Mechanism of Dnase1 Family Catalysis
Structural biology has provided key insights into the catalytic mechanism of the Dnase1 family. Improved structures of Dnase1 have helped refine and account for the catalytic mechanism of DNA cleavage used by Dnase1 family members. The Dnase1 family primarily relies on a catalytic His conserved across the broader exonuclease-endonuclease-phosphatase (EEP) superfamily to cleave double-stranded DNA in a divalent cation (Mg
2+/Ca
2+) dependent manner. The dsDNA substrate of the Dnase1 family is principally B-DNA, longer than three nucleotides, with a slight preference for non-repeating base pairs
[27]. However, there is substrate variation within the Dnase1 family, best studied between Dnase1 and Dnase1L3. Dnase1 and Dnase1L3 may cooperate due to differences in substrate activity
[28]. Dnase1L3 targets a broader range of substrates, while Dnase1 primarily cleaves naked dsDNA non-specifically. Dnase1 cleaves complexed DNA poorly, though chromatin degradation by Dnase1 is aided by heparin-mediated DNA binding protein displacement
[29]. Dnase1L3 cleaves both naked DNA and complexed DNA, including chromatin, nucleosomes, DNA in microparticles/apoptotic blebs, DNA coated with lipids, and single-stranded DNA
[11][18][30]. However, Dnase1L3 is inactivated by heparin and can be cleaved by plasmin
[29]. After Dnase1 family cleavage, DNA has one backbone ester bond nicked, leaving a 3′ OH and 5′ phosphate
[1][4]. The mechanisms proposed for Dnase1-mediated DNA cleavage will provide a foundation for analysis of further enzyme specialization in the Dnase1 family.
Generally, Dnase cleavage must balance the thermodynamically favored hydrolysis of the phosphate backbone of DNA with the kinetically challenging task of nucleophilic attack on a negatively charged phosphate
[31]. Three essential features are required to strike this balance and complete phosphate backbone hydrolysis: nucleophilic hydroxide activation, transition state stabilization, and leaving group protonation
[32]. Several mechanisms have been proposed to address these three essential features in Dnase1-like family endonuclease reactions
[33][34][35][36]. The four mechanisms that have been proposed are the Glu-His-Water, Two Histidine Acid-Base, Double-Divalent Cation, and Single-Cation Carboxylic Acid models (
Figure 1).
Figure 1. Proposed catalytic mechanisms of Dnase1. The proposed mechanisms are shown in order of first publication date. (
A) The first published mechanism, proposed based on the first X-ray structure of
[33]. The Glu-His-Water mechanism is similar to serine proteases. (
B) The Two-Histidine Acid-Base mechanism is an updated mechanism from the group behind the Glu-His-Water model. It is based on more extensive X-ray structural information
[35]. (
C) The Double-Divalent Cation mechanism builds on the Two-Histidine Acid-Base model. The Double-Divalent Cation mechanism was based on mutagenesis data
[36]. (
D) The Single-Cation Carboxylic Acid mechanism reopened the question of Dnase1 family catalysis, based on a more recent substrate-cation crystal structure
[34].
The first model for Dnase1 family DNA hydrolysis is the Glu-His-water model. The development of the Glu-His-water model was facilitated by the first crystal structure of Dnase1, and is reminiscent of the serine protease catalytic triad
[27][33]. Dnase1 was crystallized in the presence of the calcium ion conjugated nucleotide Ca-pTp. The crystal structure relied on a 4 Å Fourier difference density map for the active site substrate, which did not permit accurate determination of substrate positioning (
Figure 1A). However, it could be concluded that a divalent calcium ion permitted phosphate interaction with the otherwise negatively charged active site. The Glu-His-Water model is based on the hydrogen bond distance between the nitrogen on the deprotonated H134 and an opposing carboxyl group in E78. The carboxyl group from E78 polarizes H134, allowing H134 to remove a proton from water, which can then attack the scissile phosphate (
Figure 1A). This model accounted for alkylation experiments showing that H134 is a catalytically essential residue
[37][38]. The sensitivity of the deprotonated H134 to a lower pH is consistent with the greater activity of Dnase1 at neutral pH. The key weakness in this model is the lack of a proton donor to donate a proton to the 3′ ester bond that completes the reaction.
The Glu-His-Water model was superseded by the Two Histidine Acid-Base catalysis model, which addressed the weakness of the previous model, and extended the catalytic interface of Dnase1. These refinements were made based on the first high-resolution structure of Dnase1 bound to DNA
[35]. In this model, Dnase1 interacts with the scissile phosphate via four residues: H134, D168, N170, and H252
[35] (
Figure 1B). The updated model changed the role of H134 and assigned new roles to the other amino acids interacting with the scissile phosphate. In contrast to a deprotonated histidine removing a proton from the attacking water, H134 acts as a general acid, and completes the reaction by protonating the 3′ ester. Instead, H252 is the deprotonated histidine that activates the attacking water (
Figure 1B). This is supported by mutagenesis data that demonstrates H252 is key to the catalytic mechanism
[39]. Both histidines are polarized by negatively charged side chains, E78 for H134 as before, and D212 for H252. A third acidic side chain, E39, coordinates the divalent cation from the previous mechanism. R9 (Q9 in human Dnase1) may stabilize the transition state. Along with the active site, this model assigned roles to key DNA binding residues that provide the energy relaxation needed for all these charged elements to occupy the confined active site via hydrogen bonding and Van der Waals interactions. The defined DNA-binding residues are E13, T14, R41, S43, N74, Y76, Y175, S206, T207, and Y211
[35]. Tyr76 is particularly important because it stacks against a nitrogenous base
[35]. The key strengths of this model are the ability to explain newer mutagenesis results, a complete catalytic mechanism, and the support of a substrate-enzyme crystal structure. The principal weaknesses of this model are the unassigned roles of N170 and D168, and the lack of structural evidence for H252 activating the attacking water.
The Double-Divalent Cation catalytic model built upon the previous model. The updated model was based on mutagenesis studies performed on previously assigned active site residues, and corroborated the DNA binding residues
[36][40][41]. The key finding was that E39 and D168 were both essential for activity (
Figure 1C). D168 was assigned the role of metal binding, which developed the two-cation model active site. This model is characterized by both D168 and E39 binding cations. Between the two cations, the histidines perform acid-base catalysis exactly as described in the previous model (
Figure 1C). Strengths of the double divalent cation model include assignment of roles to the most experimentally studied residues, support from molecular dynamics simulations, and explanation of both mutagenesis and structural data
[35][36][40][41][42]. This model is supported by a similar proposed mechanism for the related EEP family member (17% sequence similarity), Human Apurinic/Apyrimidinic Endonuclease1 (hAPE1) at neutral pH
[43]. This model also is supported by pKa studies of the active site histidines, via pH dependent imidazole rescue of Dnase1 activity for H134A and H252A mutants
[40]. However, the Double-Divalent Cation model suffers from two key weaknesses. The role for the catalytically required N170 is unexplained, and there is no structural support for the two divalent cations in the Dnase1 active site.
The most recent model of Dnase1 catalysis is the Single Cation Carboxylic Acid catalysis model. This model was proposed based on a high-resolution structure solved during development of therapeutic Dnase1 for Cystic Fibrosis
[34]. This model accounts for new structural evidence from the recombinant Dnase1 crystal structure (PDB:4AWN) showing a phosphate anion and Mg
2+ cation together in the active site (
Figure 1D)
[34]. The Single Cation Carboxylic Acid catalysis model includes only one metal ion in the active site and assigns the role of the catalytic base to D168. Protonation of the leaving group is performed by a proton coordinated by E39 (
Figure 1D). This model diverges from previous models to account for the orientation of the observed phosphate and Mg
2+ in the active site. The strength of the Single Cation Carboxylic Acid catalysis model is its basis on a recent high-resolution substrate-cation structure. This model is consistent with a proposed hAPE1 mechanism
[44][45]. Key weaknesses of this model are that it does not provide a biochemical mechanism for the protonation event and discounts previous experimental results highlighting the importance of the second histidine in the active site
[39]. Furthermore, the proposed hAPE1 catalytic mechanism is controversial because the crystal structures support both a single cation-binding site and two cations binding in the active site, depending on pH
[43]. Thus, there are currently two leading models for the mechanism by which Dnase1 cleaves DNA.
There remain many opportunities to better understand the mechanism of catalysis by Dnase1 family members. To date, the catalytic mechanism relies entirely on Dnase1 because Dnase1 is the only Dnase1 family member whose structure has been solved. Since the catalytic site is well conserved, it is likely that the mechanism will remain similar for all family members. Importantly, the key catalytic residues discussed here are identical across all family members
[7]. However, despite this active site identity, Dnase1L2 is more active at an acidic pH. Additionally, Dnase1L3 has a unique C terminal domain not present in any other Dnase1 family or EEP superfamily member
[4][18]. This domain enables digestion of dsDNA complexed with proteins and lipids
[18][30], which could add to the catalytic mechanism. Finally, structural characterization of the Dnase1L3 inhibitors DR396, Pontacyl Violet, and Fmoc-D-cyclohexylalanine are still outstanding
[46][47]. Thus, there are many opportunities to better understand this important enzyme family, especially with a broader focus beyond Dnase1.