N²-Phenyl-9-(hydroxyalkyl)guanines and related compounds are substrates for Herpes simplex virus thymidine kinases

Abstract

Herpes simplex virus (HSV) types 1 and 2 thymidine kinases (TK) are responsible for phosphorylation of antiherpes acyclonucleosides such as acyclovir (ACV) and 9-(4-hydroxybutyl)guanine (HBG). Related compounds, the N²-phenyl-9-(hydroxyalkyl)guanines, are devoid of direct antiviral activity, but potently inhibit the viral TKs and block viral reactivation from latency in vivo. The similarity in structure between the acyclonucleosides and TK inhibitors raised the question of the relevance of phosphorylation of certain of the latter analogs in their mechanisms of action. Using recombinant TKs and HPLC analysis of reaction mixtures, we report that the lead TK inhibitor N²-phenyl-9 -(4-hydroxybutyl)guanine (HBPG) and its pentyl homolog (HPnPG) are excellent substrates for the enzymes, approaching the efficiency with which the natural substrate thymidine is phosphorylated, and significantly better than ACV or HBG. Other 9-hydroxyalkyl congeners are substrates for the enzymes, but with much poorer efficiency. HBPG triphosphate was a poor inhibitor of HSV DNA polymerase, the target of acyclonucleoside triphosphates, suggesting that phosphorylation of HBPG is not important in its mechanism of blocking viral reactivation in vivo. The fact that HBPG is an efficient substrate is consistent, however, with its binding mode based both on molecular modeling studies and x-ray structure of the HBPG:TK complex.

Introduction

We recently reported that a potent inhibitor of Herpes simplex virus (HSV) types 1 and 2 thymidine kinases (TK), N²-phenyl-9-(4-hydroxybutyl)guanine (HBPG; see structures), is an efficient substrate for the enzymes (Manikowski et al. 2005). In addition, this compound was the only derivative among a related series that demonstrated potent antiviral activity against experimental models of HSV reactivation and encephalitis in mice (Manikowski et al. 2005). In order to evaluate the significance of phosphorylation of HBPG in its mechanism of activity in vivo, we have compared the kinetic parameters of phosphorylation of HBPG with those for the natural substrate thymidine (TdR) and the antiherpes drug acyclovir (ACV). In addition, we screened related compounds for their ability to act as substrates for the HSV TKs. Finally, we compared the inhibitory potencies of HBPG-MP and the synthetic triphosphate HBPG-TP with that of ACV-TP on HSV-1 DNA polymerase. The results are consistent with the probability that HBPG does not owe its anti-reactivation activity in the animal models to its conversion to the monophosphate form in virus-infected cells.

Material and Methods

Chemical compounds

Synthesis of N² phenyl-9-(hydroxyalkyl)-6-oxopurines and HBPG-MP are described in Manikowski et al. (2005). 9-(4-Hydroxybutyl) guanine (HBG) was prepared as described by Larsson et al. (1983). Acyclovir mono and triphosphates (ACV-MP and ACV-TP) were prepared by the method of Sawai et al. (2002); the products were identical with those reported by Furman et al. (1979).

N²-Phenyl-9-(4-triphosphoryloxybutyl)guanine sodium salt, HBPG-TP, was prepared as follows. The free acid of HBPG-MP (47 mg, 123 mmol) was converted to its trioctylammonium salt by dissolving it in N,N-dimethylformamide (5 mL) containing trioctylamine (1.5 eq., 65 mg, 184 mmol), followed by evaporation of the solvent to dryness. A suspension of the residue in hexamethylphosphoric triamide (HMPA) (680 mL) was treated with 1,1′-carbonyldiimidazole (100 mg, 613 mmol) under N₂, and, after 3.5 h, the clear reaction mixture was quenched by the addition of MeOH (92 mL). A solution of trioctylammonium pyrophosphate (613 mmol) [prepared from tetrasodium pyrophosphate (163 mg, 613 mmol) and trioctylamine (542 mg, 1.53 mmol)] in HMPA (4.9 mL)] was added dropwise with stirring. Precipitation occurred almost immediately. After 24 h at rt, the mixture was quenched with 0.02 M aqueous triethylammonium bicarbonate (15 mL) and extracted with Et₂O (2 × 20 mL). The aqueous material was applied to a DEAE-Sephadex column (20 × 2.2 cm), bicarbonate form, and elution was carried out in a linear gradient of 0.05–1.0 M aqueous triethylammonium bicarbonate, pH 7.9, at a flow rate of 4.0 mL/min, with detection at 260 nm. The product eluted in fractions 86–97 (15 mL/fraction). The fractions were combined and coevaporated with 1-butanol at 30 °C under vacuum, and the resulting material was dissolved in water and passed through 15 mL of Dowex 50Wx8, Na⁺ form. The eluate was lyophilized, the residue dissolved in water, and passed over chelating resin (Chelex 100). The eluate was lyophilized to give 27 mg (36%) of HBPG-TP as the sodium salt. ¹H NMR (300 MHz, D₂O) d 7.84 (s, 1H), 7.49 (m, 2H), 7.38 (m, 2H), 7.14 (m, 1H), 4.08 (t, 2H, J = 7.0 Hz), 3.97 (q, 2H, J = 6.4 Hz), 1.97–1.86 (m, 2H), 1.67–1.53 (m, 2H); ³¹P NMR (121 MHz, D₂O) d −4.82 (d, J = 19.3 Hz), −9.37 (d, J = 19.3 Hz), −20.76 (t, J = 19.3 Hz).

HSV thymidine kinases and substrate assay

HSV-1 and HSV-2 TKs were expressed as His₆-tagged proteins and assayed as described (Manikowski et al. 2005). One unit (U) is the amount of enzyme which phosphorylates 1 nmol of TdR to TMP in 1 hour at 37 °C. Test compounds (200 μM) were incubated at 37 °C for various times in 25 μL of a mixture containing 30 mM HEPES-K⁺, pH 7.5, 6 mM MgCl₂, 6 mM ATP, 0.5 mM DTT and approximately 2 U of HSV-1 TK or 10 U of HSV-2 TK. Samples were heated at 100 °C for 5 min and centrifuged for 15 min at 10,000 rpm in a microcentrifuge. Supernatants were transferred to a new tube for subsequent HPLC analysis. Reverse phase chromatography employing a Shimadzu HPLC system was used to separate test compounds from phosphorylated products. A 4.6 mm × 25 cm ALL-TIMA C18-NUC 100A 5U (Alltech) column was used at rt in the following conditions: injection volume, 20 μL; detection, UV 260 nm; eluents, buffer A (20 mM KH₂PO₄, pH 7.5), buffer B (20 mM KH₂PO₄, pH 5.2, 60% MeOH); linear gradient, 30 min from 0% to 100% buffer B; 20 min 100% buffer B; flow rate, 0.5 ml/min. Retention times (R_t) for compounds and their monophosphates are summarized in Table 1.

Table 1.

Conversion of 9-hydroxyalkyl-6-oxopurines to monophosphates (MP) by HSV TKs.

Cpd acronym	HPLC Rt, min1		Relative % MP2
	Parent	MP	HSV1 TK	HSV2 TK
HBPG	38.46	31.79	1003	1003
ACV	25.18	14.56	32	5
HBG4	20.36	13.25	3	14
HPrPG	38.26	nd	nd	nd
HPnPG	44.22	36.01	124	58
HBP(N-Me)G	39.66	33.34	15	11
HBPOHx	36.23	29.82	22	16
HBPSHx	40.70	34.35	12	5

¹R_t = retention time in min; HPLC conditions in Experimental Section.

²Cpds assayed at 200 mM for 40 min in the presence of ca. 2 U of HSV-1 TK or 8 U of HSV-2 TK.

³100% for HBPG corresponds to 25% conversion to HBPG-MP for both enzymes.

⁴Assayed with 10 U of enzymes.

nd, not detected.

HSV-1 DNA polymerase and assay

HSV-1 DNA polymerase was purified from HSV-1 infected HeLa cells and assayed as previously described (Focher et al. 1992). Inhibitor studies (K_i determinations) were done using the above assay conditions with [³H]dGTP (1,500 cpm/pmol) at five concentrations (between 1.5 and 10 μM), in the presence of increasing concentrations of inhibitor.

Results

Substrate assays of HSV thymidine kinases

Thymidine kinases were prepared by expressing the cloned genes for HSV-1 and HSV-2 enzymes containing a His₆ tag (Manikowski et al. 2005). Properties, including K_m values for the substrate TdR and inhibitor sensitivities, were similar to those observed with enzymes isolated from virus-infected cells (Focher et al. 1988, Hildebrand et al. 1990).

In the assays of analogs with the enzymes, each analog (200 μM) was incubated for the indicated time in the assay mixture, and the products of the reaction were resolved by HPLC as described in the Experimental Section. Synthetic HBPG-MP was used to confirm the retention time (R_t) of the product, and other monophosphates were identified by their earlier retention times compared with those of the test analogs. Figure 2 illustrates the time-course of phosphorylation of HBPG and the antiherpes compound HBG (Larsson et al. 1983) by HSV-1 TK, resulting in nearly complete conversion of the former compound but only ca. 3% conversion of HBG, emphasizing the profound effect of the phenyl group of HBPG on substrate efficiency. Indeed, 10 units of the enzyme were used in these assays to reveal the weak substrate property of HBG.

Figure 2.

Kinetics of phosphorylation of HBPG (○) and HBG (●) by HSV-1 TK. Reactions were carried out as described in the Experimental Section, using 200 μM of compounds and 10 U of enzyme (in order to measure conversion of HBG). Aliquots of mixtures were analyzed by HPLC, and the % conversion to monophosphate is plotted against time.

Table 1 summarizes the chromatographic behavior of analog:monophosphate pairs and the degrees of the conversion of each to its monophosphate catalyzed by HSV-1 and HSV-2 TKs under standard incubation conditions after 40 minutes incubation. 100% corresponds to conversion of 5 nmoles of substrate to monophosphate. These results showed that HBPG and the related homolog HPnPG were the best substrates for both enzymes, while the lower homolog HPrPG appeared not to be phosphorylated under these conditions. The latter result is consistent with the rigid fit between the TK substrate and phosphorylation (ATP) sites in the enzyme (Bennett et al. 1999). 9-Hydroxybutyl compounds lacking the 2-NH group, the hypoxanthines HBPOHx and HBPSHx, are potent inhibitors of the TKs (Manikowski et al. 2005) and moderate substrates for the enzymes, while the N²-methylated analog HBP(N-Me)G, although a weak TK inhibitor (Manikowski et al. 2005), also was a moderate substrate for the enzymes (Table 1). These results indicate that it is the position of the ω-hydroxyl group in proximity to the γ-phosphoryl group of ATP in the enzyme:compound complex that enables the compound to be a substrate.

Comparison of kinetic parameters of ACV and HBPG with the substrate TdR

As summarized by the results of Table 2 the HSV TKs efficiently phosphorylate the natural substrate TdR, with V_max/K_m of 17 and 7.7 for types 1 and 2 enzymes, respectively. ACV is actually a poor substrate for the enzymes, as evidenced by its very low values of V_max/K_m. HBPG, on the other hand, is comparable with the natural substrate in the efficiency with which it is phosphorylated. The poor substrate property of ACV has been reported (Keller et al. 1981), and the compound owes its potent and selective antiherpes effect to the efficiency with which the limited ACV-MP formed is further activated to the triphosphate, which is a potent, dead-end substrate inhibitor of the HSV DNA polymerase (Reardon & Spector 1989). Given the efficiency with which HPBG is phosphorylated to its monophosphate by HSV TKs, it was logical to consider the results of a parallel pathway to “activation” of this compound to its triphosphate.

Table 2.

Kinetic parameters for phosphorylation of compounds by HSV TKs.

Enzyme	Cpd1	Km, mM	Vmax, mmol/h/mL	Vmax/Km
HSV1 TK	HBPG	12±3	190±15	16
	ACV	122±27	18±2	0.15
	TdR	1.0±0.2	17±2	17
HSV2 TK	HBPG	9.3±0.5	121±2	13
	ACV	200±17	9.0±0.5	0.045
	TdR	2.0±0.2	15.3±0.5	7.7

¹Assays for HBPG and ACV were done with the HPLC method; those for TdR were done with [³H]TdR and the filter method (Manikowski et al, 2005).

Inhibition of HSV-1 DNA polymerase by nucleotide analogs

In order to evaluate whether HBPG inhibits viral reactivation through the inhibition of DNA synthesis by its phosphorylated derivatives, we synthesized HBPG-MP (Manikowski et al. 2005) and HBPG-TP and tested their effects as possible inhibitors of HSV1 DNA polymerase in vitro. Each compound was tested for inhibition of viral polymerase activity in the presence of 5 μM [³H]dGTP (K_m 0.7 μM). Under these conditions HBPG-MP and HBPG-TP were weak DNA polymerase inhibitors, with IC₅₀ values of 0.07 mM and 1.1 mM, respectively. Under the same conditions, ACV-TP potently inhibited the enzyme with IC₅₀ of 7 μM. ACV-TP has been reported to be competitive with dGTP as an inhibitor of HSV DNA polymerase (Reardon & Spector 1989). However, when assayed at variable concentrations of dGTP, HBPG-MP and HBPG-TP inhibited this enzyme in a non-competitive manner, resulting in K_i values of 90 and 800 μM, respectively (data not shown). Thus, not only are the HBPG phosphates weak inhibitors of HSV-1 DNA polymerase, but they also appear to act by a mechanism inconsistent with that of antiviral triphosphates.

Discussion

In this work we report that the HSV TK inhibitor HBPG and closely related compounds are efficient substrates for the types 1 and 2 enzymes. HBPG is the prototype of a series of specific TK inhibitors including many which cannot act as substrates for the enzymes, e.g. N²-phenylguanines lacking a 9-substituent such as N²-(3-[trifluoromethyl]phenyl)guanine (cpd. 122E). It has been thought that the ability of these compounds to block HSV reactivation did not depend on the substrate property of the compounds. Indeed, the most potent guanine derivative 122E also blocked HSV reactivation in a mouse model (Yanachkov et al. 2011).

Compounds that are HSV TK substrates and have direct antiHSV activity include analogs of thymidine and “acyclonucleosides” such as acyclovir (ACV) and its carba-isostere HBG. ACV (Keller et al. 1981) and HBG (Figure 1) are indeed poor substrates for the HSV TKs (see also Table 2), but the low levels of monophosphates formed in HSV-infected cells can readily be converted to di- and tri-phosphates, the latter being the ultimate, potent active antiviral forms (Furman et al. 1979, Larsson et al. 1983). Considering that the triphosphate form of HBPG is not a potent HSV DNA polymerase inhibitor (see above) and that non-substrate inhibitors possess anti-reactivation activity in vivo (Yanachkov et al. 2011), we conclude that phosphorylation of HBPG is not required for its anti-reactivation and, possibly, anti-encephalitic activity in vivo.

Figure 1.

Structures and acronyms of TK inhibitors.

The results reported here, however, are fully consistent with the similar substrate-competent binding modes of HBPG and ACV (and Thd) observed in the solid state structures of HSV-1 TK (Bennett et al. 1999). It is also consistent with the results of molecular modeling studies (Focher et al. 2011) in which 9-substituted N²-phenylguanines were predicted to bind HSV TKs in the substrate mode.