Impact of Amino Acid Substitutions in Region II and Helix K of Herpes Simplex Virus 1 and Human Cytomegalovirus DNA Polymerases on Resistance to Foscarnet

ABSTRACT

Amino acid substitutions conferring resistance of herpes simplex virus 1 (HSV-1) and human cytomegalovirus (HCMV) to foscarnet (PFA) are located in the genes UL30 and UL54, respectively, encoding the DNA polymerase (pol). In this study, we analyzed the impact of substitutions located in helix K and region II that are involved in the conformational changes of the DNA pol. Theoretical substitutions were identified by sequences alignment of the helix K and region II of human herpesviruses (susceptible to PFA) and bacteriophages (resistant to PFA) and introduced in viral genomes by recombinant phenotyping. We characterized the susceptibility of HSV-1 and HCMV mutants to PFA. In UL30, the substitutions I619K (helix K), V715S, and A719T (both in region II) increased mean PFA 50% effective concentrations (EC₅₀s) by 2.5-, 5.6-, and 2.0-fold, respectively, compared to the wild type (WT). In UL54, the substitution Q579I (helix K) conferred hypersusceptibility to PFA (0.17-fold change), whereas the substitutions Q697P, V715S, and A719T (all in region II) increased mean PFA EC₅₀s by 3.8-, 2.8- and 2.5-fold, respectively, compared to the WT. These results were confirmed by enzymatic assays using recombinant DNA pol harboring these substitutions. Three-dimensional modeling suggests that substitutions conferring resistance/hypersusceptibility to PFA located in helix K and region II of UL30 and UL54 DNA pol favor an open/closed conformation of these enzymes, resulting in a lower/higher drug affinity for the proteins. Thus, this study shows that both regions of UL30 and UL54 DNA pol are involved in the conformational changes of these proteins and can influence the susceptibility of both viruses to PFA.

INTRODUCTION

Most antiviral agents currently approved for the prophylaxis and treatment of herpes simplex virus (HSV) and human cytomegalovirus (HCMV) infections target the viral DNA polymerase (pol), an enzyme essential for viral replication that is encoded by the UL30 and UL54 genes, respectively. First-line drugs for HSV infections include the nucleoside analogues acyclovir (ACV), its prodrug valacyclovir, and famciclovir (the prodrug of penciclovir) (1). First-line treatment of HCMV infections is based on the use of ganciclovir (GCV), a deoxyguanosine analogue, and its prodrug valganciclovir (2). These nucleoside analogues require a first phosphorylation by a viral kinase followed by two subsequent phosphorylations by cellular kinases. The triphosphorylated forms are then incorporated into replicating DNA and ultimately stop chain elongation. However, clinical failure that may be associated with drug resistance can arise in immunocompromised patients who received nucleoside analogues for long periods of time. Second-line treatments consist of the cytosine analogue cidofovir (CDV) and the pyrophosphate analogue foscarnet (PFA) (1). CDV needs to be phosphorylated only twice by cellular kinases to be active. CDV-diphosphate competes with dCTP for incorporation in elongating DNA and slows down the viral DNA pol activity. Two successive incorporations of CDV-diphosphate are needed to stop the polymerization reaction (3). PFA does not require phosphorylation; it binds directly to the viral DNA pol and prevents the release of pyrophosphate following transfer of the nucleotide to the elongating DNA. However, the emergence of HSV and HCMV isolates with amino acid substitutions in the viral DNA pol conferring resistance to one, two, or even all antiviral agents has been reported (4, 5). Recently, letermovir, an inhibitor of the HCMV terminase complex, has been approved for prophylaxis in HCMV-seropositive adult recipients of hematopoietic stem cell transplant (6).

The DNA pol of herpesviruses belong to the family B DNA pol, which also includes the gp43 DNA pol of RB69 and T4 bacteriophages (7). The gp43 DNA pol of RB69 has been crystallized in the different conformations adopted by the enzyme during the DNA polymerization process (8–10). The structure of the DNA pol has been likened to a right hand that contains different domains, i.e., the palm, thumb, fingers, exonuclease and NH₂-terminal domains (11, 12). In addition, the DNA pol of herpesviruses possesses a pre-NH₂ terminal domain (13). The DNA polymerization steps can be summarized as follows: while the DNA pol is in an open conformation, the thumb domain closes around the DNA duplex (14). The incorporation of the nucleotide n + 1 induces the fingers domain (i.e., helices N and P) to move away from helix K and get closer to the palm domain, and the enzyme adopts a closed conformation. The nucleotide is then transferred to the 3′-OH end of the elongating DNA, and the pyrophosphate is released. The fingers domain moves away from the palm domain and gets closer to the helix K, and the DNA pol returns to an open conformation.

It has been suggested that PFA binds to and traps the DNA pol in its closed conformation (15). Some amino acid substitutions located in helices N and P (fingers domain) were shown to displace the equilibrium between the open and closed conformations to a more open form, thereby reducing the affinity of PFA for the DNA pol and leading to drug resistance (15–19). We hypothesize that changes in residues located in helix K (NH₂-terminal domain) and region II (palm domain) that are involved in the open-to-closed conformational changes of the DNA pol may also affect the susceptibility of HSV-1 and HCMV to PFA. To investigate the impact of amino acid substitutions localized in these two regions on the susceptibility of herpesviruses to PFA, we used a strategy based on sequence alignments of the DNA pol of HSV-1 and HCMV, which are susceptible to PFA, and those of RB69 and T4 bacteriophages, which are naturally resistant to this antiviral. The proof of concept of this strategy was already demonstrated, as the change of the V478 residue in helix N of the fingers domain of RB69 enzyme by the tryptophan located at an equivalent position in UL30 (W781) and UL54 (W780) renders the DNA pol of RB69 susceptible to PFA (18). Conversely, the opposite substitutions W781V (UL30) and W780V (UL54) lead to resistance of HSV-1 and HCMV to PFA (16). We thus identified a series of residues in helix K and region II that are different between the DNA pol of herpesviruses and those of bacteriophages (Fig. 1A and B). We characterized drug susceptibilities of recombinant HSV-1 and HCMV harboring these theoretical substitutions and that could be recovered in cell culture. We then evaluated the replicative capacity of recombinant viruses carrying substitutions associated with PFA resistance. We also determined the kinetics parameters of UL30 and UL54 DNA pol activity and their inhibition by PFA in enzymatic assays. Finally, we analyzed the impacts of these substitutions on the structure of DNA pol by using three-dimensional molecular modeling.

FIG 1.

Alignment of amino acid sequences of helix K (A) and region II (B) of DNA pol of RB69 and T4 bacteriophages (gp43) with DNA pol of HSV-1 (UL30) and HCMV (UL54). Amino acids that are different between bacteriophages and herpesviruses DNA pol (in bold) were selected, and theoretical substitutions were introduced into recombinant HSV-1 and HCMV. Both tables summarize drug susceptibility profiles of recombinant HSV-1 and HCMV mutants recovered in cell culture as well as those that were unable to grow. Amino acid substitutions in red are those which confer resistance or hypersusceptibility to foscarnet (PFA) and which were further investigated. 1, Three different cosmid and plasmid transfections were attempted without success for each of the two independent clones. 2, Three different bacmid transfections were attempted without success for each of the two independent clones. 3, Data are from reference 20. 4, Data are from reference 21. R, resistant; S, susceptible; HS, hypersusceptible; ACV, acyclovir; GCV, ganciclovir.

RESULTS

Generation of recombinant viruses.

Most recombinant viruses harboring theoretical substitutions located in helix K (Q618I, I619K, T623A, and L625I for HSV-1 and Q579I, I580K, T584A, and L586I for HCMV) and in region II (V694Q, Q697P, P712Y, V715S, and A719T for HSV-1 and V694Q, Q697P, V715S, and A719T for HCMV) that were identified based on the sequence alignment of DNA pol of herpesviruses and bacteriophages could be generated as summarized in Fig. 1A and B. The substitutions Q579I (in helix K of UL54) and V715S (in region II of UL30 and UL54) substitutions have already been characterized by our group (20, 21). Some recombinant viruses were too much impacted by the substitution to be able to grow in cell culture. Indeed, recombinant viruses harboring substitutions located in helix K of the enzyme, such as Q617P, R620T, and L626F for HSV-1 and their counterparts Q578P, R581T, and L587F for HCMV, were unable to grow after three independent transfection assays with two different clones of each mutant virus (Fig. 1A and B). Similarly, recombinant HSV-1 and HCMV harboring the substitution F718L in region II of the DNA pol demonstrated a defective growth phenotype in cell culture. In contrast to its HSV-1 counterpart, recombinant HCMV carrying the P712Y substitution in region II was unable to grow.

Modeling of DNA polymerase mutants resulting in a viral growth defect.

Based on our 3D models of the closed forms of UL54 and UL30 DNA pol (derived from the DNA pol δ of Saccharomyces cerevisiae bound to a primer template DNA and an incoming nucleotide; PDB 3IAY), residue Q578, located in the helix K of UL54 DNA pol (or Q617 in UL30), establishes van der Waals interactions with the DNA template (Fig. 2A). The changes brought about by the substitution Q578P could affect the main-chain conformation, thereby preventing its binding to the DNA template. Moreover, the role of the Q578 side chain in stabilizing the local structure, e.g., via its H-bonding with R581 (R620 in UL30) as well as van der Waals interactions with V812 and A816 (V812 and S816 in UL30) in helix P of the fingers domain, could be crucial to maintain the overall conformation of the DNA pol required for their activity.

FIG 2.

Local environments of amino acid substitutions associated with a viral growth defect and located in helix K of UL30 and UL54 DNA polymerases. The cartoon representations of the UL54 closed form and UL30 open form (crystal structure PDB 2GV9) are in light blue and wheat, respectively. (A) Structural model of Q578 and surrounding residues in the closed form of UL54 DNA pol. Q578 interacts with R581 (H bond) and V812 and A816 (van der Waals interactions). The substitution Q578P would lead to the loss of these interactions and might be detrimental to DNA binding. (B) The residue R620 in the open form of UL30 DNA pol has an important role in maintaining the local structure by two H bonds (with G616 and S816) and numerous van der Waals interactions. (C) The residue R581 in the closed form of UL54 DNA pol (R620 in UL30) is involved in the stabilization of the structure to accommodate the binding of the DNA molecule. (D) Crystal structure of UL30 DNA pol showing both hydrophobic and hydrophilic residues in the immediate environment of L626. Introducing a bulkier hydrophobic phenylalanine at residue 626 will disrupt the local structure that is maintained by interactions between hydrophilic residues E294 and R302. The H bonds are shown as dashed lines. The carbon atoms of mutated residues (labeled in blue and in brackets) are in cyan. The DNA molecule is derived from the structure of the closed form of yeast DNA pol δ (PDB 3IAY). The figure was prepared by using PyMOL (43).

Furthermore, based on the crystal structure of the open form of UL30 (PDB 2GV9), the bulky and positively charged R620 is very important to strengthen the local structure, as evidenced by its van der Waals or H-bonding interactions with V601, V812, V813, S816, G616, and T612 (Fig. 2B). A similar role for the corresponding residue R581 was also demonstrated in the closed form of UL54 (Fig. 2C). All these interactions may be pivotal for the integrity of the local structure, which establishes important direct contacts with both the fingers domain (helix P) and the DNA molecule. Replacing the arginine with a much smaller threonine residue would inevitably cause the disruption of the local structure, leading to an inactive DNA pol.

The introduction of a phenylalanine instead of the leucine at codon 626/587 in helix K (UL30/UL54) may also disrupt the local environment. In fact, the crystal structure of the open form of UL30 shows that introducing a bulkier side chain at this position seems unlikely to be compatible with the surrounding residues, including E294, V297, R302, and L305, which are located in the NH₂-terminal domain (Fig. 2D). Therefore, this residue would be detrimental to the appropriate conformational changes required for the function of the DNA pol.

Residue P712 is located in region II of both UL30 and UL54 DNA pol. It is not directly involved in the interactions with the DNA molecule. The recombinant HSV-1 harboring the P712Y substitution is still viable, likely because this amino acid change might be easily accommodated, as it faces a more flexible segment (A899-A900-G901) in the closed form of UL30 (Fig. 3A) than its counterpart (P923-Q924-A925) in UL54 (Fig. 3B) DNA pol. The P712Y substitution in the latter could not be tolerated, as this change would cause steric hindrance between Y712 and its surrounding residues, such as P923 and V955, resulting in the collapse of the local structure, which will render the DNA pol either unstable or unable to bind the DNA molecule.

FIG 3.

Local environments of amino acid substitutions associated with a viral growth defect and located in region II of UL30 and UL54 DNA polymerases. The cartoon representations of the UL30 closed form, UL54 closed form, and UL30 open form (crystal structure; PDB 2GV9) are in pale green, light blue, and wheat, respectively. (A) Local structure of P712 in the closed form of UL30 DNA pol. The substitution P712Y may be accommodated here, as P712 faces a segment containing short side chains (A899-A900-G901), which might be more flexible. (B) P712 in the closed form of UL54 DNA pol faces the segment P923-Q924-A925, which may not be flexible enough to be compatible with the bulky tyrosine introduced by the P712Y substitution. (C) Crystal structure of UL30 DNA pol showing a compact hydrophobic pocket for F718. The substitution F718L will be detrimental to the local structure, as the nearby secondary structures could not be maintained in the presence of a smaller leucine at this position. The carbon atoms of mutated residues (labeled in blue and in brackets) are in cyan. The figure was prepared by using PyMOL (43).

Residue F718, in both UL30 and UL54 DNA pol, establishes numerous van der Waals interactions with surrounding residues, including L721, M844, L845, F918, I922, and L924, as shown in the crystal structure of the open form of UL30 (Fig. 3C). The shorter side chain of leucine compared to phenylalanine would most likely shift the main-chain atoms in order to maintain the favorable interactions with the above-mentioned hydrophobic residues, which may cause its main carbonyl chain group and the carboxylate group of neighboring D717 to move away from their optimal positions for anchoring Mg²⁺ ions, which are critical for DNA pol activity (8, 13).

Thus, these results showed that these amino acids are important for the local structure of UL30 and UL54 DNA pol and that specific substitutions can lead to defective viruses.

Antiviral drug susceptibilities.

The different recombinant HSV-1 and HCMV mutants that were able to grow in cell culture were further evaluated for drug susceptibility. The EC₅₀s (effective concentrations that reduce plaque numbers by 50% compared to cells without drugs) of PFA and ACV against the wild-type (WT) recombinant HSV-1 were 52 ± 20 μM and 0.039 ± 0.018 μM, respectively (Table 1). Recombinant HSV-1 harboring substitutions Q618I, T623A, and L625I (in helix K) as well as V694Q, Q697P, and P712Y (in region II) were all susceptible to PFA (0.18-, 0.96-, 1.6-, 1.7-, 1.3-, and 0.62-fold changes in mean EC₅₀s compared to WT, respectively). The substitution I619K (in helix K) was associated with both PFA and ACV resistance (2.5- and 2.1-fold increases in mean EC₅₀s compared to those of WT, respectively). Substitutions V715S and A719T (in region II) were associated with resistance to both PFA (5.6- and 2.0-fold increases in mean EC₅₀s compared to WT, respectively) and ACV (9.2- and 3.5-fold increases compared to WT, respectively) (21). All drug susceptibility testing was confirmed with a second clone (which corresponds to a second recombinant mutant virus generated independently from the first clone). Thus, the substitution I619K in helix K and the substitutions V715S and A719T in region II were found to confer PFA resistance to HSV-1.

TABLE 1.

Antiviral drug susceptibilities of wild-type HSV-1 and recombinant mutants^a

Location of mutation	Recombinant virus	EC50b (fold changec) [n]
		Foscarnet	Acyclovir
None	rvHSV-1 WT	52 ± 20 (1) [23]	0.039 ± 0.018 (1) [18]
Helix K	rvHSV-1 Q618I-1	9.3 ± 1.1 (0.18) [3] ns	0.055 ± 0.007 (1.5) [3] ns
	rvHSV-1 Q618I-2	9.20 (0.16) [1]	0.063 (1.6) [1]
	rvHSV-1 I619K-1	129 ± 69 (2.5) [3]***	0.08 ± 0.02 (2.1) [3]**
	rvHSV-1 I619K-2	123 (2.3) [1]	0.083 (2.1) [1]
	rvHSV-1 T623A-1	49 ± 6.4 (0.96) [3] ns	0.051 ± 0.008 (1.4) [3] ns
	rvHSV-1 T623A-2	74 (1.4) [1]	0.038 (1.01) [1]
	rvHSV-1 L625I-1	82 ± 21 (1.6) [3] ns	0.09 ± 0.015 (2.4) [3]***
	rvHSV-1 L625I-2	46 (0.87) [1]	0.105 (2.8) [1]
Region II	rvHSV-1 V694Q-1	91 ± 26 (1.7) [3] ns	0.048 ± 0.011 (1.3) [3] ns
	rvHSV-1 V694Q-2	55 (1.05) [1]	0.042 (1.1) [1]
	rvHSV-1 Q697P-1	66 ± 15 (1.3) [3] ns	0.056 ± 0.010 (1.5) [3] ns
	rvHSV-1 Q697P-2	79 (1.5) [1]	0.071 (1.8) [1]
	rvHSV-1 P712Y-1	33 ± 5.5 (0.62) [3] ns	0.14 ± 0.021 (3.7) [3]***
	rvHSV-1 P712Y-2	28.1 (0.5) [1]	0.26 (6.8) [1]
	rvHSV-1 V715Sd	289 ± 24 (5.6) [3]***	0.36 ± 0.12 (9.2) [3]***
	rvHSV-1 A719T-1	104 ± 32 (2.0) [3]**	0.13 ± 0.005 (3.5) [3]***
	rvHSV-1 A719T-2	170 (3.2) [1]	0.19 (5.1) [1]

^aEC₅₀, 50% effective concentration, in micromolar units; WT, wild type; rv, recombinant virus.

^bOne-way ANOVA with Dunnett’s posttest (GraphPad Prism, version 8.00) was used to compare EC₅₀s of mutant viruses to those of the WT. **, P < 0.01; ***, P< 0.001; ns, nonsignificant compared to the WT. EC₅₀s that are significantly different from those of the WT were considered to induce drug resistance (bold) (this correlates with an increase of the EC₅₀s of ≥2-fold compared to those of the WT counterparts). Results are means ± SD from 3 to 23 independent experiments (n) for clone 1; clone 2 was tested only once.

^cFold change compared to the WT rvHSV-1.

^dEC₅₀s from reference 21.

The EC₅₀s (effective concentrations that reduce the Gaussia luciferase [GLuc] activity by 50% compared to cells without drug) of PFA and GCV against the wild-type (WT) recombinant HCMV were 28 ± 6.7 μM and 2.2 ± 1.2 μM, respectively (Table 2). Recombinant HCMV harboring substitutions I580K, T584A, L586I (in helix K), and V694Q (in region II) were susceptible to PFA (1.1-, 0.7-, 1.03-, and 1.6- fold changes in mean EC₅₀s compared to WT, respectively). The substitution Q579I located in helix K was associated with a phenotype of hypersusceptibility to PFA (mean EC₅₀s of 0.17-fold compared to WT) (20). The substitutions Q697P and A719T, located in region II, conferred resistance to PFA with increases in mean EC₅₀s of 3.8- and 2.5-fold compared to those of WT, respectively. However, recombinant HCMV carrying the substitutions Q697P and A719T were either susceptible or resistant to GCV (2.4-fold increase in the mean EC₅₀ compared to the WT). We also previously showed that HCMV with the substitution V715S is resistant to PFA and GCV (2.8- and 2.7-fold increases in EC₅₀s compared to those of the WT, respectively) (21). All drug susceptibility testing were confirmed with a second clone, except for the substitution I580K, for which the second clone was susceptible to GCV (0.95-fold increase in EC₅₀ compared to the WT, whereas the first clone was at the limit of resistance). Thus, the substitution Q579I in helix K of UL54 DNA pol was associated with hypersusceptibility to PFA, whereas the substitutions Q697P, V715S, and A719T in region II conferred resistance to the drug.

TABLE 2.

Antiviral drug susceptibilities of the wild type and recombinant HCMV mutants^a

Location of mutation	Recombinant virus	EC50b (fold changec) [n]
		Foscarnet	Ganciclovir
None	rvHCMV WT	28 ± 6.7 (1) [18]	2.2 ± 1.2 (1) [19]
Helix K	rvHCMV Q579Id	4.8 ± 0.57 (0.17) [3]***	0.93 ± 0.21 (0.42) [3] ns
	rvHCMV I580K-1	32 ± 10.2 (1.1) [3] ns	4.6 ± 2.5 (2.1) [3]*
	rvHCMV I580K-2	27 (0.96) [1]	2.1 ± 0.012 (0.95) [2]
	rvHCMV T584A-1	19 ± 1.7 (0.7) [3] ns	2.2 ± 0.47 (1.06) [3] ns
	rvHCMV T584A-2	20 (0.71) [1]	3.1 (1.5) [1]
	rvHCMV L586I-1	29 ± 9.4 (1.03) [3] ns	1.2 ± 0.41 (0.56) [3] ns
	rvHCMV L586I-2	38 (1.3) [1]	1.9 (0.89) [1]
Region II	rvHCMV V694Q-1	42 ± 9.3 (1.6) [3] ns	1.05 ± 0.1 (0.49) [3] ns
	rvHCMV V694Q-2	48 (1.7) [1]	3.7 (1.7) [1]
	rvHCMV Q697P-1	104 ± 38 (3.8) [3]***	1.6 ± 0.41 (0.71) [3] ns
	rvHCMV Q697P-2	70 (2.5) [1]	3.8 (1.8) [1]
	rvHCMV V715Se	78 ± 14 (2.8) [4]***	5.9 ± 1.6 (2.7) [4]**
	rvHCMV A719T-1	69 ± 12 (2.5) [3]***	5.2 ± 2.4 (2.4) [4]**
	rvHCMV A719T-2	66 (2.4) [1]	5.01 (2.3) [1]

^aEC₅₀, 50% effective concentration, in micromolar units; WT, wild type; rv, recombinant virus.

^bA one-way ANOVA with Dunnett’s posttest (GraphPad Prism, version 8.00) was used to compare EC₅₀s of mutant viruses to those of the WT. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ns, nonsignificant compared to the WT. EC₅₀s that are significantly different from those of the WT were considered to induce drug resistance or hypersusceptibility (bold) (this correlates with an increase [resistance] or a decrease [hypersusceptibility] of the EC₅₀s of ≥2-fold compared to those of the WT counterparts). Results are means ± SD from 3 to 19 independent experiments (n) for clone 1; clone 2 was tested only once or twice (I580K).

^cFold change compared to the WT rvHCMV.

^dEC₅₀s from reference 20.

^eEC₅₀s from reference 21.

Viral replicative capacity.

We then evaluated the replicative capacities of the WT and recombinant HSV-1 and HCMV mutants that were resistant to PFA. The replicative capacities of the WT and recombinant HSV-1 harboring I619K and A719T were determined in Vero cells using real-time cell analysis (RTCA). The cell indexes were recorded for 3 days postinfection. Results showed a significant decrease in the viral replicative capacity of recombinant I619K and A719T mutants on day 3 postinfection (P < 0.01 and P < 0.05, respectively) (Fig. 4A). We previously reported that the replicative capacity of recombinant HSV-1 harboring the substitution V715S was decreased by 1.5-fold compared to the WT on day 3 postinfection (21).

FIG 4.

Replicative capacity of the wild type (WT) and different recombinant HSV-1 and HCMV mutants. (A) The replicative capacity of the WT and the different recombinant HSV-1 mutants harboring substitutions conferring resistance to foscarnet (I619K and A719T in UL30) were determined by real-time cell analyses. Vero cells were infected with the different recombinant viruses at a multiplicity of infection (MOI) of 0.01. Cell indexes were recorded every 30 min for 3 days. We used the value of 1/cell index on selected days postinfection to analyze the impact of each substitution on the replicative capacity of recombinant viruses compared to the WT. Data are means and standard deviations (SD) for six replicates and are representative of three independent experiments. (B) The replicative capacity of the WT and the different recombinant HCMV harboring substitutions conferring resistance to foscarnet (Q697P and A719T in UL54) were evaluated by the Gaussia luciferase (GLuc) reporter-based assay. MRC-5 cells were infected with the different recombinant viruses at an MOI of 0.001. Supernatants were collected daily for 8 days and the GLuc activity was measured. Results are expressed as relative light units (RLU). Data are means and SD for six replicates and are representative of three independent experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

The replicative capacities of the WT and recombinant HCMV carrying the substitutions Q697P and A719T were evaluated by using the GLuc reporter-based assay. The GLuc activity was measured in cell culture supernatants collected daily for 8 days. The viral growth of the Q697P mutant was significantly decreased compared to that of the WT on days 5 (P < 0.01), 6, 7 (both P < 0.001), and 8 (P < 0.01) postinfection (Fig. 4B). The recombinant virus with the substitution A719T exhibited an impaired growth in cell culture compared to the WT on days 5, 6, 7 (P < 0.001 for the three), and 8 (P < 0.05) postinfection. Furthermore, we previously showed that the substitutions Q579I and V715S were also associated with a decrease in the replicative capacity of HCMV mutants by 7.7- and 1.9-fold, respectively, compared to the WT on day 8 postinfection (20, 21). Thus, substitutions located in helix K and region II of the DNA pol conferring resistance (UL30/UL54) or hypersusceptibility (UL54) to PFA were associated with a decrease in viral replicative capacity.

Enzymatic activity of DNA polymerases and inhibition by foscarnet.

The impact of the amino acid substitutions conferring resistance to PFA on the kinetics parameters of DNA pol activity and their drug inhibition was evaluated by a filter-based assay. The UL30 protein with the substitution I619K showed a 0.42-fold decrease in the polymerase activity efficiency (ratio of apparent maximal velocity [V_max(app)] to apparent Michaelis-Menten constant [K_m(app)]) compared to the WT, whereas that of the A719T mutant remained unchanged (0.97-fold) (Table 3). We previously showed that UL30 recombinant protein harboring the substitution V715S is also associated with a 0.50-fold decrease in the polymerase activity efficiency compared to the WT (21). As expected from the drug susceptibility testing, the apparent constant of inhibition (K_i(app)) of PFA against UL30 proteins with the substitutions I619K, V715S, and A719T increased by 4.6-fold, 20.6-fold (21), and 5.0-fold, respectively, compared to that of the WT.

TABLE 3.

Steady-state kinetics parameters and drug inhibition of the wild type and recombinant DNA pol mutants exhibiting resistance to foscarnet^a

Recombinant protein	Steady-state kinetics parameters				Inhibition by PFA
	Vmax(app) (fmol/min)	Km(app) (μM)	Vmax(app)/ Km(app)	Fold changeb	Ki(app) (μM)c	Fold changeb
UL30-WT	26.39 ± 1.6	0.98 ± 0.27	27.04	1.00	0.22 ± 0.07	1.00
UL30-I619K	27.7 ± 11.96 ns	2.45 ± 1.19 ns	11.30	0.42	0.97 ± 0.28*	4.6
UL30-A719T	71.2 ± 11.09**	2.71 ± 0.85 ns	26.29	0.97	1.08 ± 0.54*	5.0
UL54-WT	5.41 ± 2.47	0.60 ± 0.16	8.98	1.00	0.25 ± 0.04	1.00
UL54-Q697P	9.30 ± 4.37 ns	3.5 ± 0.14*	2.7	0.29	0.83 ± 0.08**	3.4
UL54-A719T	22.7 ± 9.1*	3.5 ± 1.2*	6.5	0.72	1.1 ± 0.15**	4.6

^aWT, wild type; V_max(app), apparent maximum velocity; K_m(app), apparent Michaelis-Menten constant; K_i(app), apparent constant of inhibition. A one-way ANOVA with Dunnett’s posttest (GraphPad Prism, version 8.00) was used to compare the enzymatic parameters of mutant proteins to those of the WT. *, P < 0.05; **, P < 0.01; ns, nonsignificant compared to WT. Results are means ± SD from 3 independent experiments.

^bFold change compared to WT values.

^cK_i values that are significantly different from those of the WT were considered to induce drug resistance.

The apparent polymerase activity efficiencies of UL54 proteins with the substitutions Q697P and A719T were 0.29- and 0.72-fold lower, respectively, than that of the WT (Table 3). The polymerase activity efficiencies of recombinant UL54 proteins containing the substitutions Q579I and V715S were also decreased by 0.06- and 0.2-fold, respectively, compared to that of the WT, as reported elsewhere (20, 21). In accordance with the EC₅₀s of PFA against the corresponding recombinant HCMV mutants, the apparent K_i of PFA against UL54 proteins with the substitutions Q697P, V715S, and A719T increased by 3.4-fold, 16.0-fold (21), and 4.6-fold, respectively, compared to the WT. We also determined that the UL54 protein containing the substitution Q579I was hypersusceptible to PFA, as its apparent K_i value was also decreased by 0.37-fold compared to WT (20). Thus, the resistance or hypersusceptibility of HSV-1 and HCMV to PFA conferred by substitutions located in helix K and region II of the DNA pol is reflected at the protein level, as demonstrated by the changes in apparent K_i values of the drug, and both were associated with decreased polymerase activity efficiencies.

Modeling of DNA polymerase mutants resistant to foscarnet.

Based on the crystal structure of the open form of UL30 DNA pol (PDB 2GV9), the I619K substitution might introduce an additional salt bridge with D615 side chain or an H bond with its main carbonyl chain, which might affect the salt bridge between D615 and K516 residues seen in the WT protein (Fig. 5A). The interaction between R611 and D298 is not affected by the substitution. In the closed-form model, however, K619 might affect R611, resulting in an interaction with D615 which may affect the orientation of the main chain of D615 compared to the open form (Fig. 5B). Another consideration is that, in the closed form we modeled, the beta-strand segment from R500 to K516 contains several positively charged residues (R500, R512, K514, and K516), which might shift the conformation of D615 and make the interaction with K619 less desirable than with R611. Thus, the I619K mutant may adopt an open rather than a closed conformation of UL30 DNA pol. In contrast, the substitution Q579I, which was previously associated with hypersusceptibility of HCMV to PFA, stabilizes the closed conformation of UL54 and shows a higher affinity of the drug for the protein (20).

FIG 5.

Local environments of amino acid substitutions conferring resistance to foscarnet in the helix K of UL30 DNA polymerase. The cartoon representations of UL30 open form (crystal structure; PDB 2GV9) and UL30 closed form are colored in wheat and pale green, respectively. (A) The I619K substitution in UL30 DNA pol may stabilize the open conformation due to the additional salt bridge formed between K619 and D615. (B) Model of the closed form of UL30 DNA pol indicating that the I619K substitution may destabilize the local structure due to its proximity to R611, because D615 may rotate toward a cluster of positively charged residues (e.g., R500 and R512) upon the transition from the open-to-closed conformation. The H bonds are shown as dashed lines. The carbon atoms of mutated residues (labeled in blue and in brackets) are in cyan. The DNA molecule is derived from the structure of the closed form of yeast DNA pol δ (PDB 3IAY). The figure was prepared by using PyMOL (43).

In UL54 DNA pol, the Q697 residue is located in the palm domain, which is important to anchor the backbone of the template DNA in the replication mode (closed form). While V694 is more distantly located from the DNA template in our model, Q697 is in the vicinity of the bound DNA molecule, as is the nearby G698, which establishes an H bond with the DNA backbone (Fig. 6A). Based on our model, Q697 forms an H bond with R839 in UL54 DNA pol. In contrast, the corresponding residue in UL30 is a shorter threonine (T839) that is not able to establish such an H bond. This explains why the Q697P substitution in UL54 will probably make the DNA pol less stable in the closed form, leading to an increased PFA EC₅₀ for the HCMV mutant.

FIG 6.

Local environments of amino acid substitutions conferring resistance to foscarnet in region II of UL54 DNA polymerase. The cartoon representation of the UL54 closed form is in light blue. (A) The residue Q697 is in the vicinity of the DNA molecule (G698 establishes an H bond with the DNA backbone), whereas V694 is distant from the bound DNA molecule. Q697 forms an H bond with R839 in the palm domain in UL54, whereas this interaction is absent in UL30 due to a much shorter T839 residue at the corresponding position. (B) The environment of A719 is shown in the closed form model of UL54 DNA pol. The salt bridge formed between R788 and E949 seems important to bring the fingers domain (R788) closer. The substitution A719T may result in new H bonds (shown as green dashed lines) between T719 and E949 and could thus decrease the interaction of the latter with R788. The H bonds are shown as dashed lines. The carbon atoms of mutated residues (labeled in blue and in brackets) are in cyan. The DNA molecule is derived from the structure of the closed form of yeast DNA pol δ (PDB 3IAY). The magnesium ions are shown as green spheres, while the foscarnet molecule (PFA) is shown in stick mode (phosphate, carbon, and oxygen atoms are represented with orange, purple, and red sticks, respectively), both of which were derived from the crystal structure of the RB69 DNA pol (PDB 3KD5) upon superposition. The figure was prepared by using PyMOL (43).

Based on the DNA pol structures in the replication mode (closed form), residue A719, located in the palm domain, is close to the DNA binding site. In the immediate environment of A719, upon the open-to-closed transition in UL54 (or UL30), the positively charged residue R788 (or R789 in UL30) in the fingers domain moves toward the palm domain and forms a salt bridge with E949 (or E925 in UL30) (Fig. 6B). This salt bridge likely contributes to the stabilization of the closed conformation, as indicated by our models of both UL30 and UL54 in closed forms. The A719T substitution in both UL54 and UL30 will likely disturb the above-mentioned salt bridge, as the hydroxyl group of T719 would be within the H-bonding distance of the carboxylate group of E949 (or E925) and may change the conformation of the latter. Therefore, the substitution A719T would have deleterious effects on the transition of open-to-closed conformation or lead to a less stable closed form, which would result in increased EC₅₀s of PFA for both mutant viruses. Substitution V715S, located in region II of DNA pol, is also associated with PFA resistance for both viruses. The V715 residue is close to the catalytic center of UL30 and UL54 DNA pol. The V715S substitution may disrupt local hydrophobic interactions. The destabilization of the local structure of the enzyme may affect the D717 residue that is critical to anchor the Mg²⁺ ions required for the binding of PFA, as reported by our group (21).

Thus, UL30 and UL54 DNA pol harboring amino acid substitutions in the helix K and region II conferring resistance to PFA are less stable in the closed conformation and favor a more open form of the protein for which the drug has a lower affinity. In contrast, the Q579I substitution in helix K of UL54 DNA pol associated with hypersusceptibility to PFA stabilizes the closed conformation, which increases drug binding.

DISCUSSION

Amino acid substitutions conferring PFA resistance to HCMV are mainly distributed in the palm, fingers, and NH₂-terminal domains of UL54 DNA pol, whereas those associated with cross-resistance to PFA and GCV are located in the fingers domain (22). This suggests an important role of the fingers domain in the susceptibility of HCMV to PFA. The fingers domain is involved in the open-to-closed conformational changes of the DNA pol that occur following binding of the incoming nucleotide to the enzyme. Several studies have demonstrated that amino acid substitutions located in helices N and P of the fingers domain of UL54 DNA pol affect the susceptibility of HCMV to PFA (16, 17, 19). Most of these substitutions in helices N (W780V, V781I, and V787E) and P (L802M and A809V) confer resistance to PFA by favoring the open form of the enzyme, to which the antiviral binds with a lower affinity. In contrast, the K805Q substitution in helix P is more stable in the closed conformation and is associated with a hypersusceptibility phenotype. In this study, we evaluated whether theoretical amino acid substitutions located in the helix K (NH₂-terminal domain) and region II (palm domain) that interact with the fingers domain during the open-to-closed conformational changes of the DNA pol affect the susceptibility of HSV-1 and HCMV to PFA. Our results showed that some amino acid substitutions in helix K and region II of UL30 and UL54 DNA pol can stabilize the proteins either in their open or closed conformation and thereby influence the susceptibility of HSV-1 and HCMV to PFA.

As already reported for amino acid changes in the helix P of the fingers domains (PFA resistance for L802M and A809V versus hypersusceptibility for K805Q) (17), substitutions located in helix K of the NH₂-terminal domain can also be associated with two opposite drug susceptibility phenotypes. Indeed, the substitution I619K in UL30 DNA pol confers PFA resistance to HSV-1. Examination of the 3D models suggests that the mutated protein may adopt an open form rather than a closed conformation, which decreases the binding affinity of PFA. In contrast, the substitution Q579I in UL54 DNA pol, and its equivalent, Q618I, in UL30, conferred increased PFA susceptibility to HCMV and HSV-1 by favoring a closed conformation, thereby increasing drug binding to the enzyme (20).

In conserved region II, a series of amino acid substitutions in UL30 (i.e., R700G, V715G, and S724N) and UL54 (i.e., T700A, V715A/M, and I726V) DNA pol were previously reported to confer PFA resistance to HSV-1 and HCMV (23–29). In our study, the substitutions Q697P (UL54) and A719T (both UL30 and UL54) were associated with a decreased susceptibility to PFA. Three-dimensional modeling analyses comparing the mutated residues in the open and closed forms of the DNA pol revealed that these substitutions are less stable in the closed conformation and promote the open form of the enzyme with a reduced affinity for PFA.

The palm domain of the DNA pol of herpesviruses, which is composed of conserved regions II, I, and VII, corresponds to the catalytic center of the enzyme. Region II contains a highly conserved DXXSLYPS motif (amino acids [aa] 717 to 724) (30). This motif, and especially the D717 and F718 residues, are implicated in the coordination of the two metal ions that are essential for the reaction of DNA polymerization (31). It was reported that PFA establishes interactions with residues R784 and K811 of the fingers domain of UL54 DNA pol (corresponding to residues R785 and K811 in UL30) and chelates metal ion B, which interacts with the triphosphate tail of the incoming nucleotide (15). We previously reported that recombinant HCMV mutants harboring substitutions V715M/S are resistant to PFA (21). The V715 residue is close to the catalytic aspartate D717. Three-dimensional modeling analysis showed that the substitutions V715M/S in UL54 DNA pol may hinder the conformational change of the catalytic site of the enzyme that brings the carboxylate residue of D717 in contact with metal ion B, thereby reducing the binding of PFA. In HSV-1, the substitutions V715G/S in UL30 DNA pol were also associated with resistance to PFA by a similar mechanism.

Several parameters involved in the DNA polymerization process (e.g., binding affinity of primer-template DNA and/or nucleotide for the enzyme, efficiency of nucleotide incorporation, fidelity of DNA replication, and processivity of viral DNA synthesis) could be altered in UL30 and UL54 DNA pol mutants, leading to reduced viral fitness compared to WT counterparts, as shown in this study. Furthermore, a series of recombinant HSV-1 and HCMV strains harboring theoretical amino acid substitutions in helix K (Q617P, R620T, and L626F in UL30 and their equivalents Q578P, R581T, and L587F in UL54) and region II (F718L in UL30 and P712Y and F718L in UL54) were unable to grow in cell culture. Despite the fact that the recombinant mutant harboring the Q578P substitution exhibited a growth defect in cell culture, HCMV strains carrying other amino acid substitutions at the same position in UL54 DNA pol, such as Q578H/L, displayed normal growth (23, 32). This suggests that the nonconservative change of the Q578 residue to a proline could markedly alter the structure of the DNA pol compared with a more conservative change for a valine or a leucine. Three-dimensional modeling analyses revealed that this substitution may disrupt the local structure of the DNA pol and affect the binding of the DNA molecule. A lack of viral growth was also observed for several recombinant HCMV mutants harboring the substitutions R581T, F718L, and F718S in UL54 DNA pol (23). The R581 residue is involved in the binding of the DNA molecule, and its change to a threonine may alter the local structure of the protein, thereby reducing this interaction. Due to its close proximity to the catalytic aspartate D717, the substitution F718L may alter the interaction and the correct positioning of metal ion B, which is essential for the enzyme activity (31). Overall, these amino acid changes may disrupt the local structure of the DNA pol and lead to reduced enzyme activity, which may explain the defect in viral growth. We cannot exclude the possibility, however, that the alterations of the local structure of the DNA pol could also prevent its interactions with other proteins involved in the replication complex, such as their processivity factors (UL42 and UL44 for UL30 and UL54 DNA pol, respectively).

Here, we emphasize the importance of the interactions between the fingers, the NH₂-terminal, and the palm domains during the open-to-closed conformational changes of UL30 and UL54 DNA pol. As previously reported for helices N and P of the fingers domain, our results demonstrate that destabilization of the DNA pol conformation to the open form induced by amino acid substitutions in helix K and region II could be a potential mechanism of PFA resistance. In helix K, amino acid changes could also favor the closed conformation of the DNA pol to which PFA binds with a higher affinity, resulting in a hypersusceptibility phenotype, as previously shown for the K805Q substitution. The identification of key amino acids that are critical for the correct open-to-closed conformational changes of the DNA pol could allow the design of new antivirals aimed at preventing the fingers domain dynamics required for optimal activity of the enzyme.

MATERIALS AND METHODS

Cells.

African green monkey kidney cells (Vero) and human lung fibroblasts (MRC-5) were obtained from the American Type Culture Collection (Manassas, VA). Cells were maintained in minimal essential medium (MEM) plus 10% fetal bovine serum (FBS) (both from Gibco/Invitrogen, Burlington, Ontario, Canada) in a 5% CO₂ atmosphere.

Generation of recombinant HSV-1.

The HSV-1 genome (strain 17), split into 5 overlapping DNA cosmids (cos24, cos32, cos48, cos51, and cos71, obtained from C. Cunningham [MRC Virology Unit, Glasgow, United Kingdom]) (33), was used in this study. The cosmid 71 was replaced by three plasmids (pNEB23, pNEB10, and pPOL6) (34) with pPOL6 plasmid containing the UL30 gene. Selected mutations were introduced into the pPOL6 plasmid using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). To reconstitute the entire HSV-1 genome, Vero cells seeded in 6-well plates were cotransfected with the set of overlapping cosmids and plasmids using Lipofectamine 2000 (Life Technologies, Burlington, Ontario, Canada). For each substitution, two sets of cosmids and plasmids were cotransfected in Vero cells to generate two independent recombinant viruses (two independent clones). The entire UL30 gene in each mutated pPOL6 plasmids and recombinant viruses was sequenced to verify that only the expected mutation was introduced. Virus titers were determined by plaque assays on Vero cells.

Generation of recombinant HCMV.

The pHB5 bacmid, containing the genome of the HCMV strain AD169 and a bacterial artificial chromosome (BAC) (obtained from the laboratory of M. Messerle [Max von Pettenkofer Institut, Munich, Germany]) was used in this study (35, 36). We previously integrated the Gaussia luciferase gene in the pHB5 bacmid (pHB5-GLuc) under the control of the HCMV major immediate early promoter (37). Mutations were introduced into the UL54 gene using “en passant mutagenesis” (37). For each substitution, two bacmids were transfected in MRC-5 cells to generate two independent recombinant viruses (two independent clones). MRC-5 cells were cotransfected with WT or mutant pHB5-GLuc and the pBKCMV82 plasmid, which contains the UL82 gene, encoding the pp71 protein. This protein was shown to increase the infectivity of the virus in vitro (38). The entire UL54 gene in each mutated pHB5-GLuc and recombinant viruses was sequenced to verify that only the expected mutation was introduced. Virus titers were determined by plaque assays on MRC-5 cells.

Susceptibility of HSV-1 to antivirals.

Drug susceptibilities of recombinant HSV-1 were determined by plaque reduction assays on Vero cells. Briefly, cells seeded in 24-well plate (at 80% confluence) were infected with recombinant viruses at an MOI (multiplicity of infection) of 0.0001 for 1.5 h at 37°C in a 5% CO₂ atmosphere. Cells were then incubated with increasing concentrations of PFA (0 to 200 μM [drug-susceptible viruses] or 0 to 400 μM [drug-resistant viruses]) or ACV (0 to 1 μM) (both from Sigma-Aldrich, St. Louis, MO) in MEM plus 2% FBS containing 0.4% SeaPlaque agarose (Lonza, Rockland, ME). After 3 days, cells were fixed with 4% formalin and stained with crystal violet. Cytopathic effects were counted under an inverted microscope to determine the EC₅₀ that corresponds to the concentration of antiviral that reduces plaque numbers by 50% compared to cells without drugs. We considered that a recombinant HSV-1 mutant was resistant to a drug when its mean EC₅₀ was significantly different from that of the WT as assessed by a one-way analysis of variance (ANOVA) with Dunnett’s posttest (see below).

Susceptibility of HCMV to antivirals.

Drug susceptibilities of recombinant HCMV were determined on MRC-5 cells using the GLuc reporter-based assay (37). Cells seeded in 24-well plate (at 80% confluence) were infected with recombinant viruses at an MOI of 0.001 for 1.5 h at 37°C in a 5% CO₂ atmosphere. Cells were then incubated with increasing concentrations of PFA (0 to 400 μM) or GCV (0 to 64 μM) (Sigma-Aldrich) in MEM plus 2% FBS. Volumes of 30 μl of cell culture supernatants were transferred to a black 96-well plate. The coelenterazine (NanoLight Technology, Pinetop, AZ) substrate (50 μl) was added to each well (37), and the GLuc activity was measured using a Victor 3000 plate reader (Victor³; PerkinElmer, Waltham, MA) with an acquisition time of 1 s. The EC₅₀, corresponding to the concentration of antiviral that reduces the GLuc activity by 50% compared to cells without drugs, was determined. We considered that a recombinant HCMV mutant was resistant to a drug when its mean EC₅₀ was significantly different from that of the WT as assessed by one-way ANOVA with Dunnett’s posttest (see below).

Replicative capacity of HSV-1 by real-time cell analysis.

The RTCA system (xCELLigence; ACEA Biosciences, Inc., San Diego, CA) was used to determine the replicative capacity of the WT and the different HSV-1 mutants in Vero cells (39). Cells were seeded in a 96-well E-plate that contains a microelectronic sensor array which allows the system to measure the impedance. The differences in the impedance over time reflect changes in the cell growth and morphology and are represented by a parameter called the cell index (CI). After 48 h, cells (at 80% confluence) were infected with recombinant viruses at an MOI of 0.01 in MEM plus 2% FBS for 1.5 h at 37°C in a 5% CO₂ atmosphere. Viral suspensions were then removed, and cells were incubated with MEM plus 2% FBS for an additional 3 days. The impedance was recorded every 30 min. Data were determined as 1/cell index, measured at selected time points to better compare the replicative capacity of the different recombinant viruses.

Replicative capacity of HCMV by the Gaussia luciferase reporter-based assay.

The replicative capacity of recombinant HCMV was determined using the GLuc reporter-based assay (37). MRC-5 cells were seeded in a 24-well plate. After 48 h, cells (at 80% confluence) were incubated with the WT and the different recombinant mutants (MOI of 0.001) for 1.5 h at 37°C in a 5% CO₂ atmosphere. Viral suspensions were then removed and replaced with MEM plus 2% FBS. Cell culture supernatants (30 μl) were collected daily for 8 days and transferred to a black 96-well plate to determine the GLuc activity as described above.

Enzymatic activity of DNA polymerases and inhibition by foscarnet.

The genes UL30 (from HSV-1 strain 17; GenBank number X14112.1) and UL54 (from HCMV strain AD169; GenBank number X17403.1) were cloned into the pCITE4a plasmid (EMD BioScience, San Diego, CA) using NdeI/NcoI and NdeI/SacI restriction sites, respectively (16, 40). Selected mutations were introduced into UL30 or UL54 using a QuikChange site-directed mutagenesis kit (Stratagene). Both proteins were expressed by an in vitro transcription-translation system using reticulocyte lysates of rabbits (Promega Biosystems, Sunnyvale, CA) as previously reported (41). Protein expression was performed at 37°C for 3 h, and 0.2 mg/ml of bovine serum albumin was added at the end of the reaction.

The steady-state kinetics parameters of viral DNA pol and their inhibition by PFA were then determined by a filter-based assay (17). Activated calf thymus DNA was used as the template. The reaction was performed in a mixture containing 25 mM Tris-HCl (pH 8.0), 90 mM NaCl, 0.5 mM dithiothreitol, 5% glycerol, 1 μM [³H]dGTP (specific activity, 8 to 13 Ci/mmol; Moravek, Brea, CA), increasing concentrations (0 to 5 μM) of the other deoxynucleoside triphosphates (dNTPs) (dATP, dCTP, and dTTP; New England BioLabs, Ltd., Whitby, Ontario, Canada), 0.1 mg/ml activated calf thymus DNA (Amersham Biosciences, Piscataway, NJ), and 1 μl of the enzyme-containing transcription-translation mixture. The total volume was then adjusted to 25 μl with water. The reaction was started by the addition of 10 μl MgCl₂ (10 mM final concentration), and mixtures were incubated at 37°C for 30 min; reactions were stopped by the addition of 600 μl prechilled 5% trichloroacetic acid (TCA)–1% sodium pyrophosphate (NaPP_i). Reaction mixtures were maintained on ice for 10 min to precipitate the DNA and then filtered through a Millipore Multiscreen 1.2-μm-pore-size glass fiber C 96-well filtration plate (EMD Millipore, Billerica, MA). Filters were then washed twice with 250 μl of prechilled 5% TCA–1% NaPP_i and with approximately 350 μl of 99% ethanol. Drug inhibition of DNA pol activity was determined by adding in the reaction mixture described above increasing concentrations of PFA (0 to 4 μΜ) while maintaining a total volume of 25 μl. At the end of the reaction, filters were removed and placed in vials containing Ecolite(+) scintillation liquid cocktail (MP Biomedicals, Solon, OH) to measure radioactivity.

The average observed rates of enzymatic activity versus substrate concentrations data (in the absence and presence of PFA) were fitted simultaneously according to the nonlinear equation for noncompetitive inhibition to calculate the apparent maximal velocity (V_max(app)), the apparent Michaelis-Menten constant (K_m(app)), and the apparent constant of inhibition (K_i(app)) by the use of GraphPad Prism software (version 8.00; GraphPad Software Inc., San Diego, CA). The polymerase activity efficiency was calculated as the ratio of V_max(app) to K_m(app).

Three-dimensional modeling.

The crystal structure of the open form of HSV-1 UL30 DNA pol (PDB 2GV9) was used. The open form of HCMV was obtained by homology modeling based on the 2GV9 structure. The structure of herpesvirus DNA pol bound to a primer-template has not been resolved yet. The closed forms of both HCMV and HSV-1 DNA pol were thus built by homology modeling with the crystal structure of the closed form of the DNA pol δ of Saccharomyces cerevisiae (which is closely related to the family B DNA pol) bound to a primer-template DNA and an incoming nucleotide (PDB 3IAY). The DNA molecule was derived from the structure of the closed form of yeast DNA pol δ (PDB 3IAY). In some analyses, the magnesium ions and the PFA molecule derived from the crystal structure of the RB69 DNA pol (PDB 3KD5) were included in the closed form model of UL54 upon superposition. The homology modeling was performed using the Swiss-Model server (42).

Statistical analyses.

The GraphPad Prism software (version 8.00) was used for all statistical analyses. A two-way ANOVA with Bonferroni’s post hoc test was used to compare the replicative capacities of recombinant mutant viruses to those of the WT. A one-way ANOVA with Dunnett’s posttest was used to compare EC₅₀s and kinetics parameters of mutant viruses to those of the WT. A P value of <0.05 was considered statistically significant.

Data availability.

The data sets generated and analyzed during the current study are available from the corresponding author upon request.