N-Hydroxypiridinedione: A Privileged Heterocycle for Targeting the HBV RNase H

Hepatitis B virus (HBV) remains a global health threat. Ribonuclease H (RNase H), part of the virus polymerase protein, cleaves the pgRNA template during viral genome replication. Inhibition of RNase H activity prevents (+) DNA strand synthesis and results in the accumulation of non-functional genomes, terminating the viral replication cycle. RNase H, though promising, remains an under-explored drug target against HBV. We previously reported the identification of a series of N-hydroxypyridinedione (HPD) imines that effectively inhibit the HBV RNase H. In our effort to further explore the HPD scaffold, we designed, synthesized, and evaluated 18 novel HPD oximes, as well as 4 structurally related minoxidil derivatives and 2 barbituric acid counterparts. The new analogs were docked on the RNase H active site and all proved able to coordinate the two Mg2+ ions in the catalytic site. All of the new HPDs effectively inhibited the viral replication in cell assays exhibiting EC50 values in the low μM range (1.1–7.7 μM) with low cytotoxicity, resulting in selectivity indexes (SI) of up to 92, one of the highest reported to date among HBV RNase H inhibitors. Our findings expand the structure–activity relationships on the HPD scaffold, facilitating the development of even more potent anti-HBV agents.


Introduction
Hepatitis B virus (HBV) poses a severe health burden worldwide.While a safe and effective HBV vaccine exists, current estimates reveal 1.5 million new infections per year around the globe and 296 million chronic HBV patients, resulting in 820,000 deaths annually [1,2].
Current treatments for HBV infection use two primary compound categories: nucleos(t)ide analogs (NAs) and pegylated interferon alpha [3,4].Interferon alpha treatment use is limited due to its subcutaneous injection and serious side effects including flu-like symptoms, bone marrow suppression, fatigue, and depression, which also decrease the patient adherence to the treatment [5].NAs, mainly represented by entecavir, tenofovir disoproxil fumarate, and tenofovir alafenamide, inhibit the reverse transcriptase (RT) enzymatic activity of the viral polymerase (P) protein.They are administered orally, have an excellent safety profile, and effectively suppress HBV replication, showing a high antiviral potency.Nevertheless, HBsAg clearance is achieved in only 3-5% of patients, after 10 years of NA treatment [6].Since NAs induce profound but incomplete suppression of HBV DNA replication but have no direct effect on the viral cccDNA that templates all viral transcripts, drug administration is often life-long and viral replication usually resurges on treatment cessation [7,8].Considering the limitations of existing HBV treatments, there is a compelling need to develop drugs that employ alternative stages in the virus's replication cycle [9][10][11][12][13].
HBV replicates via reverse transcription.The viral polymerase (P) protein has two distinct enzymatic domains, the RT and ribonuclease H (RNase H).The RT catalyzes synthesis of the negative polarity DNA strand [(−) DNA] using the virus pregenomic RNA (pgRNA) as a template.Once the (−) DNA strand has been synthesized inside the viral capsid, the RNase H domain cleaves the pgRNA template strand to expose the newly synthesized (−) DNA strand so it can template synthesis of the (+) DNA strand [14,15].RNase H is a metalloenzyme and acts through a metal-chelation hydrolysis mechanism which requires two Mg 2+ ions coordinated by four carboxylic amino acid moieties (the "D-E-D-D" motif) in the enzyme's active site [16].There is no crystal structure for HBV P or its RNase H domain, so our group generated and validated an HBV P folding model (including the RNase H catalytic domain) using AlphaFold [17].AlphaFold is an AI system from Google DeepMind that can reliably predict the 3D structure of a protein from its amino acid sequence, often achieving an accuracy comparable to experimental methods [18].This model accurately predicts the essential role the two Mg 2+ ions play during the pgRNA hydrolysis mechanism, enabling the application of computational chemistry techniques, for the development of potential RNase H inhibitors [18].Inhibition of the RNase H enzymatic activity prevents the synthesis of (+) DNA and therefore the synthesis of functional viral genomes.Non-functional RNA:DNA heteroduplexes accumulate in the newly synthesized nucleocapsids, terminating viral replication and cccDNA maintenance [19].The HBV RNase H is not targeted by any of the current anti-HBV drugs.
Previously, our group reported several potent HBV RNase H inhibitors belonging to four chemical classes: α-hydroxytropolones (αHTs), N-hydroxyisoqinolinediones (HIDs), N-hydroxypyridinediones (HPDs), and N-hydroxynapthyridinones (HNOs) (Figure 1) [19][20][21][22][23][24][25][26][27].All inhibitors of HBV RNase H possess three electron donors (either O or N atoms) in appropriate positions for chelating the two Mg 2+ ions in the enzyme's catalytic site [28].Among the four inhibitor chemical categories, HPDs have demonstrated the most promising results in terms of inhibiting HBV replication, as well as having favorable druglike properties [21,29,30].We have designed, synthesized, and tested HPDs belonging to two distinct chemical groups: imine HPDs and oxime HPDs.Recently, as part of our ongoing efforts to develop even more potent HBV RNase H inhibitors, we set out to further explore the HPD imine scaffold.We identified several novel potent HPD imines with EC 50 values as low as 1.1 µM and selectivity indexes (SI) of up to 58.These findings represented a substantial improvement over the previously reported most potent HPD imine [30].
Here, we implemented a medicinal chemistry approach aiming at refining the HPD oxime scaffold.Our goal was to deepen our understanding of the structure-activity relationships and to further improve the potency, toxicity, and druglike properties of this compound chemotype.We designed and synthesized 18 novel HPD oximes, leaving the main scaffold (which contains the O "trident" that chelates with the Mg 2+ ions) intact and modifying the side chain.We used A24, the most promising HPD oxime (in terms of SI value, SI = 352 [29]) as our lead compound, and we developed HPDs mostly bearing an aromatic side substitution (Figure 2).We also synthesized HPD oxime analogs with two aromatic side rings as well as alkynes as a side substitution.Additionally, we altered the linker between the main HPD scaffold and the side group.Furthermore, to explore the role of the oxygen "trident" in activity, we synthesized two structurally related barbituric acid analogs and four analogs bearing the main scaffold of the known drug minoxidil.The minoxidil main pyrimidine ring contains three atoms (two N and one O atom) in suitable positions to chelate the Mg 2+ ions of the enzyme active site and has already been proved able to chelate divalent cations [31,32].All new compounds were computationally docked in the RNase H active site of the P structural model and evaluated in cell assays for their ability to inhibit HBV replication.
Molecules 2024, 29,2942 3 of 32 in the RNase H active site of the P structural model and evaluated in cell assays for their ability to inhibit HBV replication.

Chemistry
The novel HPD compounds were synthesized using a three-step synthetic approach (Scheme 1).The first step involved the Gabriel reaction of N-hydroxyphthalimide with benzyl bromides or chlorides to afford the compounds 1-2, 4, and 9-14.The N-hydroxyphthalimides 3 and 5-8 were synthesized using the Mitsunobu reaction, where triphenylphosphine reacts with diisopropyl azodicarboxylate (DIAD) to form a phosphonium intermediate.This intermediate then binds to the oxygen of the alcohol, facilitating a subsequent nucleophilic substitution.Subsequently, the O-substituted N-

Chemistry
The novel HPD compounds were synthesized using a three-step synthetic approach (Scheme 1).The first step involved the Gabriel reaction of N-hydroxyphthalimide with benzyl bromides or chlorides to afford the compounds 1-2, 4, and 9-14.The N-hydroxyphthalimides 3 and 5-8 were synthesized using the Mitsunobu reaction, where triphenylphosphine reacts with diisopropyl azodicarboxylate (DIAD) to form a phosphonium intermediate.This intermediate then binds to the oxygen of the alcohol, facilitating a subsequent nucleophilic substitution.Subsequently, the O-substituted N-

Chemistry
The novel HPD compounds were synthesized using a three-step synthetic approach (Scheme 1).The first step involved the Gabriel reaction of N-hydroxyphthalimide with benzyl bromides or chlorides to afford the compounds 1-2, 4, and 9-14.The N-hydroxyphthalimides 3 and 5-8 were synthesized using the Mitsunobu reaction, where triphenylphosphine reacts with diisopropyl azodicarboxylate (DIAD) to form a phosphonium intermediate.This intermediate then binds to the oxygen of the alcohol, facilitating a subsequent nucleophilic substitution.Subsequently, the O-substituted Nhydroxyphthalimides 1-14 reacted with hydrazine monohydrate to afford the corresponding O-substituted hydroxylamines 15-28.In the final step, the suitable hydroxylamines were condensed with 5-acetyl-1-(benzyloxy)-6-hydroxy-4-methylpyridin-2(1H)-one B in absolute ethanol at reflux to yield the HPD oximes 33-50.Oximes exhibit two geometrical isomers, E and Z.In the present work, no attempt was made to separate the two isomers of the final oxime derivatives and thus the compounds were obtained as a mixture, with a different ratio of the two isomers each time.The existence of the two isomers was confirmed by one-and two-dimensional NMR experiments and the E/Z (or Z/E) isomer ratio was calculated.
Analogs 59-60 of barbituric acid were prepared via a two-step reaction process originating from barbituric acid (Scheme 1).Initially, barbituric acid underwent acetylation using acetic anhydride under reflux.Subsequently, 5-acetyl barbituric acid was combined with O-substituted hydroxylamines (which were prepared as previously outlined) through a coupling reaction in absolute ethanol in the presence of molecular sieves, resulting in the formation of the compounds 59-60.
For the synthesis of the hydroxylamines 29-31, a different synthetic procedure was employed (Scheme 2).The a-bromination of commercially available 3,5-difluoroacetophenone yielded the compound i in 92% yield.NMR analysis revealed the presence of a small quantity of unreacted starting material which was indistinguishable from the product during TLC monitoring of the reaction.Nevertheless, the purity of the crude product was acceptable (94% w/w by NMR analysis), and the impurity would be removed in the subsequent step.The application of the Delepine reaction yielded the hydrochloric salt of the amine ii in 69% yield after recrystallization.The amine was Boc-protected with Boc 2 O/NaHCO 3 in a quantitative yield to afford the intermediate iii which was reduced to the corresponding alcohol iv with NaBH 4 in a satisfactory yield of 78%.The hydroxylamine precursor moiety N-hydroxyphthalimide was then tethered to the compound iv under Mitsunobu reaction conditions to afford the intermediate v in an excellent yield (94%).Subsequently, the Boc group was cleaved with 3M HCl in AcOEt and the free amine vi was coupled to the corresponding heteroaryl acid with TBTU/DIPEA.After workup, the crude amide was subjected to hydrazinolysis or aminolysis with methylamine to afford the final hydroxylamine in good yields (67-92%) over two steps.

Efficacy against HBV Replication and Cytotoxicity
Starting from the most promising HBV RNase H inhibitor we previously reported, A24 [29], we designed and synthesized 18 novel HPD oximes that bear the same pharmacophore scaffold as the hit compound and alterations in the side oxime moiety.Our goal was to further explore the SARs of the HPD scaffold to improve the compounds potency, selectivity, and druglike properties.
The compounds 47 and 48 were the most potent among the newly synthesized compounds (EC50 of 1.1 µΜ).Both compounds feature the same benzyl group side chain as A24 but differ in their 4′-substitution on the side benzene ring; 47 bears a 4′-COOCH3 group, while 48 hosts a 4′-NO2 group.Notably, all three compounds share a polar group at the 4′-position of the side benzene, indicating the potential favorability of incorporating polar groups in this position for enhanced antiviral activity.Moreover, 47 and 48 are minimally cytotoxic, yielding SI values of 87.9 and 91.7, respectively.
The compounds 34 and 46 are characterized by a methyl group positioned on the benzylic methylene, linking the core ring to the side aromatic moiety.This linker group appears pivotal in the inhibitory activity, as evidenced by an up to sixfold reduction in antiviral efficacy (EC50 = 6.3 µΜ for 34) compared to the structurally analogous compound 47 that also features a 4′-COOCH3 substitution.
We also investigated the impact of increasing the size of the side aromatic moiety by incorporating additional aromatic and heteroaromatic rings (the compounds 37, 38, 49, and 50).Remarkably, these compounds were active against HBV replication, with EC50 values ranging from 2.3 to 4.7 µM.Moreover, three out of the four compounds with the bulkier side moiety exhibited no cytotoxic effects, thereby maintaining favorable SI values.
Analogs with a single substitution on the side chain ( 33

Efficacy against HBV Replication and Cytotoxicity
Starting from the most promising HBV RNase H inhibitor we previously reported, A24 [29], we designed and synthesized 18 novel HPD oximes that bear the same pharmacophore scaffold as the hit compound and alterations in the side oxime moiety.Our goal was to further explore the SARs of the HPD scaffold to improve the compounds' potency, selectivity, and druglike properties.
All 18 novel HPDs had EC 50 values in the low µM range (1.1-7.7 µM).Moreover, 15 out of 18 exhibited no significant cytotoxicity in vitro (CC 50 values > 80 µM), resulting in SI values (CC 50 /EC 50 ) ranging from 11.9 to 91.7 (Table 1).50), regardless of the nature of the substitution.Notably, compounds containing halogens exhibited a significant increase in inhibitory activity, particularly when the halogen is positioned at the 4′ location of the aromatic side chain (33,34,47,48).This suggests that nonpolar groups enhance potency and reduce cytotoxicity more effectively than polar groups, such as the hydroxyl group in the compound 39.This is likely due to the hydrophobic loop in the enzyme s active site, which necessitates a lipophilic moiety in the compound, thereby enhancing ligand-protein interactions.
We also tested the significance of the aromaticity by incorporating aliphatic side chains (40,41,42).The results showed the favorability of the four-carbon chain with an EC50 1.2 µΜ for the compound 41 and good SI values.
Overall, HPDs featuring aromatic side chains substituted at the 4′ position with nonpolar groups were the most effective HBV RNase H inhibitors. Conversely, larger polar groups, particularly two or three aromatic rings, as well as the inclusion of a chiral linker, are less well tolerated for antiviral efficacy.Compounds that possess halogen substitutions on the aromatic side chain and a short linker (1 C after the oxime group) between the HPD core pharmacophore ring and the side aromatic moiety exhibit the optimal combination of antiviral activity, minimal cytotoxicity, and favorable SI values.
Our next effort focused on structurally modifying the primary pharmacophore by incorporating the minoxidil structure.Given the structural similarity between minoxidil and our HPD pharmacophore, we hypothesized that the three electron donors (two N atoms and one O atom) would form a "trident" configuration capable of chelating the Mg 2+ ions at the enzyme s catalytic site.Although computational studies (see Section 2.4) suggested potential activity, the compounds 55-58 were inactive against viral replication in vitro.Several hypotheses could explain these results, including the reduced electronegativity of nitrogen compared to oxygen (existent in the HPD heteroatom "trident"), the compounds potential inability to penetrate cell membranes due to increased polarity, or the absence of E/Z isomerism, which might influence the molecule s conformation within the catalytic site.Further optimization could shed light on the potential antiviral activity of these analogs.
Despite their ability to chelate the Mg 2+ ions in the enzyme catalytic site in computational studies (see Section 2.4), the compounds 59 and 60, where the HPD ring was replaced with the pharmacophore ring of barbituric acid, lacked any inhibitory activity.This may be because the oxygen "trident" of the N-hydroxyimide group is essential for coordinating metal ions within the RNase H active site.These findings align with our previous observations for this class of compounds [30].Additionally, the increased polarity of these compounds likely reduces their inability to cross cell membranes to come into contact with the enzyme.To further address the impact of the different substitution patterns and residues in lipophilicity and druggability, we conducted a modeling calculation of druglike properties and descriptors using QikProp module of the Schrödinger platform (Table S1, Supporting Information). 1 LogP (<5) is the log of the partition coefficient of a solute between octanol and water.Predicted with the FAF4 online server [34]. 2 LogD is the log of the partition coefficient of a solute between 1octanol and water at pH 7.4, or physiological pH. 3 tPSA (<140 Å 2 ) is the topological polar surface area (Å 2 ). 4 Fsp 3 , the number of sp 3 hybridized carbons/total carbon count. 5Induced fit docking score to the HBV RNase H active site; kCal/Mol. 6pH 7.4; values in µM. 7CC50/EC50.

Compound Solubility and Apparent Passive Permeability
In total, 21 out of the 24 compounds were highly soluble (solubility limit ≥ 100 µM) in conditions reflecting tissue culture media (pH 7.4).However, the barbituric analogs were insoluble, preventing meaningful biological evaluation (Table 2).In summary, the HPD oximes and minoxidil analogs exhibit promising druglike properties (Tables 1 and S1).Additionally, out of the 24 novel compounds tested in parallel artificial membrane assays, 22 demonstrated a high apparent passive permeability by the industry standard cutoff (>1 × 10 −6 cm/s).This includes all HPD oximes and minoxidil analogs. 1 LogP (<5) is the log of the partition coefficient of a solute between octanol and water.Predicted with the FAF4 online server [34]. 2 LogD is the log of the partition coefficient of a solute between 1octanol and water at pH 7.4, or physiological pH. 3 tPSA (<140 Å 2 ) is the topological polar surface area (Å 2 ). 4 Fsp 3 , the number of sp 3 hybridized carbons/total carbon count. 5Induced fit docking score to the HBV RNase H active site; kCal/Mol. 6pH 7.4; values in µM. 7CC50/EC50.

Compound Solubility and Apparent Passive Permeability
In total, 21 out of the 24 compounds were highly soluble (solubility limit ≥ 100 µM) in conditions reflecting tissue culture media (pH 7.4).However, the barbituric analogs were insoluble, preventing meaningful biological evaluation (Table 2).In summary, the HPD oximes and minoxidil analogs exhibit promising druglike properties (Tables 1 and S1).Additionally, out of the 24 novel compounds tested in parallel artificial membrane assays, 22 demonstrated a high apparent passive permeability by the industry standard cutoff (>1 × 10 −6 cm/s).This includes all HPD oximes and minoxidil analogs. 1 LogP (<5) is the log of the partition coefficient of a solute between octanol and water.Predicted with the FAF4 online server [34]. 2 LogD is the log of the partition coefficient of a solute between 1octanol and water at pH 7.4, or physiological pH. 3 tPSA (<140 Å 2 ) is the topological polar surface area (Å 2 ). 4 Fsp 3 , the number of sp 3 hybridized carbons/total carbon count. 5Induced fit docking score to the HBV RNase H active site; kCal/Mol. 6pH 7.4; values in µM. 7CC50/EC50.

Compound Solubility and Apparent Passive Permeability
In total, 21 out of the 24 compounds were highly soluble (solubility limit ≥ 100 µM) in conditions reflecting tissue culture media (pH 7.4).However, the barbituric analogs were insoluble, preventing meaningful biological evaluation (Table 2).In summary, the HPD oximes and minoxidil analogs exhibit promising druglike properties (Tables 1 and S1).Additionally, out of the 24 novel compounds tested in parallel artificial membrane assays, 22 demonstrated a high apparent passive permeability by the industry standard cutoff (>1 × 10 −6 cm/s).This includes all HPD oximes and minoxidil analogs. 1 LogP (<5) is the log of the partition coefficient of a solute between octanol and water.Predicted with the FAF4 online server [34]. 2 LogD is the log of the partition coefficient of a solute between 1octanol and water at pH 7.4, or physiological pH. 3 tPSA (<140 Å 2 ) is the topological polar surface area (Å 2 ). 4 Fsp 3 , the number of sp 3 hybridized carbons/total carbon count. 5Induced fit docking score to the HBV RNase H active site; kCal/Mol. 6pH 7.4; values in µM. 7CC50/EC50.

Compound Solubility and Apparent Passive Permeability
In total, 21 out of the 24 compounds were highly soluble (solubility limit ≥ 100 µM) in conditions reflecting tissue culture media (pH 7.4).However, the barbituric analogs were insoluble, preventing meaningful biological evaluation (Table 2).In summary, the HPD oximes and minoxidil analogs exhibit promising druglike properties (Tables 1 and S1).Additionally, out of the 24 novel compounds tested in parallel artificial membrane assays, 22 demonstrated a high apparent passive permeability by the industry standard cutoff (>1 × 10 −6 cm/s).This includes all HPD oximes and minoxidil analogs. 1 LogP (<5) is the log of the partition coefficient of a solute between octanol and water.Predicted with the FAF4 online server [34]. 2 LogD is the log of the partition coefficient of a solute between 1-octanol and water at pH 7.4, or physiological pH. 3 tPSA (<140 Å 2 ) is the topological polar surface area (Å 2 ). 4 Fsp 3 , the number of sp 3 hybridized carbons/total carbon count. 5Induced fit docking score to the HBV RNase H active site; kCal/Mol. 6 The compounds 47 and 48 were the most potent among the newly synthesized compounds (EC 50 of 1.1 µM).Both compounds feature the same benzyl group side chain as A24 but differ in their 4 ′ -substitution on the side benzene ring; 47 bears a 4 ′ -COOCH 3 group, while 48 hosts a 4 ′ -NO 2 group.Notably, all three compounds share a polar group at the 4 ′ -position of the side benzene, indicating the potential favorability of incorporating polar groups in this position for enhanced antiviral activity.Moreover, 47 and 48 are minimally cytotoxic, yielding SI values of 87.9 and 91.7, respectively.
The compounds 34 and 46 are characterized by a methyl group positioned on the benzylic methylene, linking the core ring to the side aromatic moiety.This linker group appears pivotal in the inhibitory activity, as evidenced by an up to sixfold reduction in antiviral efficacy (EC 50 = 6.3 µM for 34) compared to the structurally analogous compound 47 that also features a 4 ′ -COOCH 3 substitution.
We also investigated the impact of increasing the size of the side aromatic moiety by incorporating additional aromatic and heteroaromatic rings (the compounds 37, 38, 49, and 50).Remarkably, these compounds were active against HBV replication, with EC 50 values ranging from 2.3 to 4.7 µM.Moreover, three out of the four compounds with the bulkier side moiety exhibited no cytotoxic effects, thereby maintaining favorable SI values.
Analogs with a single substitution on the side chain (33-35, 39-42, 47-48) demonstrate greater potency compared to those with two or three substitutions (36-38, 43-45, 49, 50), regardless of the nature of the substitution.Notably, compounds containing halogens exhibited a significant increase in inhibitory activity, particularly when the halogen is positioned at the 4 ′ location of the aromatic side chain (33,34,47,48).This suggests that non-polar groups enhance potency and reduce cytotoxicity more effectively than polar groups, such as the hydroxyl group in the compound 39.This is likely due to the hydrophobic loop in the enzyme's active site, which necessitates a lipophilic moiety in the compound, thereby enhancing ligand-protein interactions.
We also tested the significance of the aromaticity by incorporating aliphatic side chains (40,41,42).The results showed the favorability of the four-carbon chain with an EC 50 1.2 µM for the compound 41 and good SI values.
Overall, HPDs featuring aromatic side chains substituted at the 4 ′ position with nonpolar groups were the most effective HBV RNase H inhibitors. Conversely, larger polar groups, particularly two or three aromatic rings, as well as the inclusion of a chiral linker, are less well tolerated for antiviral efficacy.Compounds that possess halogen substitutions on the aromatic side chain and a short linker (1 C after the oxime group) between the HPD core pharmacophore ring and the side aromatic moiety exhibit the optimal combination of antiviral activity, minimal cytotoxicity, and favorable SI values.
Our next effort focused on structurally modifying the primary pharmacophore by incorporating the minoxidil structure.Given the structural similarity between minoxidil and our HPD pharmacophore, we hypothesized that the three electron donors (two N atoms and one O atom) would form a "trident" configuration capable of chelating the Mg 2+ ions at the enzyme's catalytic site.Although computational studies (see Section 2.4) suggested potential activity, the compounds 55-58 were inactive against viral replication in vitro.Several hypotheses could explain these results, including the reduced electronegativity of nitrogen compared to oxygen (existent in the HPD heteroatom "trident"), the compounds' potential inability to penetrate cell membranes due to increased polarity, or the absence of E/Z isomerism, which might influence the molecule's conformation within the catalytic site.Further optimization could shed light on the potential antiviral activity of these analogs.
Despite their ability to chelate the Mg 2+ ions in the enzyme catalytic site in computational studies (see Section 2.4), the compounds 59 and 60, where the HPD ring was replaced with the pharmacophore ring of barbituric acid, lacked any inhibitory activity.This may be because the oxygen "trident" of the N-hydroxyimide group is essential for coordinating metal ions within the RNase H active site.These findings align with our previous observations for this class of compounds [30].Additionally, the increased polarity of these compounds likely reduces their inability to cross cell membranes to come into contact with the enzyme.To further address the impact of the different substitution patterns and residues in lipophilicity and druggability, we conducted a modeling calculation of druglike properties and descriptors using QikProp module of the Schrödinger platform (Table S1, Supporting Information).

Compound Solubility and Apparent Passive Permeability
In total, 21 out of the 24 compounds were highly soluble (solubility limit ≥ 100 µM) in conditions reflecting tissue culture media (pH 7.4).However, the barbituric analogs were insoluble, preventing meaningful biological evaluation (Table 2).In summary, the HPD oximes and minoxidil analogs exhibit promising druglike properties (Tables 1 and S1).Additionally, out of the 24 novel compounds tested in parallel artificial membrane assays, 22 demonstrated a high apparent passive permeability by the industry standard cutoff (>1 × 10 −6 cm/s).This includes all HPD oximes and minoxidil analogs.

Computational Molecular Docking
We performed induced fit docking (IFD) experiments to evaluate the binding pattern for all compounds into the active site of RNase H domain of HBV P. IFD was restricted to produce at least five binding poses for each compound.All of the compounds chelated two Mg 2+ ions via salt bridge interactions.It was observed in the case of the co-crystal structure of HIV RNase H with β-thujaplicinol that the hydroxyl trident of the inhibitor chelates Mg 2+ ions via eight salt bridge bonds.The core of the HPDs have two adjacent hydroxyl groups able to chelate both Mg 2+ ions via 6-7 salt bridge interactions, while minoxidil derivatives (55-58) containing only one hydroxyl group can only make 2 to 4 bonds (Figure 3A,B).The compounds 59 and 60 which are analogs to barbituric acid contain an alternative carbonyl on their main core and can chelate Mg 2+ ions via their deprotonated amino group through 1 to 2 salt bridge interactions (Figure 3C).The reduced number of interactions with Mg 2+ ions are also reflected in their poor docking scores −5.7 and −5.5 kcal/mol for 59 and 60, respectively (Table 1).We also observed that the compounds 56, 57, and 58 which form 4 or fewer salt bridge interactions have EC 50 values >100 µM as docking scores range from −8.2 to −8.17 kcal/mol, whereas compounds which contain at least two hydroxyl groups on their main core have docking scores ranging from −11.1 to −8.9 kcal/mol and have EC 50 values ranging from 1.1 to 7.7 µM.-) and 1 metal coordination interaction (--).In (B,C), the compounds 56 (yellow, pKa = 7.509) and 59 (pink) make 3 salt bridge (--) interactions with Mg 2+ ions.Right; ligand interaction diagram; Left, surface diagram.A PDB file containing for the HBV RNase H used for docking can be found in [18].Docking scores are in kCal/mol.
In our previous docking results, the R-groups of most of the compounds were solvent exposed and few of them interacted in the three binding pockets which were defined based on the R-group placement [30].In our current docking study, most of the compounds R-groups were located in pocket S3 and made interactions with residues S750, N749, and H726 in the pocket (Figure 4).It seems that a longer oxime linker helps these compounds to fit better into the S3 binding pocket to facilitate the formation of interactions with residues in the binding pocket.--) and 1 metal coordination interaction (--).In (B,C), the compounds 56 (yellow, pKa = 7.509) and 59 (pink) make 3 salt bridge (--) interactions with Mg 2+ ions.Right; ligand interaction diagram; Left, surface diagram.A PDB file containing for the HBV RNase H used for docking can be found in [18].Docking scores are in kCal/mol.
In our previous docking results, the R-groups of most of the compounds were solvent exposed and few of them interacted in the three binding pockets which were defined based on the R-group placement [30].In our current docking study, most of the compounds' R-groups were located in pocket S3 and made interactions with residues S750, N749, and H726 in the pocket (Figure 4).It seems that a longer oxime linker helps these compounds to fit better into the S3 binding pocket to facilitate the formation of interactions with residues in the binding pocket.A PDB file containing for the HBV RNase H used for docking can be found in [18].Docking scores are in kCal/mol.

Chemistry-General Part
All reagents and starting materials were purchased from commercial suppliers and used without further purification.Anhydrous CH2Cl2 was obtained by distillation from calcium hydride under argon.Anhydrous THF was freshly distilled from Na and benzophenone ketyl.All non-aqueous reactions were performed under an inert atmosphere of argon.Concentrated refers to the removal of solvent with a rotary evaporator at normal water aspirator pressure, followed by further evacuation on a high-vacuum line.Thinlayer chromatography was performed using silica gel 60 Å precoated aluminum or glass- A PDB file containing for the HBV RNase H used for docking can be found in [18].Docking scores are in kCal/mol.

Synthesis of O-Substituted N-Hydroxyphthalimides 1-14
General procedure: To a solution of N-hydroxyphthalimide (500.0 mg, 3.07 mmol, 1 equiv.) in anhydrous DMF (3 mL), NaH 60% w/w (1.25 equiv.) is added at 0 • C. The mixture is stirred at rt for 30 min.Thereafter, the appropriate halogenide (1.5 equiv.) is added and the reaction is stirred at rt overnight.Then, water is added and a solid precipitate is formed.The precipi-tate is filtered under vacuum and washed with water and a solution of n-pentane/Et 2 O 7:3.The solid is dried over P 2 O 5 , to afford the desired product.

Synthesis of O-Substituted Hydroxylamines 15-28
General procedure: To a solution of the appropriate N-hydroxyphthalimide (250.0 mg, 1 equiv.) in CH 2 Cl 2 (3 mL), hydrazine monohydrate 64% w/w (2 equiv.) is added and the reaction is stirred at rt for 1-24 h.The formed white precipitate is filtered, washed with CH 2 Cl 2 , and the filtrate is concentrated to afford the corresponding hydroxylamine.

Synthesis of Barbituric Acid Analogs 59-60
The compound 5-Acetyl-barbituric acid (1 equiv.) is suspended in abs.EtOH (5 mL) at ~90 • C and molecular sieves and the appropriate hydroxylamine (1.1 equiv.)are added.The mixture is stirred for 3-4 days at 70 • C under an argon atmosphere.After 4 days, the reaction mixture color remains unchanged, and the reagents do not seem to dissolve.Moreover, TLC does not provide a clear image of whether the reaction has progressed.The suspended solid is slowly filtered under vacuum and washed with Et

Cells and Cell Culture
HepDES19 cells were incubated on collagen-coated plates at 37 • C with 5% CO 2 and saturating humidity in Dulbecco's modified Eagle's medium (DMEM/F12) (Cytiva Life Sciences, Marlborough, MA, USA) supplemented with 10% fetal bovine serum (FBS), penicillin (100 IU/mL), and streptomycin (100 µg/mL).HepDES19 cells are HepG2 cells that carry a stably transfected HBV genotype D genome under the control of a tetracycline repressible promoter [50].Cells were maintained in the presence of 1 µg/mL tetracycline to repress the expression of the stably integrated HBV genome, and HBV replication was induced by the withdrawal of tetracycline from the culture medium.

qPCR HBV Replication Inhibition Assay
HBV replication inhibition was measured in HepDES19 cells induced to replicate HBV by the removal of tetracycline using a strand-preferential quantitative PCR assay as previously described [29,51].Cells were seeded at 4 × 10 4 cells/well in 96-well plates for 48 h.Serially diluted compound was then added to cells at a final concentration of 1% DMSO for 72 h, after which the cells were lysed and qPCR was performed as described in Li et al. [51].EC 50 values were calculated from the (+) DNA data with GraphPad Prism using the four-parameter log (inhibitor) versus response algorithm with the bottom value set to zero.Three or more replicate assays were performed on different days.

MTS Cytotoxicity Assay
Cell viability was assessed using the CellTiter 96 TM Aqueous Non-Radioactive Cell Proliferation assay (MTS) (Promega, Madison, WI, USA) as described previously [29].HepDES19 cells were seeded at 1 × 10 4 cells per well in 96-well, and compound was added after 48 h.Cells incubated with serially diluted compound (1% DMSO) for 72 h.Bac absorbance values were subtracted, and data were converted to percent cell viability.The cytotoxic concentration 50% (CC 50 ) values were calculated with GraphPad Prism by using the four-parameter variable response log (inhibitor) versus response algorithm with the bottom value set to zero.Three or more replicate assays were performed on different days.

Solubility Limit Assay
Compound solubility limits were tested in DMEM-F12 without phenol red (Gibco, Grand Island, NY, USA) supplemented with 10% FBS (pH 7.2-7.4) to mimic cell culture experiments, as we have performed previously [29].Compounds were serially diluted in buffer at pH 7.4 in 384-well, optically clear plates (upper limit 200 µM, DMSO 1%) and read on a plate reader at 620 nm.Compound concentration (µM) was plotted against optical density (OD) and a solubility limit was determined by identifying an inflection point where there was a significant increase in optical density due to increasing turbidity; the concentration of the compound at the inflection point was defined as the solubility limit.Two or more replicate assays were performed on different days for each compound.

Parallel Artificial Membrane Permeability Assay
Apparent passive permeability (P app ) was assessed at pH 7.4 using a 96-well donor/ acceptor cassette (Sigma Aldrich, Burlington, MA, USA), which mimics the apical and basolateral sides of the small intestine epithelium, as we have described previously [30].Briefly, an artificial membrane composed of 1% w/v lethicin/dodecane was added to the poly (vinylidene fluoride) (PVDF) membrane filter.Once dry, compounds (200 µM) were diluted in buffer at pH 7.4 and added to the donor side of the membrane, and the same pH buffer was added to the acceptor side.The cassette was assembled and incubated at room temperature with shaking at 250 rpm for 2 h, after which 100 µL of the acceptor well was retrieved and the experimental compound absorbance was read on a plate reader between 200 and 600 nm.Compound P app s (cm/s) values were determined by normalizing to the compound absorbance at equilibrium, incubation time, and membrane porosity Two or more replicate assays were conducted on different days.

Molecular Modeling
The full-length HBV P model used in docking studies was generated through Al-phafold2 and then prepared by placing Mg +2 ions into the active site of the RNase H domain by superposition of the DEDD active site motif of RNase H onto the cocrystal structure of HIV RNase H (PDB: 3K2P) [18].A PDB file for the HBV model used for docking can be found in [18].The Ligprep and Protein Preparation Wizard programs in the Schrödinger suite (Schrödinger LLC, New York, NY, USA) were used to prepare ligands and the protein, respectively, as mentioned in Giannakopoulou et al. [30].Thirty-six conformers of each ligand were generated by the Ligprep routine and used to seed the docking algorithm.The protonation states of ligands and the protein structures were given at pH 7.5 ± 2 and were energy minimized with an OPSL4 force field.The induced fit docking (IFD) program of the Schrödinger suite (Schrödinger LLC) that allows both the ligand and protein to change shape to minimize the energy during binding was used to analyze the binding poses of HPDs, as described previously [30].A receptor grid of 10 Å was used to dock the compounds into the active site of the HBV RH domain.Refinement of the protein-ligand complex was performed using a van der Waals radius scaling of 0.5.Residues were refined within 5 Å of the ligand poses of the top 20 structures after initial docking.Redocking was performed with the top structures within a 30 kcal/mol energy window using Glide XP precision (Schrödinger LLC, New York, NY, USA, Release 2023-3).

Conclusions
This work represents part of our ongoing project of creating safe and effective HBV RNase H inhibitors.In total, 18 new HPD oximes plus 4 structurally related minoxidil and 2 barbituric acid analogs were designed, synthesized, and evaluated for their ability to inhibit HBV replication.All of the HPD oximes were active against HBV while exhibiting minimal cytotoxicity.Additionally, they had high rates of apparent passive permeability and had high solubility limits at pH 7.4.The oxime group appears to be the better choice for linking the main HPD ring to the side aromatic segment because the HPD oximes have more promising profiles than corresponding HPD imines in both in vitro and in silico studies.The benzyl moiety with small lipophilic substituents (especially in the 4 ′ -position) appears to be the best of the tested side aromatic components, whereas adding extra aromatic rings or altering the linker group does not seem to improve the compound's performance.The minoxidil and barbituric acid analogs, while promising in computational studies, did not exhibit any antiviral activity.The results of this study provide deeper insights into the SARs of the HPD compound class and reinforce HBV RNase H as a promising drug target for combating HBV.These findings will also guide the further optimization of the HPD scaffold in our ongoing efforts to develop more effective and selective HBV RNase H inhibitors. Future research will aim to further optimize the side aromatic moiety of HPDs and explore the incorporation of alternative pharmacophore rings with metal-binding groups capable of chelating Mg 2+ ions in the enzyme's RNase H active site.

Figure 3 .
Figure 3. HPDs chelating Mg 2+ ions.(A) The compound 39 (green, pKa = 9.609) docked into the active site of HBV P RNase H domain chelating both Mg 2+ ions.The ligand interaction diagram on the right shows two deprotonated hydroxyl groups on the HPD core making 7 salt bridge bonds (--) and 1 metal coordination interaction (--).In (B,C), the compounds 56 (yellow, pKa = 7.509) and 59 (pink) make 3 salt bridge (--) interactions with Mg 2+ ions.Right; ligand interaction diagram; Left, surface diagram.A PDB file containing for the HBV RNase H used for docking can be found in[18].Docking scores are in kCal/mol.

Figure 3 .
Figure 3. HPDs chelating Mg 2+ ions.(A) The compound 39 (green, pKa = 9.609) docked into the active site of HBV P RNase H domain chelating both Mg 2+ ions.The ligand interaction diagram on the right shows two deprotonated hydroxyl groups on the HPD core making 7 salt bridge bonds (--) and 1 metal coordination interaction (--).In (B,C), the compounds 56 (yellow, pKa = 7.509) and 59 (pink) make 3 salt bridge (--) interactions with Mg 2+ ions.Right; ligand interaction diagram; Left, surface diagram.A PDB file containing for the HBV RNase H used for docking can be found in[18].Docking scores are in kCal/mol.

Figure 4 .
Figure 4. Docking studies of HPDs into the active site of RNase H domain of HBV P. (A) The Rgroup of the compound 49 (white) docked into the active site of the HBV P RNase H domain makes a pi-pi cation interaction with H726 in the S3 binding pocket.(B) The compound 36 (green) R-group makes an h-bond with S750, whereas (C) 45 (cyan) makes an h-bond with N749 in the S3 binding pocket.Right panel, ligand interaction map; Left panel, surface diagram of docked compound into the active site of HBV P RNase H domain.A PDB file containing for the HBV RNase H used for docking can be found in[18].Docking scores are in kCal/mol.

Figure 4 .
Figure 4. Docking studies of HPDs into the active site of RNase H domain of HBV P. (A) The R-group of the compound 49 (white) docked into the active site of the HBV P RNase H domain makes a pi-pi cation interaction with H726 in the S3 binding pocket.(B) The compound 36 (green) R-group makes an h-bond with S750, whereas (C) 45 (cyan) makes an h-bond with N749 in the S3 binding pocket.Right panel, ligand interaction map; Left panel, surface diagram of docked compound into the active site of HBV P RNase H domain.A PDB file containing for the HBV RNase H used for docking can be found in[18].Docking scores are in kCal/mol.

Table 2 .
Compound solubility limits and apparent passive permeability at pH 7.4.

Table 2 .
Compound solubility limits and apparent passive permeability at pH 7.4.

Table 2 .
Compound solubility limits and apparent passive permeability at pH 7.4.

Table 2 .
Compound solubility limits and apparent passive permeability at pH 7.4.

Table 2 .
Compound solubility limits and apparent passive permeability at pH 7.4.