A Small Molecule, ACAi-028, with Anti-HIV-1 Activity Targets a Novel Hydrophobic Pocket on HIV-1 Capsid

ABSTRACT

The human immunodeficiency virus type 1 (HIV-1) capsid (CA) is an essential viral component of HIV-1 infection and an attractive therapeutic target for antivirals. Here, we report that a small molecule, ACAi-028, inhibits HIV-1 replication by targeting a hydrophobic pocket in the N-terminal domain of CA (CA-NTD). ACAi-028 is 1 of more than 40 candidate anti-HIV-1 compounds identified by in silico screening and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assays. Our binding model showed that ACAi-028 interacts with the Q13, S16, and T19 amino acid residues, via hydrogen bonds, in the targeting pocket of CA-NTD. Using recombinant fusion methods, TZM-bl, time-of-addition, and colorimetric reverse transcriptase (RT) assays, the compound was found to exert anti-HIV-1 activity in the early stage between reverse transcription and proviral DNA integration, without any effect on RT activity in vitro, suggesting that this compound may affect HIV-1 core disassembly (uncoating) as well as a CA inhibitor, PF74. Moreover, electrospray ionization mass spectrometry (ESI-MS) also showed that the compound binds directly and noncovalently to the CA monomer. CA multimerization and thermal stability assays showed that ACAi-028 decreased CA multimerization and thermal stability via S16 or T19 residues. These results indicate that ACAi-028 is a new CA inhibitor by binding to the novel hydrophobic pocket in CA-NTD. This study demonstrates that a compound, ACAi-028, targeting the hydrophobic pocket should be a promising anti-HIV-1 inhibitor.

INTRODUCTION

Human immunodeficiency virus type 1 (HIV-1) is a retrovirus that has affected humans for half a century, causing 700,000 deaths and 1.7 million new infections in 2019, according to the Joint United Nations Programme on HIV and AIDS (https://www.unaids.org/en). Significant progress has been made in recent years to understand HIV infection and to design drugs for counteracting multiple stages of the viral life cycle. Antiretroviral therapy (ART) has largely allowed HIV-infected patients to have the same life expectancy as uninfected individuals (1, 2). Despite of the benefits of ART, HIV-1 often acquires drug-resistant mutations, (3) resulting in treatment failure. High adherence to ART is required to sustain viral suppression during HIV clinical treatment (4). It is especially important that we continue to discover new anti-HIV-1 agents that have potent antiviral activity with different mechanisms, providing HIV patients with more options for ART treatment.

The capsid (CA) plays an essential role in both the early and late stages of the HIV-1 life cycle. One of the key periods in the early stage is HIV-1 capsid core disassembly, also known as uncoating (5, 6). After HIV-1 infection into the host cell via CD4 and CXCR4/CCR5 molecules, uncoating occurs in a regulated manner. The exact mechanism of uncoating remains unclear. Previous studies have suggested several hypotheses regarding the time and location of uncoating (7). Electron microscopy results suggested that uncoating occurs near the surface of the plasma membrane (8), while other studies have suggested that uncoating occurs in the cytoplasm or tethered to nuclear import (9). However, recent studies have indicated the possibility that uncoating actually occurs in the nucleus and might be coupled with both reverse transcription and integration (10–12). A new inhibitor of CA uncoating has the potential to provide novel insights and tools for further research in this field.

Representative anti-CA compounds, such as CAP-1 (13, 14), PF-3450074 (PF74) (15, 16), ebselen (17), BD-1, BM-1 (18), I-XW-053 (19), and C1 (20) have been discovered, although none have been approved for clinical use by the FDA. In this study, we used PF74 and ebselen as control drugs. PF74 is a lead compound of GS-CA1 (21) and lenacapavir, formerly known as GS-6207, (22, 23) that have been developed by Gilead Sciences, Inc. These compounds interact with the C-terminal domain-N-terminal domain (CTD-NTD) interface and hinges between CA monomers, inducing hyperstabilization of the HIV-1 core and inhibiting the interaction of cofactors, such as Nup153 and CPSF6 (16). Lenacapavir has advanced to phase 2/3 of the clinical CAPELLA trial (ClinicalTrials.gov identifier, NCT04150068), and preliminary reports from this trial indicate that it is effective at reducing the viral load of patients on failing treatment regimens, strongly suggesting that the CA is a viable target for the development of novel anti-HIV-1 compounds. Ebselen is also a unique CA inhibitor that covalently binds to the C-terminal domain of CA via cysteine residues, inhibiting HIV-1 activity (17).

We previously reported that the insertion of a short amino acid sequence near the Arg18/Thr19 region decreased its stability, inducing abnormal CA degradation (24), and identified an adequate hydrophobic cavity near this region on the surface of CA-NTD, which could be a potential drug target. To the best of our knowledge, none of the other published anti-CA compounds are known to target this hydrophobic cavity.

In this study, we searched for new compounds capable of interacting with the hydrophobic pocket from a library containing millions of commercially available compounds via in silico docking simulations. We identified several compounds as candidate HIV-1 inhibitors that were capable of preventing HIV-1 replication. Here, we highlight ACAi-028, a CA inhibitor with unique molecular characteristics, which possesses potent anti-HIV-1 activity. It is anticipated that both this compound and the identified binding pocket hold potential therapeutic and research applications.

RESULTS

Identification of candidate compounds that target a novel hydrophobic cavity of the HIV-1 capsid.

The CA protein consists of the N-terminal domain (CA-NTD; amino acid residues 1 to 145) and the C-terminal domain (CA-CTD; residues 151 to 231) linked via a short flexible region (residues 149 to 150) (25–27). CA-NTD comprises one β-hairpin, seven α-helices, and a cyclophilin binding loop (CypA-BL), while CA-CTD has a 3₁₀-helix and four α-helices as shown in Fig. 1A. Recently, CA inhibitors have attracted attention due to the development of lenacapavir, whose lead compound is PF74 (15), which exhibits a long-acting and strong anti-HIV profile. We have reported that the insertion of a short amino acid sequence into the CA-NTD, specifically near the R18 and T19 residues, results in spontaneous CA degradation (24).

FIG 1.

Profile of CA, target cavity, and in silico docking simulations of ACAi-028. (A) Secondary structures of HIV-1_NL4-3-derived CA which comprises a β-hairpin in green, CypA-BL in purple, 11 α-helices (H1 to H11) in pink, and 1 3₁₀-helix in yellow. 3D structures of full-length CA (PDB, 4XFX) are shown in tan, as well as a close-up image of ACAi-028 target cavity in the CA-NTD. The ACAi-028 target pocket is highlighted in cyan. (B) The workflow of in silico screening to identify candidate anti-CA inhibitors. (C) The chemical structure of ACAi-028. (D) The docking simulation result of ACAi-028 with the target cavity is shown. Hydrogen bond interactions between the molecular surface of CA-NTD_{1–146/Δ87-99G} crystal and ACAi-028 are indicated in dotted red lines. The carbons of CA and ACAi-028 are shown in tan and white colors, respectively. Nitrogen atoms, oxygen atoms, hydrogen atoms, and bromine atoms are shown in blue, red, white, and brown, respectively. ACAi-028 forms three H-bond interactions with the side chains of Gln13 and Thr19 (interatomic distances of 1.87 Å and 1.89 Å, respectively) as well as with the main chain of Ser16 (interatomic distance of 2.06 Å). Molecular graphics were created with UCSF Chimera (https://www.rbvi.ucsf.edu/chimera). All docking simulations were performed with SeeSAR and FlexX v10 (BioSolveIT GmbH, Sankt Augustin, Germany).

To identify new capsid inhibitors, we selected a novel hydrophobic pocket (Fig. 1A) near this position as a potential interaction site for drug candidates with CA-NTD. In silico docking simulations were performed to search for new compounds that interact with this pocket from a database of over eight million commercially available compounds. The selection process is illustrated in Fig. 1B. More than 40 compounds were identified as candidate HIV-1 inhibitors that prevented HIV-1_LAI (47) replication using the MTT assay.

In this study, we identified ACAi-028 (Fig. 1C), which has a small molecular weight (MW) of 381 g/mol and potent anti-HIV-1_LAI activity (50% effective concentration [EC₅₀], 0.55 μM), among the candidate HIV-1 inhibitors. Therefore, we examined the anti-HIV profile of ACAi-028.

The crystal structure of CA-NTD_{1-146/Δ87-99G} was produced (15), and the binding profile of ACAi-028 to the CA-NTD was elucidated using a docking model (Fig. 1D). ACAi-028 is estimated to form two hydrogen bonding (H-bond) interactions with the side chains of Q13 and T19 (interatomic distances of 1.87 and 1.89 Å, respectively) and one H-bond interaction with the main chain of S16 (interatomic distance of 2.06 Å). This structural arrangement suggests that ACAi-028 could potentially bind to the newly identified hydrophobic cavity in CA-NTD.

ACAi-028 inhibits the early stage of the HIV-1 life cycle.

Next, we determined the anti-HIV activity (EC₅₀s) of ACAi-028 against various HIV-1 and HIV-2 strains, in comparison with several compounds that have been previously reported as capsid inhibitors, such as PF74 (15) and ebselen (17), and an RT inhibitor, azidothymidine (AZT) (28) (Table 1).

TABLE 1.

Anti-HIV activity of ACAi-028 against HIV-1_VSVG-NL4-3, HIV-1_NL4-3, HIV-1_LAI, and HIV-2_ROD

Drug	EC50 (μM) of:
	Single-round inhibitiona	Multiple-round inhibitionb
	TZM-bl	MT-4	MT-2
	HIV-1VSV-G NL4-3	HIV-1NL4-3	HIV-1LAI	HIV-2ROD
ACAi-028	0.48 ± 0.21	0.12 ± 0.04	0.55 ± 0.04	>10
PF74	0.76 ± 0.55	0.33 ± 0.08	0.37 ± 0.50	>10
Ebselen	1.68 ± 1.45	1.83 ± 0.10	1.73 ± 0.04	n.d.c
AZT	0.14 ± 0.13	0.26 ± 0.02	0.23 ± 0.05	n.d.

^aTZM-bl cells (5 × 10⁴/ml) were subjected to HIV-1_{VSV-G NL4-3} (50 ng/ml of p24) in 96-well plates for 48 h in the presence of the tested compounds. The supernatant was removed and lysis buffer added to the samples. After the samples were shaken for 25 minutes, d-luciferin was added to each well. The luciferase intensity was measured using FluoSTAR Omega at 48 h. The assays were conducted in duplicate, and the data shown represent mean values ± SD derived from the results of two or three independent experiments.

^bMT-2 cells (10⁴/ml) were exposed to 100 TCID₅₀ of HIV-1_LAI or HIV-2_ROD and cultured in the presence of various concentrations of each compound, and the EC₅₀ values were determined by the MTT assay. For HIV-1_NL4-3, the EC₅₀ values were determined by using MT-4 cells as target cells. MT-4 cells (10⁵/ml) were exposed to 100 TCID₅₀s of HIV-1_NL4-3, and the inhibition of p24 Gag protein production by each drug was used as an endpoint. All assays were conducted in duplicate, and the data shown represent mean values ± 1 SD derived from the results of two or three independent experiments.

^cn.d., not determined.

ACAi-028 exerted potent anti-HIV-1 activity against single-round HIV-1 infection using vesicular stomatitis virus G (VSV-G) pseudotyped HIV-1_NL4–3 (HIV-1_{VSV-G dENV}), suggesting that ACAi-028 inhibits the early stage of the HIV-1 life cycle. ACAi-028 also prevented multiple-round HIV-1 infection using HIV-1 strains (HIV-1_NL4-3 and HIV-1_LAI), except for HIV-2 strain (HIV-2_ROD), at a concentration similar to that of PF74, ebselen, and AZT. In addition, ACAi-028 inhibited the replication of various types of HIV-1 strains (HIV-1_104pre and HIV-1_MDR/B) (29) in peripheral blood mononuclear cells (PBMCs) and HIV-1_ATV^R_5μM in MT-4 cells (Table 2) and had negligible cytotoxicity in all the cell lines we tested (Table 3).

TABLE 2.

Anti-HIV-1 activity of ACAi-028 against clinically isolated multi-drug-resistant HIV-1 in PBMCs and protease-inhibitor-resistant laboratory strain in MT-4 cells^a

Drug	EC50 (μM) of:
	PBMCs		MT-4
	HIV-1ERS104pre	HIV-1MDR/B	HIV-1ATVR5μM
ACAi-028	0.32 ± 0.07	0.36 ± 0.05	0.64 ± 0.09
PF74	0.29 ± 0.03	0.33 ± 0.02	0.63 ± 0.07
ATV	0.0026 ± 0.0002	0.45 ± 0.01	>1

^aPHA-PBMC (10⁶/ml) or MT-4 cells (10⁵/ml) were exposed to 100 TCID₅₀ of each virus, and the inhibition of p24 Gag protein production by each compound was used as an endpoint. All assays were conducted in duplicate, and the data shown represent mean values ± 1 SD derived from the results of two independent experiments. PHA-PBMCs were derived from a single donor in each independent experiment. HIV-1_ERS104pre was isolated from a treatment-naive patient and served as wild-type HIV-1. HIV-1_MDR/B was originally isolated from an AIDS patient who had received 9 anti-HIV-1 drugs over the 34 months and contains the following amino acid substitutions in the protease-encoding region compared to the consensus type B sequence cited from the Los Alamos database: L10I, K14R, L33I, M36I, M46I, F53I, K55R, I62V, L63P, A71V, G73S, V82A, L90M, and I93L. HIV-1_ATV^R_5μM was generated previously by a long-term selection experiment using HIV-1_NL4-3 with an increasing concentration of ATV (50) and contains the following 10 amino acid substitutions in the protease-encoding region compared with the wild-type HIV-1_NL4-3: L23I, E34Q, K43I, M46I, I50L, G51A, L63P, A71V, V82A, and T91A.

TABLE 3.

Cytotoxicity of ACAi-028, PF74, and ebselen against various cell lines^a

Cell line	CC50 (μM) of:
	ACAi-028	PF74	Ebselen
Li-7 (hepatocyte)	>100	>100	>20
HLE (hepatocyte)	72.3	>100	>20
HEK293 (kidney)	62.6	>100	>20
MT-4 (lymphocyte)	>100	>100	>20
MT-2 (lymphocyte)	86.6	>100	>20
PBMC	>100	n.d.	n.d.

^aCells were incubated with the respective compounds for 1 week before cytotoxicity was quantified using the MTT assay.

To examine the details of early-stage inhibition by ACAi-028, we performed fusion (30), TZM-bl, time-of-addition (31), and colorimetric RT activity assays. ACAi-028 did not significantly inhibit the entry step of HIV-1_LAI compared with the entry inhibitor AMD3100 (32) (Fig. 2A). ACAi-028 also displayed strong anti-HIV-1 activity against HIV-1 R5 strains, including HIV-1_Ba-L and HIV-1_JR-FL, and X4 strains, except for HIV-2_ROD, using the TZM-bl assay (Fig. 2B).

FIG 2.

The early-stage inhibition of ACAi-028 in the HIV-1 life cycle. (A) A fusion assay was conducted between 293T and COS-7 cells in the presence of ACAi-028 or AMD3100. Results of the luciferase intensity are shown as percentages compared with DMSO controls. AMD3100 is a fusion inhibitor against CXCR4 coreceptors. (B) The early-stage inhibition of ACAi-028 (10 μM) against X4-tropic (HIV-1_LAI and HIV-1_NL4-3), R5-tropic (HIV-1_Ba-L and HIV-1_JR-FL), VSV-G HIV-1_NL4-3, and HIV-2_ROD strains were measured using TZM-bl cells. Luciferase activity was measured at 48 h postinfection and is shown as a percentage normalized to DMSO controls. (C) Time-of-addition assay was conducted by the addition of ACAi-028 and various early-stage inhibitors, such as AMD3100, AZT, EFV, PF74, and RAL, to TZM-bl cells in 2-h intervals up to 10 h. Data are shown as percentage every 2 h and normalized to DMSO controls. (D) Effects of ACAi-028, PF-74, and EFV on RT activity were measured by the colorimetric reverse transcriptase assay. An optical density at 405 nm (OD₄₀₅) values were measured and shown as percentages. All assays were performed in duplicate, and error bars indicate ±SD from at least two independent experiments. Statistical significance was examined using Student’s t test; **, P < 0.005.

A time-of-addition assay was conducted to compare the inhibition time of ACAi-028 with that of the other classes of anti-HIV-1 drugs. As expected, AMD3100 ceased displaying anti-HIV-1 activity when it was added more than 2 h postinfection, while an integrase inhibitor, raltegravir (RAL), significantly prevented HIV-1 infection beyond 10 h postinfection (Fig. 2C).

We observed that ACAi-028 inhibited anti-HIV-1 activity at a similar time as RT inhibitors, namely, AZT and efavirenz (EFV) (33), as well as a CA inhibitor, PF74. Fifty percent inhibition induced by these drugs occurred between 4 and 8 h postinfection, which is consistent with a previous report (15). Moreover, we examined whether ACAi-028 prevents HIV-1 RT activity using a colorimetric RT activity assay and found that ACAi-028 and PF74 did not have a significant impact on HIV-1 RT activity compared with EFV (Fig. 2D) in vitro. This result suggests it is unlikely that ACAi-028 is an RT inhibitor. Taken together, these results indicate that ACAi-028 may have a similar level of potency as a CA inhibitor as PF74, which affects the CA uncoating process.

ACAi-028 does not affect the late stage of the HIV-1 life cycle.

To examine the effect of ACAi-028 on late-stage inhibition, we examined whether ACAi-028 affects the process of HIV-1 production, including Gag proteolytic processing and maturation. Forty-eight hours after transfection of pHIV-1_NL4-3 into 293T cells in the presence of ACAi-028 (10 μM), PF74 (10 μM), ebselen (10 μM), or a protease inhibitor, darunavir (DRV) (2 μM) (34), Gag proteins within the cells were observed by Western blotting, and viral production was evaluated by comparing to p24 expression levels in the culture supernatant. ACAi-028, PF74, and ebselen did not affect Gag proteins in the cells, unlike DRV (Fig. 3A). ACAi-028 (88.7%) and ebselen (75.2%) did not significantly interfere with viral production, whereas PF74 (62.9%), and DRV (15.7%) reduced HIV-1 production, which is consistent with previous reports (15, 17) (Fig. 3B). Additionally, the effect of ACAi-028 on Gag proteolytic processing and maturation of HIV-1 virions was investigated using Western blotting and the TZM-bl assay. ACAi-028 did not affect HIV-1 maturation, as observed via Western blotting, using anti-Gag or anti-IN antibodies as well as the dimethyl sulfoxide (DMSO) control (Fig. 3C). ACAi-028 did not reduce the viable virions (111.8%), as opposed to DRV (3.6%) (Fig. 3D), suggesting that ACAi-028 does not have any effect on HIV-1 production and maturation. On the other hand, a high concentration of PF74 (10 μM) did not affect Gag proteins in the cell lysate (Fig. 3A), but the production level of virions in the presence of PF74 significantly decreased by 62.9% compared with that of the DMSO control. Furthermore, the infectivity of the virions was reduced to 6.9% (Fig. 3B and D), suggesting that PF74 affects both the early and late stages, in agreement with previous reports (15). It has been reported that ebselen does not exhibit late-stage inhibition (17). Interestingly, premature products, such as Gag-Pol (p160), Gag-Pol intermediate (p120), Gag (p55), and Gag intermediates, were found in the virion lysate that was produced at a high concentration (10 μM) of Ebselen, via the Western blotting with anti-Gag or anti-IN antibodies, which is similar to DRV (Fig. 3C). Additionally, the infectivity of the virions produced in the presence of ebselen (10 μM) was significantly decreased by 35.4% compared with that of the DMSO control (Fig. 3D). These results suggest that ebselen may interfere with HIV-1 maturation. These results indicate that ACAi-028 is unlikely to affect the late stage of the HIV-1 life cycle.

FIG 3.

Effect of ACAi-028 on the late stage of the HIV-1 life cycle. Gag-Pol proteolytic processing and virus production were examined by Western blotting with the anti-Gag antibody in the cell lysate (A) and measuring the p24 levels in the supernatant of 293T cells which were transfected with pNL4-3 in the presence of ACAi-028 (10 μM), PF74 (10 μM), Ebselen (10 μM), or DRV (2 μM) (B). After the virions in the culture supernatant were purified, Gag-Pol proteolytic processing and HIV-1 maturation of the virions were examined by the Western blotting with anti-Gag on the left side and anti-IN antibody on the right side (C) and by TZM-bl assay at the bottom (D). All assays were performed in duplicate, and error bars indicate ±SD from three independent experiments. Statistical significance was examined using Student’s t test; *, P < 0.05; **, P < 0.005.

Conformational difference of targeting the hydrophobic pocket between HIV-1 and HIV-2.

ACAi-028 did not show anti-HIV activity against the HIV-2_ROD strain (Table 1). We compared the amino acid sequences of experimental HIV-1 strains, such as HIV-1_NL4-3, HIV-1_{HXB (LAI)}, HIV-1_Ba-L, and HIV-1_JR-FL, which are all closely related to simian immunodeficiency virus from chimpanzees (SIV cpz). Similarly, comparisons of experimental HIV-2 strains, such as HIV-2_ROD and HIV-2_EHO, which are related to SIV sooty mangabeys (SIV smn), as shown in Fig. 4A, were also undertaken. The Q13, S16, and T19 residues, which are expected to play important roles in ACAi-028 binding to CA (Fig. 1D), are conserved among all HIV-1 strains and SIV cpz, whereas most of the amino acids from residues 2 to 15 of HIV-2 are different from those of HIV-1 (Fig. 4A). Moreover, P1, Q13, S16, T19, and E45 residues that constitute the ACAi-028 target pocket were highly conserved across 6,144 sequences from all HIV-1 subtypes (HIV sequence database-filtered Web alignment) at conservation rates of 99.85%, 99.89%, 97.85%, 99.72%, and 98.56%, respectively (Fig. 4B).

FIG 4.

Amino acid residues and structures of the CA-NTD among HIV and SIV. (A) Alignment of CA-NTD (residues from 1 to 45) among laboratory HIV-1 strains, SIV_CPZ, HIV-2 strains, and SIV_smn. (B) Representation of frequencies of CA-NTD (residues from 1 to 50) from 6,144 sequences of all HIV-1 subtypes (Los Alamos HIV sequence database-filtered Web alignment) using the WebLogo 3.7.4 application (http://weblogo.threeplusone.com/create.cgi). Highlighted in red are the five residues (Pro1, Gin13, Ser16, Thr19, and Glu45) that constitute the ACAi-028-binding cavity. Above each of these residues is the percentage consensus as determined by the Jalview 2.11.1.4 program (http://www.jalview.org) (51). (C) The target cavity on the surface of the CA-NTD_HIV-1 monomer is shown in tan in the left panel. The CA-NTD_HIV-2 monomer and corresponding target cavity are shown in cyan in the right panel. The bottom panel shows the merged image of the CA-NTD of both strains. (D) The docking simulation result of ACAi-028 with the target cavity in CA-NTD_ROD is shown. Hydrogen bond interactions between the molecular surface of CA-NTD_ROD and ACAi-028 were not observed, unlike the docking result between CA-NTD_HIV-1 and ACAi-028. The carbons of CA and ACAi-028 are shown in cyan and white colors, respectively. Nitrogen atoms, oxygen atoms, hydrogen atoms, and bromine atoms are shown in blue, red, white, and brown, respectively. Docking simulations were performed with SeeSAR and FlexX v10. Molecular graphics were done with UCSF Chimera.

The crystal structures of the ACAi-028 target hydrophobic pocket of HIV-1_NL4-3 (PDB accession number 4XFX) (35) and HIV-2_ROD (PDB accession number 2WLV) (36) are shown in Fig. 4C. This HIV-1_NL4-3 pocket seems to have a sufficient volume for ACAi-028 binding, while that of HIV-2_ROD appears to be too shallow for binding (Fig. 4C). As shown in Fig. 4C, the cavity of CA-NTD_NL4-3 is covered by that of the CA-NTD_ROD in the overlay of these crystal structures, suggesting that the ACAi-028 target volume of HIV-2 is clearly smaller than that of HIV-1.

Moreover, we examined the binding ability of ACAi-028 to CA-NTD_ROD using a binding model. As shown in Fig. 4D, there were no bridging H-bonds between ACAi-028 and the amino acid residues of CA-NTD _ROD, corresponding to the lack of inhibition (EC₅₀, >10 μM) of ACAi-028 against HIV-2 _ROD (Table 1). These results indicate that ACAi-028 may fail to interact with HIV-2_ROD CA, resulting in no anti-HIV-2 activity.

Binding profiles of ACAi-028 to CA proteins.

In order to observe the direct binding of ACAi-028 to CA, we produced recombinant HIV-1_NL4-3-derived CA proteins using Escherichia coli and examined the binding of ACAi-028 to these proteins using an electrospray ionization mass spectrometry (ESI-MS) (37). The ESI-MS spectra of the CA monomer with 1% methanol revealed nine peaks of charged ions in the range of mass/charge ratio (m/z) range of 1,100 to 1,900 (Fig. 5A). The MW estimated from the peaks of charged irons (deconvoluted ESI-MS spectrum) was 25,601.9 Da, corresponding to the theoretical MW of the intact CA monomer (25,602.5 Da), as calculated by Peptide Mass Calculator v3.2 (Fig. 5A). After treatment of CA with ACAi-028, peaks associated with ACAi-028 binding to CA emerged in the m/z range of 1,300 to 1,900 next to each spectrum of the CA monomer (Fig. 5B). The deconvoluted ESI-MS spectrum revealed a peak associated with CA and ACAi-028 at 25,984.1 Da, which was similar to the sum of the MW of a CA monomer (25,601.9 Da) and ACAi-028 (381 Da) (Fig. 5B). Additionally, we examined whether ACAi-028 binds covalently to CA. After treatment of CA with ACAi-028 or ebselen, the samples were denatured by acetonitrile and trifluoroacetic acid. The binding peak of CA with ACAi-028 was not detected in Fig. 5C, while two strong peaks of CA with ebselen (25,922.03 and 26,120.59 Da, representing CA combined with one or two ebselen molecules, respectively) were observed (Fig. 5D). These results are in line with a previous report (17), suggesting that ACAi-028 binds noncovalently to CA, unlike ebselen. Thus, ACAi-028 binds directly and noncovalently to the CA monomer.

FIG 5.

ACAi-028 interacts directly and noncovalently with the CA monomer. ESI-MS spectra of ACAi-028 (50 μM) and Ebselen (50 μM) binding to CA. (A) Black arrows represent CA monomer peaks of the charged ions. Deconvoluted ESI-MS spectrum of the CA monomer is shown on the right side. (B) ESI-MS spectra of CA binding to ACAi-028 are shown in red arrows. Deconvoluted ESI-MS spectra of CA monomer or CA monomer binding to ACAi-028 is shown on the right side. (C) Under denaturing conditions, acetonitrile and trifluoroacetic acid were added to a mixture CA of ACAi-028. Black arrows indicate CA monomers failed to interact with ACAi-028. Deconvoluted ESI-MS spectra are shown on the right side. (D) Under the same conditions as C, black arrows indicate CA monomers and red arrows represent the CA monomer binding covalently to one (left) or two (right) ebselen, respectively. Each deconvoluted ESI-MS spectrum is shown on the right side.

ACAi-028 affects the molecular characterization of CA proteins via S16 and T19 residues.

ACAi-028 was shown to interact directly with the CA proteins using ESI-MS. To investigate how ACAi-028 interacts with CA proteins, we produced CA variants carrying S16E (CA_S16E) or T19A (CA_T19A) amino acid substitutions, which were intended to alter the binding ability of ACAi-028 to CA. CA_M185A carrying an M185A amino acid substitution was also produced as previously reported (38) as a control.

CA thermal stability in the presence of ACAi-028 was examined using differential scanning fluorimetry (DSF) (39, 40) (Fig. 6). The melting temperature (T_m) 50 of wild-type CA (CA_WT) increased in the presence of PF74, while that of ACAi-028 clearly decreased by 6.8°C and 7.1°C at 10 and 50 μM, respectively (Fig. 6A and B). T_m 50 of CA_S16E showed a mild reduction of 3.9°C at 50 μM, while that of CA_T19A remained unchanged in the presence of ACAi-028 (Fig. 6C and D), suggesting that S16 and T19 residues may be associated with the binding of ACAi-028 to the target pocket of CA-NTD. Additionally, the CA multimerization assay was performed in the presence of ACAi-028. The treatment of CA_WT with PF74 increased CA multimerization as previously described (15), whereas ACAi-028 decreased CA_WT multimerization at 4 and 40 μM in a concentration-dependent manner (Fig. 7A). M185A greatly decreased the CA multimerization, in agreement with a previous report (38). The addition of S16E substitution to CA proteins slightly reduced CA_S16E multimerization, while T19A slightly increased CA_T19A multimerization compared with CA_WT (Fig. 7B) at the same sodium concentration, suggesting that these residues significantly affect the CA multimerization. In the presence of ACAi-028, CA_S16E multimerization was similar to that of CA_WT (Fig. 7C). However, CA_T19A multimerization was not affected by the presence of ACAi-028, even at a higher concentration (40 μM) (Fig. 7D), suggesting that the T19A mutant may inhibit ACAi-028 binding to the target pocket or counteract the reduction of CA multimerization induced by the binding of ACAi-028 to CA under these conditions.

FIG 6.

Effect of ACAi-028 on CA thermal stability. (A) DSF was performed using SYBR-orange dye in the presence of methanol shown in black circles; 1, 10, and 50 μM ACAi-028 in light pink, pink, and red squares, respectively; and 10 μM PF74 in blue Xs to examine CA stability. (B) Graphical representation of T_m changes (ΔT_m 50) from A are shown for ACAi-028 and PF74. (C) DSF of CA_S16E and CA_T19A were tested in high concentrations of ACAi-028 (10 and 50 μM shown in pink and red squares, respectively). Black circles indicate methanol as a control. (D) Graphical representation of ΔT_m 50 from C are shown for (10 and 50 μM) ACAi-028. All assays were performed in triplicate, and error bars indicate ±SD from three independent experiments. Statistical significance was examined using Student’s t test; *, P < 0.05; **, P < 0.005.

FIG 7.

Effect of ACAi-028 on CA multimerization. (A) CA multimerization assay was performed by the addition of a high sodium buffer (the ratio of 150 mM sodium phosphate to 5 M sodium phosphate buffer is five to five) in the presence of 4 and 40 μM ACAi-028 shown in pink and red, respectively; PF74 (4 μM) in blue; or RAL (40 μM) in green lines. Turbidity of the mixtures was measured at OD₃₅₀ over a period of 60 minutes. PF74 and RAL are used as positive or negative controls, respectively. Representative data are shown from three independent experiments. (B) CA_WT multimerization was compared with CA_T19A, CA_S16E, and CA_M185A multimerization (the ratio of 150 mM to 5 M sodium is five to five). Effects of ACAi-028 (4 and 40 μM shown in pink and red lines, respectively) on CA_S16E multimerization in higher sodium concentration (the ratio of 150 mM to 5 M sodium is four to six) (C) and CA_T19A multimerization (the ratio of 150 mM to 5 M sodium is five to five) (D) are seen. Representative data are shown from three independent experiments.

The results of both CA DSF and multimerization assays suggested that ACAi-028 may interact with CA proteins via S16 and T19 residues in the target pocket of CA-NTD.

ACAi-028 potentially interacts with the binding pockets of CA in the hexameric state.

ACAi-028 exerted anti-HIV-1 activity in the early stages of the HIV-1 life cycle. In the early stage, CA proteins constitute a capsid lattice (HIV core), which is composed of approximately 250 hexamers and exactly 12 pentamers (41). To confirm the location and space of the binding pocket in a CA hexamer, we analyzed a CA hexamer model in complex with or without ACAi-028, based on CA crystal structures (PDB accession number 3H4E [42] and 5MCX [31]). In the model, the ACAi-028 can bind to the pocket located inside the CA hexamer (PDB, 3H4E) (Fig. 8A). Additionally, the ACAi-028 binding profile to the pocket in a CA dimer extracted from the CA hexameric state (PDB, 5MCX) was predicted using our docking model. ACAi-028 can putatively interact with the pocket in the CA dimer of the CA hexameric state without affecting the paired CA monomer via two H-bonds with L43 and E45 (Fig. 8B), suggesting that ACAi-028 can potentially interact with the binding pockets of the CA hexameric state in the HIV-1 core. Taken together, ACAi-028 is a novel capsid inhibitor that binds to the new hydrophobic pocket in CA-NTD, thereby inhibiting the early stage of HIV-1 replication.

FIG 8.

Interaction of ACAi-028 with the binding pockets of CA in the hexameric state. (A) Left panel shows the structure of a CA hexamer (PDB, 3H4E). Right panel shows the CA hexamer with docking poses of ACAi-028. (B) The structure of a CA dimer extracted from a CA hexamer (PDB, 5MCX) and the docking result of ACAi-028 to the CA dimer are shown. ACAi-028 formed two H-bond interactions with the main chains of Leu43 and Glu45 residues which are located at the monomer-monomer interface of one CA monomer in the CA dimer extracted from the CA hexamer. Docking simulations were performed with SeeSAR and FlexX v10. Molecular graphics were created with UCSF Chimera.

DISCUSSION

Based on the results described above, we concluded that ACAi-028 is a small molecular CA inhibitor of HIV-1 that interacts with CA via the novel region, which has not been previously reported (Fig. 9 and Table 4). The ACAi-028 binding pocket is formed by key residues, namely, Pro1, Gln13, Ser16, Thr19, and Glu45, which constitute the β-hairpin end, flexible linker, and front edge of α-helix 1 (Fig. 1). To understand the mechanisms of action of CA inhibitors previously described (43), we categorized several candidate CA inhibitors based on their molecular characterization into groups A, B, C, and D (Fig. 9 and Table 4). CAP-1 (group A) is a late-stage inhibitor, without early-stage inhibition of the HIV-1 life cycle (13, 14), possibly because the hexameric association of CA proteins in the mature HIV-1 core requires tight molecular packing, which prevents access of CAP-1 to this pocket. Indeed, the other small molecules, including BD-1 and BM-1 (18), which share the CAP-1 binding region are all late-stage inhibitors. Therefore, the inhibitory mechanism of group A is likely different from that of ACAi-028. PF74 (group B) interacts with a region distinct from ACAi-028 and CAP-1, as shown in Fig. 9 and Table 4, and binds to the pocket formed between CA-NTD and the CA-CTD of an adjacent CA monomer. PF74 is known to inhibit both the early and late stages of the HIV-1 life cycle (15), whereas ACAi-028 inhibited only the early stage (Fig. 2 and 3). It is necessary to elucidate the inhibitory mechanism of the capsid inhibitors in the late stage (during HIV-1 production and maturation) to distinguish the anti-CA mechanism between ACAi-028 and PF74. As shown in the CA multimerization and DSF assay (Fig. 5), ACAi-028 has a significant effect on the CA multimerization and thermal stability, which are completely opposite to PF74, suggesting that ACAi-028 and PF74 represent different classes of CA inhibitors (Fig. 9 and Table 4). I-XW-053 (19) (group C) binds to and disrupts CA NTD-NTD interactions in CA hexamers during the early stages of HIV-1 infection. By surface plasmon resonance (SPR) analysis of I-XW-053 binding with CA mutants, the proposed binding sites of I-XW-053 were found to involve I37 and R173, which are different from the amino acids identified within the target binding region of ACAi-028. C1 (group C) is also a late-stage assembly inhibitor that interacts with the CA-NTD at E98, H120, and I124 residues (20). Ebselen (group D) is a covalent inhibitor of HIV-1 CA by forming a selenosulfide bond with C198 and C218 residues in the CA-CTD. ACAi-028 possesses distinct properties, in comparison to the inhibitors of groups C and D, suggesting that ACAi-028 targeting the hydrophobic pocket is likely to be different from any previously discovered CA inhibitors.

FIG 9.

Profiles of ACAi-028 and CA inhibitors. ACAi-028 and representative CA inhibitors previously reported are categorized into groups new, A, B, C, and D. Names, chemical structures, effects of CA multimerization, and inhibition stages of the HIV-1 life cycle, as well as putative and representative binding regions of the CA inhibitors, are shown.

TABLE 4.

Profiles of ACAi-028 and the other CA inhibitors^a

CA	Group	Name	Binding region(s)	C.M.b	Stage of inhibition	Reference(s)
NTD	New	ACAi-028	Q13, S16, T19	↓	Early
	A	CAP-1	E28, E29, F32, V59, H62	↓	Late	13, 14
		BD-1, BM-1	F32, H62	↓	Late	18
	B	PF74	Q67, K70, T107	↑	Early/late	15
		(GS-CA1) lenacapavir	Q67, K70, T107	↑	Early/late	21, 22
	C	I-XW-053	I37	↓	Early	19
		C1	E98, H120, I124	↓	Late	20
CTD	D	Ebselen	C198, C218	↑	Early/late	17
	B	PF74	Y169, L172, R173, Q179	↑	Early/late	15
		(GS-CA1) lenacapavir	Y169, L172, R173, Q179	↑	Early/late	21, 22
	C	I-XW-053	R173	↓	Early	19

^aThe binding region of ACAi-028 was determined using virtual docking simulation (Fig. 1). ACAi-028 is an early stage inhibitor (Fig. 2) that induces reduction of capsid multimerization (Fig. 7). Please refer to the references listed for details of the mechanisms of the other CA inhibitors.

^bC.M., CA Multimerization; ↑, increase; ↓, decrease.

Amino acid substitutions that alter CA multimerization have been previously reported (38). Substitution of amino acid residues at S16 and T19 altered the CA multimerization; S16E mutants decreased the CA multimerization, whereas T19A mutants increased the CA multimerization. Additionally, the E45 residue is a residue predicted to be located in the ACAi-028-binding pocket, and E45A substitution is known to increase the CA multimerization (44). These amino acids might constitute an important site for CA dimerization in the formation of CA hexamers (52) (Fig. 8). The binding of ACAi-028 to the pocket could potentially interfere with CA dimerization because this pocket is located within an inward-facing portion of the CA hexamer (Fig. 8A), which is supported by the reduction of CA multimerization in the presence of ACAi-028 (Fig. 5A). To investigate the binding profile of ACAi-028 to CA-NTD, we produced crystals of CA-NTD_{1-146/Δ87-99G} proteins according to a previous report (15) and utilized this structure for the preparation of a docking model for ACAi-028 binding to CA. Unfortunately, when crystals of CA-NTD_{1-146/Δ87-99G} were produced in the presence of ACAi-028, significant precipitations occurred without the emergence of a crystal in the droplet, suggesting that ACAi-028 may cause instability or destabilization of the CA protein. Thus, we were unable to acquire actual binding profiles of ACAi-028 to CA. Moreover, when we have selected HIV-1 variants resistant to ACAi-028 with increasing concentrations of up to 20 μM, we did not detect ACAi-028-related strong resistant mutations in the CA (Gag) regions, suggesting that ACAi-028 might still possess an unknown mechanism for HIV-1 inhibition. This represents a limitation of the present study and requires further investigation.

Recently, lenacapavir was reported to be a powerful anti-HIV compound with broad-spectrum inhibition even against multidrug-resistant HIV-1, HIV-2, and SIV (21) for long-acting HIV-1 treatment (21, 22) (phase 2/3 CAPELLA trial). Furthermore, lenacapavir has also been predicted to be effective in HIV prevention as an HIV pre-exposure prophylaxis (PrEP). These findings strongly suggest that CA is an attractive therapeutic target for the development of novel antivirals.

In conclusion, we have identified ACAi-028 as a small molecular anti-HIV-1 CA inhibitor that targets a novel hydrophobic CA-NTD pocket and exerts the early-stage inhibition of the HIV-1 life cycle with EC₅₀ of 0.55 μM. Further research is under way to understand the role of this region in HIV-1 replication. The novel hydrophobic pocket identified here should be a viable target for the development of new synthetic CA inhibitors. Furthermore, ACAi-028, a potent CA inhibitor targeting this novel pocket, could have valuable therapeutic and research applications.

MATERIALS AND METHODS

Cells and viruses.

MT-2 and MT-4 cells (Japanese Collection of Research Bioresources Cell Bank [JCRB Cell Bank], Japan) were cultured in RPMI 1640 medium (Gibco, Thermo Fisher Scientific, USA) with fetal bovine serum (FBS; Gibco, Thermo Fisher Scientific, USA), penicillin (P), and kanamycin (K). 293T, Li 7, HLE, and COS7 cells obtained from JCRB Cell Bank and TZM-bl cells obtained from the NIH AIDS Research and Reference Reagent Program were cultured in low-glucose Dulbecco’s modified Eagle’s medium (DMEM) with l-glutamine and phenol red (Fujifilm Wako Pure Chemical Corporation, Japan), as well as FBS, P, and K. Phytohemagglutinin (PHA)-PBMCs were derived from a single donor in each independent experiment. The research protocols described in the present study were carried out in accordance with relevant guidelines and regulations and were approved by the Ethics Committee for Epidemiological and General Research at the Faculty of Life Sciences, Kumamoto University. The HIV-1 strains used in our experiments have already been established previously, including HIV-1_NL4-3, HIV-1_LAI, HIV-1_Ba-L, HIV-1_JR-FL, HIV-1_ERS104pre which was isolated from clinical HIV-1 strains of drug-naive patients with AIDS (45), and HIV-1_MDR/B which was originally isolated from AIDS patient who had received 9 anti-HIV-1 drugs over the 34 months and was genotypically and phenotypically characterized as a multi-drug-resistant HIV-1 variant (46, 47). HIV-1_{VSV-G dENV} was produced by cotransfection of a pCMV-VSV-G vector (addgene) and pNL4-3 _dENV with a deleted KpnI-NheI site in the Env region into 293T cells.

Plasmid constructs.

Full-length CA sequences derived from pNL4-3 were introduced to pET30a vectors (Novagen-Merck KGaA, Germany), producing a pET30a CA vector. The site-directed mutagenesis was performed using PrimeSTAR Max (TaKaRa Bio, Inc., Japan) to introduce S16E, T19A, E45A, and M185A into the pET30a CA, producing pET30a CA_S16E, CA_T19A, CA_E45A, and CA_M185A vectors, respectively. The CA-NTD₁₄₆ deletion sequence carrying a single glycine residue instead of CypA-BL (residues 87 to 99) (the amino acid sequence is based on data in reference 15) was introduced to pET30a vectors with 6×His at N terminus, producing a pET30a His-CA_{1–146/Δ87-99G} vector.

Protein expression and purification.

CA proteins were produced from the pET30aCA in E. coli Rosetta (DE3) pLysS competent cells (Novagen) grown in LB medium supplemented with K and chloramphenicol at 37°C and induced with 1.0 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) for 3 to 5 h at 37°C. The bacterial cells were harvested and stored at –80°C. The pellets of CA were resuspended and sonicated in CA buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, and 5 mM 2-mercaptoethanol [BME]) supplemented with 0.5 mM phenylmethylsulfonylfluoride. The lysates were cleared by centrifugation for 15 min at 3,500 rpm at 4°C. After 5 M NaCl was added to the supernatants, the samples were cleared by centrifugation for 15 min at 3,500 rpm again. The precipitate was resuspended in the CA buffer. The sample was precleared for 15 min at 15,000 rpm at 4°C, and the supernatants were filtered through a 0.45-μm filter. The CA proteins were loaded onto a HiLoad 16/60 Superdex 200 column (GE Healthcare, USA) and were eluted with CA loading buffer (20 mM Tris-HCl [pH 7.5], 150 mM NaCl, and 5 mM BME) using AKTAprime plus (GE Healthcare). CA monomer fractions were concentrated using an Amicon Ultra-10K device (Merck Millipore) in the CA buffer. The protein concentration was determined using a bicinchoninic acid (BCA) protein assay reagent kit (Thermo Fisher Scientific) and stored at −80°C.

In silico simulation and docking model.

The crystal structures of CA-NTD_{1-146/Δ87-99G} produced by our method and various full-length CA proteins (PDB accession number 4XFX, 3H4E, and 5MCX) from the RCSB Protein Data Bank (http://www.rcsb.org) were utilized for the docking simulation. Hydrogens atoms were added to the 2D structure of ACAi-028, and the structures were energy minimized with the MMFF94x force field as implemented in MOE (Chemical Computing Group, QC, Canada). All docking simulations were performed with SeeSAR and FlexX v10 (BioSolveIT GmbH, Sankt Augustin, Germany). Molecular graphics and analysis were performed with UCSF Chimera (https://www.rbvi.ucsf.edu/chimera).

Fusion assay.

A fusion assay was performed as previously described (30). In brief, 293T cells were transfected with pHIV-1_NL4-3 Tat with or without pHIV-1_NL4-3 Env, while COS-7 cells were transfected with CD4, CXCR4, and LTR-luciferase. After 24 h of incubation at 37°C and 5% CO₂, the transfected 293T cells were mixed with the transfected COS-7 cells in the presence or absence of the tested compounds for 6 h. The luciferase of the samples was detected using firefly luciferase reporter assay kit I (PromoCell GmbH, Germany), and the luciferase intensity was normalized to the negative control. Ratios of luciferase intensity of the samples were compared.

Time-of-addition assay.

Time-of addition assay was previously reported (31). Briefly, TZM-bl cells (5 × 10⁴/ml) were subjected to HIV-1_LAI (50 ng/ml of p24) in 96-well white plates. Drugs at the indicated concentrations were also added to the set of wells demarcated for 0 h. After 2 h of incubation at 37°C and 5% CO₂, the supernatant was removed and the cells were washed once to remove all traces of the virus. Drugs at the respective concentration were added back to the cells marked for 0 h as well as the ones marked for 2 h. Subsequently, drugs at the correct concentration were added every 2 h up to 10 h during incubation. The cells were incubated at 37°C and 5% CO₂ until 48 h. Luciferase intensity was measured using a FluoSTAR Omega instrument (BMG Labtech GmbH, Germany).

TZM-bl assay.

TZM-bl cells were obtained from the NIH AIDS Research and Reference Reagent Program. The TZM-bl assay was performed in a 96-well white plate using the firefly luciferase reporter assay kit I. Briefly, the supernatant medium was removed and lysis buffer added to the samples. After the samples were shaken for 25 minutes, d-luciferin was added to each well. After the shaking step, the luciferase intensity was measured using a FluoSTAR Omega instrument.

Virus quantification.

Virus samples were measured by an HIV-1 p24 enzyme-linked immunosorbent assay (ELISA) using Lumipulse G1200 (Fujirebio, Inc., Japan) and normalized to determine the viral concentration.

Reverse transcriptase assay.

A colorimetric reverse transcriptase assay was performed (Roche, Switzerland). Briefly, recombinant HIV-1 reverse transcriptase was added to the tested compounds dissolved in the lysis buffer and subsequently incubated at 37°C and 5% CO₂ for 3 to 4 h. After samples were washed, a peroxidase-conjugated anti-digoxigenin antibody solution was added to the samples and incubated for 1 h at 37°C and 5% CO₂. After another washing step, a peroxidase substrate ABTS solution with enhancer was added to the samples. The optical density was measured at 405 nm using a Versamax microplate reader (Molecular Devices, USA).

Western blotting.

Western blotting was described previously (40). Briefly 293T cells were plated and incubated for 24 h at 37°C in 5% CO₂. Cells were transfected with pNL4-3 vectors using the Attractene transfection reagent (Qiagen, Germany). After 8 h, the medium was changed, the tested compounds were added, and then samples were incubated for 48 h. Subsequently, the viruses were filtered, purified by ultracentrifugation in 15% sucrose-phosphate-buffered saline (PBS), normalized by the p24 levels, and stored in PBS at −80°C. The cells were lysed in M-per buffer (Thermo Fisher Scientific) supplemented with halt protease inhibitor cocktail (Thermo Fisher Scientific). The samples were titrated using BCA protein assay kit and stored at −80°C. The samples were prepared and separated with SDS-PAGE (5 to 20% Extra PAGE one precast gel; Nacalai Tesque) and transferred onto a nitrocellulose membrane. The samples were detected with an anti-HIV-1 Gag (p55 + p24 + p17) antibody (catalog number ab63917; Abcam), anti-HIV-1 IN antibody (catalog number ab66645; Abcam), second mouse or rabbit antibody (MBL, Co., Ltd.), and anti-beta actin antibody (horseradish peroxidase [HRP] conjugated) (Abcam) and then visualized using SuperSignal West Pico chemiluminescent substrates (Thermo Fisher Scientific).

CA multimerization assay.

CA multimerization assays were performed as previously reported (38). The compounds were added to 30 μM of CA proteins in 50 mM sodium phosphate (pH 8.0) supplemented with 50 mM NaCl. Capsid assembly was initiated by the addition of 50 mM sodium phosphate (pH 8.0) supplemented with 5 M NaCl. Optical density at 350 nm was measured using a FLUOstar Omega reader (BMG Labtech) for 1 h.

Differential scanning fluorimetry.

The DSF method was previously described (39, 40). In brief, recombinant CA proteins (25 μM) were prepared in PBS. After CA treatment with tested compounds for 5 to 8 h on RT, SYPRO orange (Life Technologies) was added to the samples (final concentration of SYPRO orange, 5×). The samples were successively heated from 25 to 95°C, and the increasing fluorescence intensities were measured by the real-time PCR system 7500 Fast (Applied Biosystems, Thermo Fisher Scientific). Data were indicated as a relative ratio between the minimum and maximum intensity of SYPRO orange from 25 to 95°C detected for each sample.

ESI-MS.

The ESI-MS protocol was previously described (37). In brief, MS spectra of CA in the presence of ACAi-028 were obtained using an electrospray ionization (ESI) quadrupole time of flight (QTOF) mass spectrometer (impact II; Bruker Daltonics). Each sample solution under native conditions was introduced to the ESI-QTOF mass spectrometer through an infusion pump at a flow rate of 3.3 μl/min. To detect the denatured samples, an analysis was done using the QTOF mass spectrometer equipped with a CaptiveSpray electrospray ionization platform with liquid chromatography (Ultimate 3000 high-pressure liquid chromatography [HPLC]; Thermo Fisher Scientific). The following ion source parameters were applied: dry heater, 150°C; dry gas, 8.0 liters/min; capillary voltage, 1,000 V; and end plate offset, −500 V. MS scans were acquired at a spectra rate of 1 Hz at a mass range from 100 to 3000 m/z. Molecular weights by protein deconvolution were determined using Data Analysis 4.4 (Bruker Daltonics). The MW of CA proteins was calculated using Peptide Mass Calculator v3.2 (http://rna.rega.kuleuven.be/masspec/pepcalc.htm).

Crystallization and X-ray data collection.

The crystallization procedure was performed according to methods in a previous report (15) (PDB, 2XDE). Briefly, crystallization was performed by the hanging-drop vapor diffusion method using EasyXtal 15-well tools (Qiagen). CA-NTD_{1-146/Δ87-99G} proteins were expressed and purified as described above. The protein concentration was adjusted to 2 mg/ml. The reservoir solution consists of 100 mM phosphate-citrate (pH 4.2), 200 mM NaCl, and 20% polyethylene glycol 8000 (PEG 8000). The crystals reached 0.2 to 0.4 mm within 1 week at 10°C. The crystals were transferred to a reservoir solution supplemented with 25% glycerol and flash frozen at 100 K. Then, X-ray diffraction experiments were carried out. Data collection and refinement statistics were subsequently examined.

Drug susceptibility assay.

The susceptibility of HIV-1_LAI and HIV-2_ROD to ACAi-028 and control drugs/compounds were determined as previously described (48). Briefly, MT-2 cells (10⁴/ml) were exposed to 100 × 50% tissue culture infectious dose (TCID₅₀) of HIV-1_LAI or HIV-2_ROD in the presence or absence of various concentrations of compounds in 96-well plates and were incubated at 37°C for 7 days. After incubation, 100 μl of the medium was removed from each well and MTT solution was added to each well in the plate, followed by incubation at 37°C for 1.5 to 4 h. After incubation to dissolve the formazan crystals, acidified isopropanol containing 4% (vol/vol) Triton X-100 was added to each well and the optical density measured in a kinetic microplate reader (V_max; Molecular Devices, Sunnyvale, CA). All assays were performed in duplicate. The 50% cytotoxic concentration (CC₅₀) of compounds for each cell line was also evaluated by MTT assay. To determine the sensitivity of primary HIV-1 isolates to compounds, PHA-PBMCs (10⁶/ml) were exposed to 50 TCID₅₀ of each primary HIV-1 isolate and cultured in the presence or absence of various concentrations of drugs in 10-fold serial dilutions in 96-well plates. In determining the drug susceptibility of certain laboratory HIV-1 strains, MT-4 cells were employed as target cells as previously described (49, 50) with minor modifications. In brief, MT-4 cells (10⁵/ml) were exposed to 100 TCID₅₀ of drug-resistant HIV-1 strains in the presence or absence of various concentrations of compounds and were incubated at 37°C. On day 7 of culture, the supernatants were harvested and the amounts of p24 (CA) protein were determined by using the Lumipulse G1200 system. Drug concentrations that suppressed the production of the p24 Gag protein by 50% (50% effective concentration [EC₅₀]) were determined by comparison with the p24 production level in drug-free control cell culture. PHA-PBMCs were derived from a single donor in each independent experiment.

Compounds.

ACAi-028 was purchased from ChemBridge (San Diego, CA, USA). AZT, PF74, and EFV were purchased from Sigma-Aldrich, Ebselen from AdipoGen Life Sciences, and AMD3100 from Selleck Chemicals. Atazanavir (ATV) was kindly provided by Bristol Myers Squibb (New York, NY).