Vibriophage phiVC8 DpoZ

Introduction

The vibriophage ΦVC8 DNA polymerase DpoZ is DNA polymerase belonging to the PolA (link) family and the phiVC8-like DpoZ subfamily, a group currently only identified in certain species of bacteriophages (Czernecki). DpoZ consists of two subfamilies: phiVC8-like and Wayne-like. These polymerases confer selectivity in addition of the nucleobase 2-aminoadenine (Z) over adenine (A), though dATP is still present in these phages despite complete ablation of A from their genomes. Z forms a non Watson-Crick base pair with thymine (T) consisting of three hydrogen bonds as opposed to the two present in A-T base pairing (C). Z is a relatively novel discovery, having only recently had its biosynthetic pathway described in detail (Zhou et al, 2021). DNA modifications in bacteriophages usually confer selective advantages by allowing phages to avoid host cell restriction enzyme digestion of their genomes (1). The phage S-2L, which encodes a PrimPol (link) polymerase, contains a Z-specific analog of the purine nucleotide enzyme PurA (link) known as PurZ. Polymerases specific to Z are required to incorporate the nucleotide completely over A into phage genomes, and as noted include DpoZ polymerases as well as the as-yet uncharacterized PrimPol identified in phage S-2L. The mechanisms by which these polymerases carry out these functions are still under investigation, though specific structural feature and putative specificity mechanisms are highlighted below.

Structure

The 2.8A crystal structure solved by Czernecki et al contains two main domains: a polymerase domain (scene) and a 3'-5' exonuclease domain (scene). The polymerase closely resembles E. coli polymerase I Klenow fragment (link), containing distinct palm (scene), thumb (scene), and fingers (scene) subdomains in addition to the exonuclease domain. The enzyme exhibits the typical fold of PolA polymerases including E. coli Klenow fragment and T7 DNA polymerase (link). The palm subdomain contains the polymerase active site (scene), where the thumb and fingers clamp onto a DNA substrate to hold it in place. The structure from Czernecki et al contains two conformations: thumb-exo open (scene) and thumb-exo closed (scene). These conformations involve movement of the thumb and exonuclease domains, and the closed conformation excludes binding of DNA when modeled with dsDNA (Czernecki). The residues K162 and G276 appear to have the largest positional shifts between the two conformations.

When aligned with other PolA polymerases in the Czernecki et al paper, three unique insertions in the polymerase domain are present. The first ranges from the residues 349-353 (scene) in the palm subdomain. The second is in helix O in the fingers subdomain from residues 442-448 (scene). The third is between helices O1 and P on the tip of the fingers, from residues 473-491 (scene). The longest of these insertions is very flexible, not being well defined on the electron density map (scene?) and indicating that DNA binding may be necessary for stabilization of that loop. The first and third insertions likely interact with dsDNA, though solving a ternary complex structure is required to confirm this. The second insertion (residues 442-448 (scene)) adds a loop in the helix structure that normally interacts with dNTPs (Miller, 2015).

The exonuclease domain is larger and contains noncanonical alpha helices compared to other members of the PolA family. These helices are referred to as E1 and E2 (scene). This domain also contains an insertion between helices alpha4 and alpha5 (scene) present in other phages such as T7 and implicated in shuttling the 3' end of the DNA primer between the polymerase and exonuclease sites (Juarex-Quinteros, 2021). Some critical residues for catalysis such as a universally conserved HD catalytic dyad (scene) are present in phiVC8 DpoZ, as these are required for polymerase activity. Some crucial residues are lacking in the structure, which may result in loss or alteration of some proofreading activity of the enzyme. However, the critical arginine residue responsible for stabilizing the gamma phosphate of incoming dNTPs is present at position 440 (scene). The enzyme also includes point mutations present in other DpoZ including L455, F459, and G548 (scene). These residues are present near or in the polymerase domain active site, and residues corresponding to these exist in PolA polymerases that do not incorporate Z. L455 and G548 (scene) do not appear in any known Wayne-like DpoZ subfamily structures, though the F459 (scene) residue is present. F459 is normally a tyrosine residue that acts as a steric gate for distinguishing dNTPs from NTPs, though this function may be conserved as was shown for T7 phage with a Tyr-->Phe mutation. In any case, these individual changes from other polymerases may help account for specificity in Z nucleobase recognition, as it is likely that it is not a single mutation in the enzyme that accounts for the selectivity.

Function

The primary unique feature of phiVC8 DpoZ and other DpoZ family members is its selectivity in Z vs A incorporation. The exact mechanism by which this occurs is still under investigation and cannot be definitively described, though there is a proposed mechanism based off of the currently solved structure. The noncanonical helices E1 and E2 (scene) comprise an unusually mobile portion of the exonuclease domain that may contribute to proofreading properties. Residues R161 and M165 of E2 (scene) have been shown to contact the base of leaving nucleotides when modeled. Mutating these residues in a double mutation to alanine resulted in a detrimental effect in exonuclease activity and about equivalent efficiencies of Z and A incorporation, though interestingly also resulted in a higher incorporation of Z over A that has not been investigated or accounted for. These residues may still have an important function in Z vs A incorporation that requires further investigation.

Czernecki et al propose that selectivity of Z vs A may result from the enzyme exhibiting selectivity in incorporating base pairs with three hydrogen bonds and excluding those with two (A-T). This could occur by polymerase backtracking, a feature previously thought to be distinct to RNA-dependent DNA polymerases but has also been shown to occur in some DNA-dependent DNA polymerases (source is somewhere in the Czernecki paper). Polymerase backtracking occurs when the enzyme disengages from the nascent 3' end of DNA and moves backwards on the template strand. This proposed mechanism may allow the enzyme to be more sensitive to base-pairing interactions and differentiate between two vs three hydrogen bonds in a base pair. This ability would be enhanced in the enzyme compared to other polymerases and the energy barrier between backtracking conformation and forward moving polymerization would need be to significantly lowered. This may work through an allosteric mechanism that gradually allows the transition of the enzyme from the polymerase domain to the exonuclease domain to excise any bases with fewer than three hydrogen bonds. The insertion at residues 117-127 (scene) may be involved in such a mechanism, though this is yet to be tested and would require analysis of a ternary complex. Further investigation into the mechanisms behind Z incorporation and selectivity is ongoing and may soon shed more light on the specific structural feature necessary to facilitate such specificity.