form and X-ray structures of D132N mutant in complex with RNA/DNA hybrids. D132N complexes with RNA/DNA oligonucleotide duplexes contained either a single Mg^2+^ ion bound at site M_B_, two Mg^2+^ ions bound at sites M_A_ and M_B_, or Mn^2+^ and Mg^2+^ bound at both metal sites, respectively. All active site carboxylates belonging to the catalytic DEDD motif (D71, E109, D132, D192) and the nearby E188 were found to be not protonated, thus negatively charged, in the apo-WT neutron structure. The inactivating D132N mutation did not substantially perturb the active site geometry but resulted in the protonation of E109 and reorientation of some water molecules hydrating the carboxylate groups (Fig. 3B and D). Accordingly, E109 was observed to have a neutral carboxylic side chain in apo-D132N neutron structure. Binding of Hybrid1 duplex to D132N mutant using 5 mM Mg^2+^ led to the product-like D132N/Mg/Hybrid1 complex containing the metal ion only at site M_B_. Conversely, when complementary Hybrid2 duplex was used at the same Mg^2+^ concentration, a Michaelis-like D132N/2Mg/Hybrid2 complex formed in which both metal sites were occupied by Mg^2+^ ions. These observations indicate that metal binding at site M_A_ may be weaker than at site M_B_ in the product state and M_A_ may leave the active site first after RNA cleavage but before the enzyme dissociates from the RNA/DNA product. Indeed, adding metal ions exogenously to the crystals of D132N/Mg/Hybrid1 complex led to not only the extra metal binding at site M_A_ to create a mixed-metal complex but also triggered sliding of the Hybrid1 duplex by one nucleotide to present a scissile phosphate to the now properly constructed dimetal catalytic site. This observation confirms the two-metal catalytic mechanism of BhRNase H1, with the metal presence at site M_A_ being essential for both activation of the catalytic water molecule and for transient binding of additional monovalent and divalent metal ions during the hydrolysis reaction as previously proposed (Samara and Yang, 2018).

The earlier and current crystal structures, and our previous observation of nucleic acid main-chain phosphate protonation (Nowotny et al., 2005; Vandavasi et al., 2018) support the catalytic water deprotonation by the upstream RNA nucleotide phosphate that is located 4 Å away from M_A_ but is within an H bond distance from the catalytic water (Fig. 1B), as suggested by earlier theoretical calculations (Rosta et al., 2011). Direct protonation of pro-R_P_ oxygen on the scissile phosphate is unlikely because the latter bridges M_A_ and M_B_, whose combined +4 charge stabilizes the negative charge on the phosphate and creates a strong electrostatic field that would decrease the pro-R_P_ oxygen's proton affinity (i.e. pK_a_). Equally doubtful is the protonation of the product's leaving O3′ group by a proton transfer from the already protonated D132 suggested by earlier theoretical calculations (Rosta et al., 2011) because metal binding would lower the carboxylic side-chain pK_a_ further from its solution value of ∼4. The only active site residue that appears capable of changing its protonation state is E109. We therefore propose that O3′ is protonated by O2′ hydroxyl located on the same ribose ring in the 5′ direction by hopping through the E109 carboxylate. O2′ would then be re-protonated through a Grotthuss mechanism utilizing a chain of H-bonded water molecules, Ser133 and Thr135 connecting the O2′ hydroxyl group to the bulk solvent (Fig. 5). Importantly, it was shown that replacement of O2′ hydroxyl by fluorine inhibits RNase H activity suggesting its possible involvement in catalysis (Uchiyama et al., 1994a, Uchiyama et al., 1994b; Pallan et al., 2016). A similar proton transfer pathway, excluding E109 participation, was previously suggested but not investigated further (Rosta et al., 2011).Fig. 5. Proposed proton transfer pathway to protonate the O3’ leaving group by O2′-OH assisted by E109 and the subsequent re-protonation of O2’ through a Grotthuss mechanism involving a chain of H-bonded water molecules and enzyme residues. Proton hopping is depicted by blue curved arrows. All distances are given in Angstrom.Fig. 5

Conclusion

4

By using neutron protein crystallography at near-physiological (room) temperature we mapped the H atom positions in the wild-type and D132N apo-BhRNase H1. X-ray crystallography at both room- and cryo-temperatures revealed monometal, dimetal and mixed dimetal active sites of BhRNase H1 complexed with two different RNA/DNA hybrids, mimicking Michaelis and product complexes. Based on the observation that E109, a residue within the catalytic DEDD motif, can change its protonation state, we proposed a proton transfer pathway to deliver an H atom to the leaving O3′ group by the O2′ hydroxyl by hopping through the E109 carboxylate. The M_A_ binding site is confirmed to be essential for catalysis as the RNA/DNA duplex slides by one nucleotide transforming a product mimic structure, where the O3′ nick is bound in the active site, into a Michelis-like complex upon metal binding to site M_A_ where the scissile phosphate is presented to and bridges M_A_ and M_B_ metal ions. Neutron structures of BhRNase H1 in complex with RNA/DNA hybrids are being pursued to gain a deeper understanding of this enzyme function.

Methods

5

General information

5.1

The DNA and RNA oligonucleotides were purchased from W.M. Keck Foundation Biotechnology Resource Laboratory at Yale University (New Haven, CT, USA). Columns for protein purification were purchased from Cytiva (Piscataway, New Jersey, USA). His-tagged Tobacco Etch Virus (TEV) protease was produced in-house. Crystallization reagents and supplies were purchased from Hampton Research (Aliso Viejo, California, USA). Crystallographic supplies for crystal mounting and X-ray and neutron diffraction data collection at room temperature were purchased from MiTeGen (Ithaca, New York, USA) and Vitrocom (Mountain Lakes, New Jersey, USA).

Expression and purification

5.2

The Bacillus halodurans gene rnhA (EC:3.1.26.4) truncated to encode the C-terminal BhRNase H1 fragment (residues 59-196) (Fig. S2) was codon optimized, synthesized, and cloned into kanamycin-resistance plasmid, pD451-SR (ATUM, Newark, CA; sequence confirmed using Sanger sequencing by Atum), in addition to a DNA sequence encoding for an N-terminal polyhistidine-(His_6_)-tag with a 34 amino acid long linker. A TEV protease tag ENLYFQ was introduced at the BhRNase H1 N-terminus sequence so that after cleavage the enzyme sequence contained an extra N-terminal serine (Fig. S2). Because residues 59-61 are not visible in the BhRNase H1, the extra N-terminal residue does not interfere with the protein crystallization or nucleic acid binding. The plasmid was transformed into the competent E. coli BL21(DE3) expression vector. The cells were grown in Luria-Bertani (LB) media with 50 μg/mL kanamycin at 37 °C to an optical density of 0.8–1.0 and then induced with 1 mM isopropyl ß-D-1-thiogalactopyranoside (IPTG). The protein expression was performed overnight at 30 °C. The next day the cells were harvested by centrifugation at 5660 rpm at 4 °C and the cell pellets were resuspended in the lysis buffer made with 40 mM sodium dihydrogen phosphate pH 7.5, 75 mM NaCl, and 0.1 mM EDTA at a ratio of 5 mL of the buffer per gram of wet cell paste. The cells were disrupted by sonication and the lysate was clarified by centrifugation at 17,000 rpm for 40 min. The His-tagged BhRNase H1 was purified via affinity chromatography using a HisTrapFF (5 mL) nickel column equilibrated with 40 mM NaH_2_PO_4_ pH 7.0, 100 mM NaCl, and 10 mM imidazole. The protein was eluted using a linear gradient of 40 mM NaH_2_PO_4_ pH 7.0, 100 mM NaCl, and 500 mM imidazole. TEV protease (1 mg TEV protease/50 mg of tagged protein) was added to purified BhRNase H1 and the mixture was dialyzed against 40 mM NaH_2_PO_4_ pH 7.0, 150 mM NaCl, and 0.5 mM EDTA overnight at room temperature. The resultant solution containing TEV protease, untagged BhRNase H1 and the remaining unreacted tagged protein was loaded onto a HisTrapFF (5 mL) nickel column and untagged BhRNase H1 was collected in the flow-through. Pure BhRNase H1 was dialyzed against 40 mM NaH_2_PO_4_ pH 7.0, 75 mM NaCl and 5% glycerol, and concentrated to ∼20 mg/mL using Amicon 10,000 MWCO centrifugal filter (Millipore Sigma, Burlington, MA) and stored at −30 °C. The protein concentration was determined using absorption at 280 nm (A280) on a Nanodrop spectrophotometer with the calculated E_1%_ of 2.53.

Crystallization, soaking, and H/D-exchange

5.3

For protein crystallization, frozen wild-type and D132N mutant BhRNase H1 samples were thawed, dialyzed overnight against 20 mM HEPES pH 7.0, 100 mM NaCl, and filtered through a 0.2 μm centrifuge filter at 7,000g. Both wild-type and D132N mutant BhRNase H1 were then crystallized by sitting drop vapor diffusion methodology using 0.1 M NaOAc pH 5.0, 0.2 M (NH_4_)2_SO_4, and 20% PEG 3350 as the precipitant solution at 18 °C. To obtain large crystals of the enzyme apo-forms for neutron diffraction, the crystallization drops were increased to 40 μL using 1:1 ratio (v/v/) of protein-to-well solution and reducing the temperature to 14 °C over the period of 3 months. For neutron diffraction, the crystals of wild-type and D132N mutant BhRNase H1 were mounted in 2 mm-inner diameter quartz capillaries with the liquid plugs of 0.1 M NaOAc pH 5.0 and 22% PEG 3350 prepared in 99.8% D_2_O to perform H/D-vapor exchange. To make the enzyme/nucleic acid complex crystals, the oligonucleotides were annealed in a 1:1 M ratio by heating the corresponding mixtures to 90 °C for 1 min and then allowing them to cool slowly to room temperature prior to crystallization with D132N mutant BhRNase H1. The resulting duplexes Hybrid1 and Hybrid2 were mixed with the protein (5 mg/mL) in a 1:1 M ratio in the presence of 5 mM MgCl_2_. The complexes were then crystallized by sitting drop vapor diffusion methodology using 0.1 M HEPES pH 7.5, 0.2 M NaCl, and 12-16% PEG 10,000 as the precipitant solution at 20 °C.

X-ray diffraction data collection and structure refinement

5.4

Room temperature and 100K X-ray diffraction data collection on BhRNase H1 crystals were performed on a Rigaku HighFlux HomeLab instrument equipped with a MicroMax-007 HF X-ray generator, Osmic VariMax optics, a DECTRIS Eiger R 4 M detector, and Oxford Cryosystems cryostream at ORNL. The data were indexed and integrated using the CrysAlis Pro software package (Rigaku, The Woodlands, TX), and the data were reduced and scaled in the AIMLESS program in the CCP4 software suite (Winn et al., 2011; Evans and Murshudov, 2013). The X-ray structures were solved by molecular replacement in PHASER (McCoy et al., 2007) using phases from PDB codes 1ZBF for the apo-enzymes and 2G8U for the complexes with oligonucleotides. The structures were then refined with phenix.refine in the PHENIX suite (Adams et al., 2010; Liebschner et al., 2019). The room temperature X-ray structures of apo-wild-type and D132N mutant BhRNase H1 were subsequently used in joint X-ray/neutron refinements. The X-ray diffraction data collection statistics are presented in Tables S1 and S2.

Neutron diffraction data collection

5.5

Neutron diffraction was tested at room temperature on the IMAGINE (Coates et al., 2018; Schröder et al., 2018; Meilleur et al., 2013, 2018 2020) single-crystal diffractometer beamline located at the High Flux Isotope Reactor (Oak Ridge National Laboratory) using the broad bandpass functionality with neutron wavelengths between 2.8 and 10 Å. A complete room-temperature neutron diffraction dataset for the D132N mutant BhRNase H1 in the apo-form was collected on IMAGINE using a neutron wavelength range of 2.8-4.5 Å. Each neutron image was composed of a 30 h exposure of the crystal held in a stationary position. The crystal was rotated along the vertical axis (Δφ = 8°) before collecting each successive image. In total, 17 neutron diffraction images were collected. Neutron diffraction data processing was performed with a version of LAUEGEN (Campbell, 1995; Campbell et al., 1998) modified to account for the geometry of the cylindrical image plate detector. The wavelength-normalization curve was determined using the intensities of symmetry-equivalent reflections at different wavelengths in LSCALE (Arzt et al., 1999). No explicit absorption corrections were applied. The data were scaled and merged in SCALA (Weiss, 2001). Room-temperature neutron diffraction data for the wild-type BhRNase H1 in the apo-form were collected on the instrument MaNDi (Coates et al., 2015; Coates and Sullivan, 2020) at the Spallation Neutron Source (SNS) at ORNL. The crystal was held stationary for 24 h exposures and rotated 10° around the φ-axis before collecting the next image. All neutrons between 2 and 4.16 Å were used to collect the frames, with a total of 8 images collected. Neutron diffraction data collection on MaNDi were processed and integrated with 3D time-of-flight profile fitting in Mantid (Arnold et al., 2014; Sullivan et al., 2018). Wavelength normalization of the data was performed with LAUENORM (Helliwell et al., 1989; Campbell et al., 1998) and the data were scaled and merged in SCALA (Weiss, 2001). Neutron data collection statistics for both datasets are given in Table S1.

Joint X-ray/neutron (XN) refinement

5.6

Joint XN refinements of both wild-type and D132N mutant BhRNase H1 structures in the apo-form were carried out using nCNS (Mustyakimov and Langan, 2007; Adams et al., 2009), a patch of the Crystallography & NMR Systems (CNS) (Brunger et al., 1998) software suite. A single rigid body refinement was the first step in the refinement procedure. Subsequently, several rounds of atomic position, atomic displacement parameter, and D atom occupancy refinements were performed. In between the rounds of the joint XN refinement, the structures were visualized in the molecular graphics program COOT (Casanal et al., 2020) to confirm correct side chain modeling and direct the rotation of side chain hydroxyl, thiol, and ammonium groups as well as water molecules to construct accurate H bonding networks. Water molecules were modeled and refined as D_2_O due to H/D-vapor exchange. BhRNase H1 enzyme molecules were modeled with H atoms at non-exchangeable positions because hydrogenated protein was used in the experiment, while labile positions were initially modeled as D atoms. After D atom occupancy refinement, the exchangeable sites were modeled based on individual site occupancies, where an occupancy of −0.56 is indicative of a pure H atom and an occupancy of 1.00 reflects a pure D atom. Sites partially occupied by both H and D atoms were given two atom records with the partial occupancies adding up to 1.00. The percentage of D atom occupancy at a specific site is calculated according to the following formula: % D = (occupancy(D) + 0.56)/1.56. Joint X-ray/neutron refinement statistics can be found in Table S1.

Author contributions

Oksana Gerlits: Conceptualization, Methodology, Validation, Investigation, Writing – Original draft, Project administration, Funding acquisition. Aliyah Collins: Investigation, Methodology, Formal analysis, Writing – Review & Editing. Andrey Kovalevsky: Validation, Formal analysis, Resources, Data Curation, Writing – Original draft, Writing – Review & Editing, Visualization, Supervision.

Funding information

This research was supported in part by funding from the 10.13039/100021542Appalachian College Association and by the 10.13039/100006151Office of Basic Energy Sciences, U.S. Department of Energy.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.