Introduction

Hepatitis B virus (HBV) infection is a global health problem because infected individuals are at risk of liver cirrhosis and hepatocellular carcinoma (HCC). Despite the availability of effective vaccines, it is estimated that 257.5 million people were infected with hepatitis B virus (HBV) chronically1. Nucleos(t)ide analogs (NAs) have been widely used in the treatment of chronic hepatitis B (CHB), and the current guidelines for CHB recommend entecavir (ETV), tenofovir disoproxil fumarate (TDF), and tenofovir alafenamide fumarate (TAF) as the first-line drugs for NAs for CHB2. These drugs inhibit reverse transcriptase and are highly effective in normalizing alanine aminotransferase (ALT) and suppressing HBV DNA. However, the long-term goal of eliminating hepatitis B surface antigen (HBsAg), functional cure, is achieved in only 3–5% in 10 years3, and the probability of achievement of functional cure is extremely low. These therapies are inadequately effective, and novel mechanisms of action are being developed to cure HBV infection.

Mature infectious HBV particles, also called Dane particles, are 42 nm diameter particles composed of the envelope and the nucleocapsid. The envelope consists of a cellular lipid bilayer and large/middle/small hepatitis B surface (LHBs/MHBs/SHBs) proteins. Other HBV-related particles, such as HBsAg subviral particles (SVPs), which do not contain nucleocapsids, and particles containing HBV RNA are also released from hepatocytes in large numbers. They are considered to act as decoys from the immune system4. LHBs/MHBs/SHBs proteins in these particles are detected as HBsAg and are widely assayed as a marker of HBV infection. The nucleocapsid consists of hepatitis B core (HBc) proteins, which form dimers and encapsulate HBV DNA and polymerase.

The interaction between the preS1/S2 region on the N-terminal side of LHBs protein and HBc protein is thought to be necessary for the envelope formation process of HBV5. In this study, we hypothesized that this interaction could be a new target for anti-HBV therapy, and conducted a high throughput screening to identify compounds to inhibit this interaction.

Results

Establishment of high throughput screening (HTS) method to evaluate the interaction between LHBs and HBc

At first, we determined the condition of the NanoBRET Assay (Promega, Madison, WI). A principle of the assay is illustrated in Fig. 1A. The results of the combination of NanoLuc donor and HaloTag acceptor in the NanoBRET Assay, which evaluates the interaction of LHBs protein and HBc protein, are shown in Fig. 1B. The combination with the highest bioluminescence resonance energy transfer (BRET) ratio was the addition of NanoLuc to the N-terminal side of the LHBs protein and HaloTag to the N-terminal side of the HBc protein, when the assay was performed with NanoLuc to HaloTag plasmid DNA ratio of 1:10. To optimize the ratio, NanoLuc to HaloTag ratios of 1:1, 1:10, 1:100 and 1:1000 in the above combinations were tested (Fig. 1C). The BRET ratios were generally similar at each concentration. Because the acceptor signal seemed to be saturated at the plasmid DNA ratio of 1:1, ratio of 1:10 was considered to be ideal and was used for the primary screening. When the BRET ratio was measured using an HBc mutant L60A as a control, it was significantly lower than that of wild-type HBc (Figure S1).

Fig. 1
figure 1

Establishment of NanoBRET Assay. (A) Schematic illustration of the principle of the NanoBRET Assay. NL, NaloLuc; HT, HaloTag; HL, HaloTag NanoBRET ligand. (B) Result of BRET ratio to evaluate the strength of interaction of large hepatitis B surface (LHBs) and hepatitis B core (HBc) proteins. NL and HT were tagged with the target proteins at N-terminus or C-terminus, and 8 different combinations was evaluated (n = 4). Protein-tag combinations are shown in N-terminal-C-terminal order. (C) Result of optimization of NL donor/HT acceptor-expressing plasmid DNA ratio (n = 4). BRET ratios were measured with plasmid DNA ratios of 1:1, 1:10, 1:100, and 1:1000 for expressing NL-LHBs and HT-HBc.

Full size image

Primary screening

Using the HTS method established above, a primary screening was conducted with 3,200 compounds. Forty hit compounds were obtained that met the conditions of BRET ratio ≤ 0.6 and cell viability ratio ≥ 0.8 (Fig. 2). Z-scores for these compounds are shown in Table S2.

Fig. 2
figure 2

Results of the primary screening using 3,200 compounds (10 µM). Hit compounds that were defined as a BRET ratio ≤ 0.6 and a cell viability ratio ≥ 0.8 are indicated by a colored area.

Full size image

Secondary screening

Secondary screening with 40 hit compounds was performed using HepG2.2.15.7 cells. 10 compounds significantly reduced the amount of Dane particles, evaluated as the amount of HBV DNA after anti-HBs immunoprecipitation, with a cell viability ratio greater than 0.8 (Fig. 3A). When the negative control was considered as 100%, hit compound 2 reduced HBV DNA to 78 ± 9%, compound 9 to 59 ± 9%, compound 16 to 61 ± 14%, compound 17 to 70 ± 10%, compound 20 to 75 ± 10%, compound 24 to 65 ± 17%, compound 28 to 62 ± 5%, compound 36 to 62 ± 25%, compound 38 to 63 ± 14%, and compound 39 to 61 ± 12% (all p < 0.05). Hit compound 28 is a reactant used in peptide synthesis and reacts with an amino group (-NH2) to form an amide bond (peptide bond). This compound was excluded because it reacts with the lysine residue of the protein, resulting in nonspecific binding. The top 5 compounds with the lowest mean values in HBV DNA and with the lowest standard deviation were used in the further analyses as candidates. No hit compound significantly reduced HBsAg without decreasing cell viability (Fig. 3B). We similarly measured HBV DNA by PCR using the pre-immunoprecipitation supernatant and found that none of the compounds significantly reduced HBV DNA while maintaining viability (Figure S2).

Fig. 3
figure 3

Results of the secondary screening that was performed three times. (A) HBV-expressing HepG2.2.15.7 cells were treated with 40 compounds (10 µM) and the amount of enveloped Dane particles in the culture supernatant was quantified (n = 3). It was evaluated with a relative change of the HBV DNA level measured by real-time PCR after immunoprecipitation with anti-HBs. The green dots indicate the mean of cell viability. (B) Results of CLIA assay for HBsAg in the same culture supernatant as in A are shown (n = 3). * p < 0.05.

Full size image

Dose dependence of anti-HBV activity

The dose-dependent anti-HBV effect of the 5 hit compounds obtained in the secondary screening was examined. The IC50 values were 0.9 µM for hit compound 16, 10.7 µM for hit compound 17, 1.0 µM for hit compound 38, and 18.6 µM for hit compound 39 (Fig. 4). The IC50 could not be calculated for hit compound 9. The CC50 values were 8.0 × 105 µM for hit compound 9, 43.4 µM for hit compound 16, 20.3 µM for hit compound 17, and 131 µM for hit compound 39. That of hit compound 38 was not calculated due to the weak cytotoxicity even at 100 µM. As a control, ETV and a capsid assembly modulator NVR 3-778 was used (Figure S3). Based on these results, hit compound 16 of 10 µM and hit compounds 38 and 39 of 30 µM were used in the subsequent analysis. Hit compounds 9 and 17 was excluded from candidates because of no concentration-dependent effect or closeness of IC50 and CC50. Hit compound 38 (Library ID: 6472) was a compound named α-isocupreine whose clinical action is unknown (Fig. 5B). Hit compound 39 (Library ID: 6474) was quinidine, which is clinically used as an antiarrhythmic agent (Fig. 5C). Interestingly, these compounds were structurally similar. Hit compound 16 (Library ID: 6625) was a novel compound that has not been reported so far (Fig. 5A).

Fig. 4
figure 4

Verification of dose dependent effects on HBV with hit compounds. HepG2.2.15.7 cells were treated with the compounds at doses of 0.1 µM, 0.3 µM, 1 µM, 3 µM, 10 µM, 30 µM, and 100 µM. The levels of HBV DNA after the immunoprecipitation are shown on the left and the levels of cell viability are shown on the right (n = 3). Based on these results, 50% inhibitory concentration (IC50) and 50% cytotoxic concentration (CC50) were calculated. Results of hit compound 9 (A), hit compound 16 (B), hit compound 17 (C), hit compound 38 (D), and hit compound 39 (E) are shown. SI stands for selectivity index, calculated by CC50/IC50.

Full size image
Fig. 5
figure 5

The structural formulas of the candidate compounds after secondary screening. Structures of hit compound 16 (A), hit compound 38 (B), hit compound 39 (C) are shown.

Full size image

Effects of candidate compounds on intracellular HBV protein and RNA

Results of western blotting analysis using HepG2.2.15.7 cells with hit compounds 16, 38, and 39 are shown in Fig. 6A, B and Figure S4. There were no significant differences of LHBs protein and HBc protein after the addition of hit compounds 16, 38, and 39.

Figure 6C shows the results of HBV-related RNA quantification when the above compounds were added to HepG2.2.15.7 cells. Hit compounds 38 and 39 showed no significant difference in the levels of precore mRNA and pgRNA compared to a control. Hit compound 16 decreased precore mRNA (0.84-fold, p = 0.04) and tended to increase pgRNA (1.47-fold, p = 0.27).

Fig. 6 
figure 6

Effects of the 3 hit compounds on HBV proteins and RNAs. (A) Representative results of western blot analysis for LHBs, HBc, and actin using HepG2.2.15.7 cells are shown. Cells were treated with hit compound 16 of 10 µM, hit compound 38 of 30 µM, and hit compound 39 of 30 µM. (B) Comparison of the levels of LHBs and HBc in western blot analysis quantified by densitometry (n = 3). (C) Comparison of the levels of precore mRNA or precore mRNA plus pregenomic RNA (pgRNA) in HepG2.2.15.7 cells with the 3 candidate compounds (n = 3). * p < 0.05.

Full size image

Change of virus particle density after addition of the compound

The culture supernatants from HepG2.2.15.7 cells with the addition of hit compound 16 were fractionated by sucrose density gradient centrifugation to compare the sedimentation velocity of particles. The negative control showed a bimodal graph of HBV DNA with a peak at the lower density fraction (1.21 g/mL) corresponding to Dane particles and another peak at the higher density fraction (1.24 g/mL) corresponding to the naked nucleocapsids (Fig. 7A), as reported previously6. In contrast, compound 16 showed a single peak of HBV DNA at the higher density fraction of 1.24 g/mL, suggesting that this compound inhibits the envelope formation of Dane particle (Fig. 7B). Because Southern blot could not be performed due to the low concentration of HBV DNA in each fraction, to deny only HBV DNA fragment was amplified, we confirmed that the similar results were obtained by real-time PCR using different primers targeting the C gene (Fig. 7C and D). The peaks of subviral particles, which were detected as HBs, in both conditions were at 1.14–1.16 g/mL. Detection of HBc by dot blot with our polyclonal antibody did not work well.

Fig. 7
figure 7

Effects of hit compound 16 on HBV particles and the HBs protein. (A, B) HepG2.2.15.7 cells were treated with DMSO (A) or hit compound 16 (B) and the culture supernatants were fractionated with sucrose density gradient centrifugation. HBV DNA in each fraction was determined by real-time PCR targeting the S gene and HBs was determined by dot blot. (C, D) The culture supernatants from cells treated with either DMSO (C) or hit compound 16 (D), were fractionated as above, and were analyzed by real-time PCR using a different set of primers targeting the C gene to quantify HBV DNA.

Full size image

Effects of the compound in an HBV infection model

Using HepG2-hNTCP-C4 cells with HBV infection, HBV suppression was tested by adding hit compound 16, that showed the lowest IC50 in HepG2.2.15.7 cells. The IC50 and CC50 of compound 16 in NTCP cells were 0.011 µM and 110 µM, respectively (Fig. 8A and B). Compound 16 showed antiviral effects at lower concentrations in the infection model than in HepG2.2.15.7 cells. These results suggest that the new compound, hit compound 16, may be more effective in HBV-infected cells than in HBV-expressing cells.

Fig. 8
figure 8

Effects of the hit compound on HBV particles in an HBV infection model. HepG2-hNTCP-C4 cells were infected with HBV and treated with hit compound 16. The IC50 (A) and CC50 (B) are calculated from the dose dependent effects on HBV (n = 3). * p < 0.05.

Full size image

Discussion

NAs and pegylated interferon (Peg-IFN) are currently used for the treatment of CHB. However, the achievement rate of HBsAg clearance, the long-term goal of treatment, is very low. Such current conditions have prompted the development of therapeutic agents with new mechanisms of action against CHB. Beprovirsen, an antisense oligonucleotide that targets HBV mRNA, was reported to be effective in lowering HBsAg in a phase 2 clinical trial7. In addition, Bersacapavir, a core protein allosteric modulator (CpAM) that binds directly to the HBc protein and promotes abnormal nucleocapsid formation, showed efficacy in combination with NAs in a phase 2 clinical trial8. While these are the direct acting antiviral agents to the virus, therapeutic agents targeting host immunity are also being developed. NASVAC, a therapeutic vaccine administered intranasally with a mixture of HBs and HBc antigens, was shown to be more effective in suppressing HBV DNA 24 weeks after completion of treatment in a phase 3 clinical trial in Bangladesh, compared to IFN9. Furthermore, after 18 months of follow-up, 26.9% of patients showed a reduction in HBsAg. Thus, although various clinical trials currently undergo, there is still no drug that can achieve a high rate of HBsAg loss.

The interaction between LHBs protein and HBc protein, on which we focused in this study, has been shown to promote accurate formation and stabilization of viral particles10 and may be a target for a novel HBV therapy. However, there is only one report of a compound that inhibits this interaction5and such a compound has not yet been fully investigated. In that paper, the antigen-antibody reaction is used to measure the interaction between LHBs and HBc proteins, but we used the NanoBRET system to evaluate the interaction. The NanoBRET system is a method for measuring protein-protein interactions based on the principle of bioluminescence resonance energy transfer, BRET, in which energy is transferred when a bioluminescent donor and a fluorescent acceptor are in close proximity. The advantage of this system is that HTS can be performed under conditions close to those of living organisms, since protein binding in the cell can be evaluated. We established this HTS method using wild-type LHBs and HBc. A mutant form of HBc, L60A, has been reported to inhibit envelope formation11,12, and the intensity of the signal in NanoBRET was reduced when this mutant was used. However, the difference was modest and it was considered that the tag proteins might affect the effect of the mutation. Therefore, we consider this assay to be a suitable method to evaluate the compound’s inhibition of envelope formation. HaloTag and NanoLuc do not interact directly but their proximity results in the robust signal. Therefore, the hit compounds are considered not to affect the signal from these tags. So far, some studies had used the NanoBRET Assay for the drug screening without a concern for such direct interaction of these tags13,14.

Using this HTS method, a primary screening was conducted and 40 compounds met the criteria for hit compounds. After secondary screening and following evaluation of dose-dependent antiviral effect, 3 candidate compounds were obtained. In the culture supernatants from HBV-expressing cells with the addition of these compounds, the level of mature virus particles was reduced. These compounds might inhibit the interaction between LHBs protein and HBc protein, resulting in the disturbance of HBV envelope formation, and the secretion of Dane particles to the extracellular space might be reduced. The decrease of HBV DNA in fractions with density corresponding to enveloped HBV particles after the addition of compound 16 (Fig. 7B) supports this mechanism. However, it should be noted that some levels of HBV DNA were detected in the fractions that contain subviral particles. In general, subviral particles do not contain HBV DNA but a small part of Dane particles might be contained in the subviral particle fractions during the fractionation performed manually. Actually, fraction 6 is on a border of subviral particles and Dane particles and might contain both particles. On the other hand, none of the compounds reduced HBV DNA in the supernatant before immunoprecipitation. We hypothesized that this was because, although the release of Dane particles decreased due to inhibition of envelope formation, an increased number of exposed nucleocapsids were released into the supernatant, leaving the total amount of DNA unchanged. Because the exposed nucleocapsids are not infectious, their increase is considered to be harmless. As for HBsAg, it is weakly detected in fractions 10–13, the naked capsid fraction in Fig. 7B, which may be due to incompletely enveloped HBV particles due to the compound. The amounts of intracellular viral proteins were not changed (Fig. 5B): this result might be due to the excess of viral proteins that are not utilized for the viral particle formation. Interestingly, precore mRNA was slightly decreased and pgRNA was slightly increased after the addition of compound 16. Such a change was observed in a mutant strain of the core promoter region often found in chronic hepatitis B patients15,16,17suggesting that compound 16 might have some effects on the core promoter region. The increase of pgRNA might potentially increase HBc level but such a result was not obtained (Fig. 6). The mechanisms have not been investigated, but it is assumed that hit compound 16 might affect the post-transcriptional regulation of pgRNA or mRNA stability. Otherwise, the inhibition of envelope formation might enhance the release or degradation of naked nucleocapsid to some extent, resulting in the HBc level in the cells remaining unchanged. Further experiments are needed to clarify this phenomenon. The IC50 of the compound 16 in HBV-infected model using HepG2-hNTCP-C4 cells was lower than in HBV-expressing cells (Fig. 8A). This might be due to the lower HBV level in the HBV-infected cells, and in addition, the possible difference of ratio of LHB and HBc might affect.

Compound 16 is an artificially altered structure of a naturally occurring compound, and is not currently in clinical use. On the other hand, compound 38 is a derivative of quinine, a plant-derived alkaloid, named α-isocupreine, and is commercially available as a research reagent, but its clinical effects are unknown. Compound 39 is quinidine, a stereoisomer of quinine, and is an antiarrhythmic drug. These compounds had no structure in common with the hit compounds shown in the previous report5, and we considered that they inhibited the interaction between core and surface proteins by different mechanisms of action. Hit compounds 38 and 39 both have a quinoline skeleton, a heterocyclic compound in which a benzene ring and a pyridine ring are fused together. The quinoline skeleton is a structure commonly found in many antibacterial and antiprotozoal drugs, and these compounds are effective against Plasmodium falciparum and other microorganisms. This action is thought to be due to the quinoline skeleton inhibiting physiological processes in microorganisms, and it may be possible that a similar mechanism is responsible for their antiviral effects against HBV. A further study is needed to clarify the detailed mechanisms in the future. For the clinical application of these compounds, it is necessary to clarify the compound-protein interaction. Once these are determined, more detailed estimation of the mechanism of action will be possible, and consideration of concomitant use with compounds having other mechanisms of action may be performed more efficiently. It is also known that quinidine and quinine have many side effects, and it is necessary to seek compounds with fewer side effects and stronger HBV inhibitory effects through structural development such as molecular conversion and modification for clinical application. The effective blood concentration of quinidine for arrhythmia is 2.3–5.0 µg/mL and the IC50 against HBV calculated in this study was 2.8 µg/mL. Therefore, this concentration is considered that to be achievable by oral administration. In addition to experiments in vitro, it is necessary to investigate the antiviral effect, side effects, and pharmacokinetics in animal models such as chimeric mice with humanized liver in the future.

HBV is thought to maintain persistent infection by repeating the cycle of being released from hepatocytes into the blood and then infecting other hepatocytes. Inhibiting envelope formation reduces mature viral release and may control infection by suppressing this cycle. Nucleic-acid polymers (NAP) have been reported as agents that inhibit the release of HBsAg18,19. Because HBsAg is thought to attenuate the immune response to HBV, NAP has been suggested to improve the immune response, and its usefulness in combination with Peg-IFN has been reported20,21,22. Although the compounds identified in this study did not reduce HBsAg in the present study in cultured cells, they may reduce HBsAg if administered long-term by inhibiting the spread of infected hepatocytes. It is also possible that the addition of drugs with immunostimulatory effects, such as Peg-IFN, may be effective in the direction of HBV elimination if infected hepatocytes are decreased with the candidate compounds. Because the candidates in this study have a different mechanism of HBV inhibition than other drugs in clinical use or in development, and because combination with antiviral agents with other mechanisms of action may also be useful23 it is required to investigate the combination with NAs and other anti-HBV drugs in clinical trials such as CpAM.

The first limitation of this study is that the number of compounds screened was small and the hit compounds are relatively weak. To identify more potent compounds, it is necessary to test further diverse compounds and to try structural developments from the findings in this study. Second, we could not confirm whether HT-HBc is assembled into the capsids and whether NL-LHBs is incorporated into the envelope. We considered that compounds inhibiting the interaction between LHBs and HBc, regardless of the formation of envelope and capsid, might potentially inhibit the interaction of assembled capsid and envelope. Third, this study employed a high-throughput fluorescence screening approach to identify compounds that inhibit the interaction between the LHBs and HBc proteins. It should be noted that HTS methods, although efficient for initial compound selection, may have limitations in specificity and accuracy. In this study, the inhibitory interaction was inferred indirectly through cellular and virological assays. However, direct biophysical validation methods such as biolayer interferometry, surface plasmon resonance or isothermal titration calorimetry were not conducted. Therefore, we did not evaluate which proteins, LHBs or HBc, the hit compounds bind to. Further molecular-level validation is needed to conclusively confirm the direct binding inhibition between LHBs and HBc proteins. Fourth, we have not performed more detailed molecular analysis of HBV replication intermediates. Finally, we have tried immunoprecipitation to show the inhibition of the interaction between LHBs and HBc using HepG2.2.15.7 cells and the compound 16, but we did not obtain conclusive data, probably due to the difference in condition between the cytosol and the experimental buffer.

In conclusion, we established an HTS method to screen for compounds that inhibit the interaction of LHBs protein and HBc protein, and identified compounds that reduce mature HBV particles in cultured cells. It is possible that the compound inhibits envelope formation, and further structural developments and in vivo studies are needed for clinical application.

Methods

Compounds and reagents

A total of 3,200 compounds from a compound library that were provided by Dr. Takayuki Doi (Tohoku University Graduate School of Pharmaceutical Sciences) were used for the screening. Compounds were dissolved in dimethyl sulfoxide (DMSO) and stored at −20 °C at a concentration of 1 mM. An HBV capsid assembly modulator NVR 3-778 was purchased from MedChemExpress (Monmouth Junction, NJ). A polyclonal anti-HBs antibody (horse) was a kind gift from the Institute of Immunology (Tokyo, Japan). A polyclonal anti-HBc antibody (rabbit) and polyclonal β-actin antibody (rabbit) were purchased from Dako (Carpinteria, CA) and Cell Signaling Technology (Danvers, MA), respectively.

Cell culture

A hepatoma cell line HepG2 was obtained from Cell Resource Center for Biomedical Research, Institute of Development, Aging and Cancer, Tohoku University. HepG2.2.15.7 cells, which were derived from HepG2.2.15 cells and stably express HBV at a high level, were kindly provided by Dr. Koichi Watashi at the National Institute of Infectious Diseases (Tokyo, Japan). HepG2-hNTCP-C4 cells expressing the receptor for HBV24 were also kindly provided by Dr. Koichi Watashi.

NanoBRET assay

An interaction of LHBs and HBc was evaluated using the NanoBRET Assay (Promega). In the assay, NanoLuc, a small luciferase protein, and HaloTag, an acceptor protein, are tagged to the target proteins. When the target proteins bind to each other, the NanoLuc donor and HaloTag acceptor come into close proximity and produce a fluorescent signal by BRET. The BRET ratio is calculated as a signal intensity from acceptor divided by a signal intensity from donor, and the ratio reflects the strength of binding of target proteins (Fig. 1A). There are 8 possible combinations of NanoLuc/HaloTag and target proteins with different tagged positions, N-terminus or C-terminus. DNA sequences of the target proteins, LHBs and HBc, were amplified using total DNA from culture supernatant of HepG2.2.15 cells and specific primers (Supplementary Table 1). The expression plasmid DNAs were constructed using Flexi Vector Systems (Promega). Additionally, a mutant HBc sequence was amplified using a L60A mutant construct (kindly gifted from Dr. Hugues de Rocquigny, Inserm, France), which was reported to have reduced binding to LHBs25and used as a control.

Plasmid DNA was transfected into HepG2 cells, an HCC-derived cell line, for the above 8 combinations and assayed using the GloMax Discover Microplate Reader (Promega) according to the NanoBRET Assay protocol. The combination with the strongest fluorescent signal was identified and the NanoLuc to HaloTag plasmid DNA ratio was optimized to minimize unbound donors. TransIT-LT1 (Mirus Bio LLC, Madison, WI) and Opti-MEM I Reduced Serum Medium, no phenol red (Thermo Fisher Scientific) were used for transfection medium. The signal was measured with GloMax Discover Microplate Reader (Promega) and the acceptor/donor signal (BRET ratio) was calculated to evaluate protein-protein interactions.

Primary screening

Compounds that inhibit the interaction of LHBs protein and HBc protein in cells were tested using NanoBRET Assay whose condition was determined as described above. HepG2 cells that were transiently transfected with plasmids expressing LHBs/HBc with tags were seeded on 96-well plates at 2 × 104 cells/well and incubated for overnight. Then, each compound was added to each well at a final concentration of 10 µM and incubated for overnight. Finally, fluorescence signals were measured according to the NanoBRET Assay protocol. Cell viability was then measured using CellTiter-Glo Luminescent Cell Viability Assay (Promega) and a cell viability ratio to a negative control was calculated. A compound with a BRET ratio ≤ 0.6 and a cell viability ratio ≥ 0.8 was considered as a hit compound. The threshold for a BRET ratio was set to select the top 1% of the 3200 compounds.

Secondary screening

HepG2.2.15.7 cells were seeded on a 24-well plate at 1.5 × 105 cells/well and incubated for overnight. After washing with 500 µl of PBS, 500 µl of fresh medium with a compound at a final concentration 10 µM was added and incubated for 48 h. The culture supernatant was collected to determine the levels of Dane particles and HBsAg.

Assay of Dane particles and HBsAg

To measure Dane particles with the envelope in the culture supernatant, HBV DNA was quantified by real-time PCR using primers and a probe targeting the S gene26 (Supplementary Table 1) after immunoprecipitation with anti-HBs antibody according to a previous report6. HBsAg in culture supernatants was measured by chemiluminescence immunoassay (CLIA) using the ARCHITECT HBsAg reagent kit (Abbott Laboratories, Chicago, IL).

Western blot analysis

Cells were lysed in Radioimmunoprecipitation assay (RIPA) buffer and soluble protein was resolved by SDS-PAGE and transferred onto a polyvinylidene difluoride (PVDF) membrane. After blocking with ECL Prime Blocking Reagent (Cytiva, Tokyo, Japan), the membrane was incubated with primary antibodies at 4 °C for overnight and incubated with secondary antibodies at room temperature for 1 h. Signals were obtained with ECL Prime (Cytiva).

Measurement of intracellular RNA

Intracellular total RNA extracted with TRIzol reagent (Thermo Fisher Scientific) was reverse transcribed using SuperScript IV VILO Master Mix with ezDNase Enzyme (Thermo Fisher Scientific), and the levels of HBV pregenomic RNA (pgRNA) and precore mRNA were measured by real-time PCR using the SYBR Green method according to previous reports6. Primers PCP and BC1 were used for the detection of precore mRNA, and PGP and BC1 were used to detect both precore mRNA and pgRNA. GAPDH-F and GAPDH-R were used to detect glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as an internal control. Primer sequences are shown in Table S1.

Sucrose density gradient centrifugation

To determine the buoyant density of HBV-associated particles, culture supernatants were fractionated by sucrose density gradient centrifugation as previously reported6. Total DNA was extracted from each fraction and the level of HBV DNA was measured by real-time PCR using primers (HBV-S-Forward, HBV-S-Reverse) and a probe targeting the S gene. To confirm the results, real-time PCR using the SYBR Green method with primers (HBVF and HBVR) targeting the C gene27 was performed. Each fraction was subjected to dot blot using anti-HBs antibody as described previously28.

HBV infection model in vitro

HepG2-hNTCP-C4 cells were infected with HBV as previously reported29. Cells were seeded on 96-well plates at 2.0 × 104 cells/well and incubated overnight. Then, after washing, a mixture of 6.0 × 109 copies/well of genotype D HBV from HepAD38 cells, 8 µl of 40% PEG8000 and 72 µl of fresh medium was added. Plates were centrifuged at 1,000 g for 30 min to increase the infection efficiency30. After 16 h, cells were washed and fresh medium was added. Additional medium exchange was performed on days 5 and 9. On day 9, the compounds were added at their optimal concentrations, and on day 11, the supernatant was collected and the level of Dane particle was measured as above.

Statistical analysis

The quantitative data were presented as mean ± standard deviation. For statistical comparisons, the Kruskal-Wallis test was used to compare each combination in determining the HTS method, and the Dunn test was used for post hoc analysis. Other comparisons between the two groups were made with the Mann-Whitney U test. A p-value less than 0.05 was considered to be significant.