Discovery of novel 1,2,4-triazole phenylalanine derivatives targeting an unexplored region within the interprotomer pocket of the HIV capsid protein

Abstract

HIV capsid (CA) protein is a promising target for developing novel anti-HIV drugs. Starting from highly anticipated CA inhibitors PF-74, we used scaffold hopping strategy to design a series of novel 1,2,4-triazole phenylalanine derivatives by targeting an unexplored region composed of residues 106–109 in HIV-1 CA hexamer. Compound d19 displayed excellent antiretroviral potency against HIV-1 and HIV-2 strains with EC₅₀ values of 0.59 μM and 2.69 μM, respectively. Additionally, we show via surface plasmon resonance spectrometry (SPR) that d19 preferentially interacts with the hexameric form of CA, with a significantly improved hexamer/monomer specificity ratio (Ratio = 59) than PF-74 (Ratio = 21). Moreover, we show via SPR that d19 competes with CPSF-6 for binding to CA hexamers with IC₅₀ value of 33.4 nM. Like PF-74, d19 inhibits the replication of HIV-1 NL4.3 pseudo typed virus in both early and late stages. In addition, molecular docking and molecular dynamics simulations provide binding mode information of d19 to HIV-1 CA and rationale for improved affinity and potency over PF-74. Overall, the lead compound d19 displays a distinct chemotype form PF-74, improved CA affinity, and anti-HIV potency.

1. Introduction

The human immunodeficiency virus (HIV) can attack the immune system cells, eventually leading to acquired immunodeficiency syndrome (AIDS), which remains a significant threat to public health.¹ HIV includes two subtypes: HIV-1 is the major contributor to the pandemic with high infectivity and mortality, while HIV-2 is mainly found in West Africa with reduced virulence.² Nonetheless, it cannot be ignored that HIV-2 infection has been observed worldwide.^3,4 Although highly active antiretroviral therapy (HAART) can improve the patient’s symptoms, there is no cure for AIDS.^5–7 This requires the development of novel antiretroviral drugs with new targets and unique mechanisms.⁸ The HIV-1 capsid (CA) protein has been regarded as an attractive target due to its critical role in numerous steps of viral replication.^9.10

The HIV CA is a self-assembling protein and has multiple oligomeric forms. The CA monomer consists of an N-terminal domain (NTD, residues 1–145) and a C-terminal domain (CTD, residues 150–231), connected by a flexible linker.¹¹ In the early stage of the viral replication cycle, HIV completes nuclear import and uncoating orchestrated by CA and host factors. Such factors include cleavage and polyadenylation specificity factor-6 (CPSF6), nucleoporin 153 and 358, and cyclophilin A.¹² In the late stage of the HIV replication cycle, individual CA monomers are assembled into pentamers and hexamers, which form the fullerene capsid cone. A mature HIV capsid contains about 250 hexamers and 12 pentamers associated through CA-CA interactions.¹³ The proper assembly and integrity of HIV-1 capsid are significant for viral infectivity.¹⁴ Therefore, CA-targeting small molecules have received extensive attention as they could interfere with multiple processes of virus replication. ^{15, 16}

Compound PF-74 (Figure 1) is a classic HIV CA inhibitor interfering with CA-CA and CA-host factors interactions.^17–19 However, PF-74 suffers from low anti-HIV-1 potency (EC₅₀ = 0.52 – 0.90 μM) and undesirable metabolic stability (T_1/2 = 0.5 – 1.3 min), making it unsuitable for clinical application.^20–23 Extensive decoration of the PF-74 scaffold by Gilead Sciences created GS-6207 (Figure 1) that has picomolar anti-HIV-1 activity and is being evaluated in clinical phase III.^{24, 25} Nevertheless, limited by the complicated structure and labor-intensive synthesis,²⁶ GS-6207 displays extremely poor water solubility, resulting in only injectable use.²⁷ This necessitates further structural optimization of PF-74 to identify novel HIV-1 capsid protein inhibitors with properties commensurate with inclusion in the regular cART regimens.

Figure 1.

Structures of PF-74 and GS-6207.

Structures of PF-74 bound to the CA hexamer offers insights and opportunity for further evolution of the PF-74 scaffold.²⁸ As shown in Figure 2, the co-crystal structure of PF-74/CA hexamer indicates that the phenylalanine core interacts with the NTD (green) in one CA monomer. At the same time, the methylindole group mainly acts on the CTD (purple) of adjacent monomers. To improve the PF-74 and NTD-CTD interface interactions, we replaced the acylamino group in the phenylalanine backbone with a 1,2,4-triazole group via a scaffold-hopping strategy. In addition, various thioether substituents were introduced into the 5-position of the triazolyl to take advantage of potential additional interactions with amino acid (aa) residues 106–109, which the parental PF-74 compound does not contact.

Figure 2.

The binding mode of PF-74 (cyan) within the interprotomer pocket (NTD in green and CTD in purple) of a CA hexamer (PDB code: 5HGL). Dashed circles highlight inhibitor design strategy/location. The purple dashed circle highlights the cyclization strategy, and the red dashed circle (residue 106–109) location for chemical expansion and exploration of new contacts within the interprotomer pocket.

Herein, 25 novel 1,2,4-triazole phenylalanine derivatives were designed, synthesized and evaluated for their anti-HIV potency. From this group, representative compounds were selected to study the mechanism of action using surface plasmon resonance (SPR) interaction analysis and competition assays, combined with early and late-stage potency assessment in the single-round infection assay. Molecular docking and molecular dynamics (MD) simulation studies were conducted to indicate the potential binding mode of compound d19 within the NTD-CTD interface. Finally, we provide the metabolic stability profiles of representative compounds in human liver microsomes, highlighting slight improvements over PF-74.

2. Results and discussion

2.1. Synthesis of the target molecules

As shown in Scheme 1, L-phenylalanine methyl ester hydrochloride (a) was reacted with 2-(2-methyl-1H-indol-3-yl)acetic acid to yield the intermediate b, which was converted to compound c via a hydrazinolysis reaction.^{29, 30} Next, compound c was reacted with 1-isothiocyanato-4-methoxybenzene, and then the mixture was treated with NaOH to obtain target compound d1.³¹ Finally, other target compounds d2–25 were provided by the reaction of d1 with various chlorine or bromine substituents.³²

Scheme 1.

Reagents and Conditions. (i) 2-(2-methyl-1H-indol-3-yl)acetic acid, HATU, DIEA, CH₂Cl₂, 0℃-r.t., 8~10 h; (ii): H₄N₂·H₂O, EtOH, 78℃, 10 h; (iii) 1) 1-isothiocyanato-4-methoxybenzene, DMF, 80℃, 6 h; 2) NaOH, H₂O, 100℃, 6 h; 3) HCl, H₂O, neutralized; (iv): various chlorine or bromine substituents, Cs₂CO₃, 50℃, 8 h.

2.2. Biological activity

The target compounds d1–25 were evaluated for their antiretroviral potency against wild-type HIV-1 and HIV-2 strains in MT-4 cells via MTT method,³³ with PF-74 as a control. The biological evaluation results expressed as EC₅₀ (anti-HIV-1/2 activity), CC₅₀ (cytotoxicity), and selectivity index (SI, CC₅₀/EC₅₀ ratio), are shown in Table 1.

Table 1.

Anti-HIV-1/2 activity and cytotoxicity of novel 1,2,4-triazole phenylalanine derivatives d1-d25

Compound	R	EC50 (μM)a		CC50 (μM)b	SIc
		HIV-1	HIV-2		HIV-1	HIV-2
d1	H	4.94 ± 1.60	10.60 ± 1.51	133.24 ± 9.75	27	13
d2		3.01 ± 1.26	6.59 ± 0.91	82.27 ± 40.08	27	12
d3		11.39 ± 2.13	19.24 ± 5.14	> 214.14	> 19	> 11
d4		2.89 ± 0.34	4.27 ± 1.67	24.84 ± 2.41	9	6
d5		2.61 ± 0.39	4.16 ± 1.27	24.25 ± 2.28	9	6
d6		2.99 ± 0.12	> 7.81	23.99 ± 1.91	8	NA
d7		3.87 ± 0.94	15.95 ± 4.63	162.45 ± 25.95	42	10
d8		1.24 ± 0.57	4.07 ± 0.87	25.00 ± 1.45	20	6
d9		2.54 ± 0.40	7.33 ± 1.71	24.36 ± 1.59	10	3
d10		0.79 ± 0.29	4.98 ± 2.08	81.87 ± 28.18	103	16
d11		0.88 ± 0.26	3.74 ± 0.78	118.47 ± 5.34	135	32
d12		2.87 ± 0.14	> 30.83	30.83 ± 7.11	11	NAd
d13		0.59 ± 0.08	3.73 ± 1.02	48.52 ± 19.89	83	13
d14		NDe	NDe	> 205.65	NAd	NAd
d15		1.81 ± 0.74	7.43 ± 1.62	105.91 ± 16.39	58	14
d16		0.85 ± 0.24	> 22.95	22.95 ± 0.38	27	NAd
d17		2.43 ± 0.21	7.89 ± 0.65	171.90 ± 7.42	71	22
d18		2.47 ± 0.11	3.42 ± 1.93	> 187.46	> 76	> 55
d19		0.59 ± 0.06	2.69 ± 0.17	95.80 ± 69.57	163	36
d20		2.05 ± 0.47	3.94 ± 1.17	24.65 ± 0.75	12	6
d21		1.64 ± 0.47	6.65 ± 1.13	32.33 ± 6.97	20	5
d22		2.71 ± 0.19	> 20.76	20.76 ± 0.87	8	NAd
d23		0.90 ± 0.34	3.98 ± 2.02	33.87 ± 7.12	38	9
d24		2.82 ± 0.16	≥ 7.94	33.00 ± 6.03	12	NAd
d25		2.88 ± 0.18	> 22.58	22.58 ± 0.63	8	NAd
PF-74		0.79 ± 0.09	3.79 ± 0.39	> 293.75	> 371	> 77

^aEC₅₀: concentration required to achieve 50% protection of MT-4 cell cultures against HIV-1 induced cytopathicity, as determined using the MTT method.

^bCC₅₀: concentration required to reduce the viability of mock-infected cell cultures (cytotoxicity, CC) by 50%, as determined using the MTT method.

^cSI: selectivity index, the ratio of CC₅₀/EC₅₀.

^dNA: no applicable.

^eND: not determined.

Except for d14, all target compounds showed from moderate to excellent anti-HIV-1 activity with the EC₅₀ values between 11.39 μM and 0.59 μM. When the substituents are N-propyl-morpholinyl and cyanomethyl, compounds d13 (EC₅₀ = 0.59 ± 0.08 μM) and d19 (EC₅₀ = 0.59 ± 0.06 μM) displayed the most potent activity against HIV-1, which were slightly better than the lead PF-74 (EC₅₀ = 0.79 ± 0.09 μM). Besides, four compounds d10 (R = N-ethyl-morpholinyl, EC₅₀ = 0.79 ± 0.29 μM), d11 (R = acetylaminyl, EC₅₀ = 0.88 ± 0.26 μM), d16 (R = N,N-dimethylethanaminyl, EC₅₀ = 0.85 ± 0.24 μM) and d23 (R = 4-methyl-thiazolyl, EC₅₀ = 0.90 ± 0.34 μM) demonstrated outstanding anti-HIV-1 potency comparable to PF-74. Other compounds possessed single-digit micromolar anti-HIV-1 activity apart from d2 (R = methyl propionatyl, EC₅₀ = 11.39 ± 2.13 μM).

Despite only accounting for a fraction of the global HIV burden, we tested these compounds against HIV-2. Patients infected with this virus have limited treatment options compared to those infected with HIV-1. This is because HIV-1 has an intrisic resistance to non-nucleoside reverse transcriptase inhibitors.³⁴ 18 compounds displayed micromolar antiretroviral efficacy against HIV-2, with the EC₅₀ values ranging from 19.24 μM to 2.69 μM. Particularly, compound d19 (R = cyanomethyl, EC₅₀ = 2.69 ± 0.17 μM) displayed slightly potency enhancement over PF-74 (EC₅₀ = 3.79 ± 0.39 μM). In addition, compounds d4 (R = 4-methylpyridinyl), d5 (R = 3-methylpyridinyl), d8 (R = fluoroethanyl), d10 (R = N-ethyl-morpholinyl), d11 (R = acetylaminyl), d13 (R = N-propyl-morpholinyl), d18 (R = 4-sulfonamide-benzyl), d20 (R = propenyl) and d23 (R = 4-methyl-thiazolyl) showed the similar anti-HIV-2 potency to PF-74. The remaining compounds exhibited decreased or lost anti-HIV-2 activity at the tested concentration.

Compound d19 displayed slightly better anti-HIV-1/2 potency than PF-74, with relatively low cytotoxicity (CC₅₀ = 95.80 μM). Thus, d19 is a promising lead compound with a novel structure that deserves further study.

2.3. Surface plasmon resonance (SPR) binding assay

In order to verify the target specificity of newly designed compounds, the surface plasmon resonance (SPR) binding assay was performed.³⁵ According to the results of anti-HIV-1 activity, the most active compounds d13 and d19 were selected to determine the affinity with HIV-1 CA monomer and hexamer, respectively. As described in Table 2, the affinity of compounds d13 (K_D = 2366.2 ± 1850.2 nM) and d19 (K_D = 3498.7 ± 1691.0 nM) for CA monomer is weaker than PF-74 (K_D = 968.5 ± 446.7 nM). Significantly, d13 (K_D = 117.7 ± 7.2 nM) and d19 (K_D = 59.0 ± 1.0 nM) displayed excellent affinity binding to CA hexamer, which is slightly weaker or equivalent to PF-74 (K_D = 47.0 ± 0.6 nM). These results displayed that this series of compounds are more inclined to interact with CA hexamer, and the selectivity of d19 for the two protein constructs is significantly better than that of PF-74. Hence, these compounds could be identified as typical HIV-1 CA inhibitors with a brand-new structure. The amino acid sequences (GenBank accession number: M30895) and NTD 3-D structure (PDB code: 2WLV) of HIV-2 CA protein are significantly similar to that of HIV-1 (GenBank accession number: AF324493; PDB code: 3H4E). In particular, the amino acids at the binding site of PF-74 are the same, except for amino acids 69 and 70. Therefore, these compounds could also bind to HIV-2 capsid protein.

Table 2.

SPR results of d13, d19, and PF-74 binding to CA monomer and hexamer

Compds	KD (nM)a		Ratiob
	Monomer	Hexamer
d13	2366.2 ± 1850.2	117.7 ± 7.2	20
d19	3498.7 ± 1691.0	59.0 ± 1.0	59
PF-74	968.5 ± 446.7	47.0 ± 0.6	21

^aAll values represent the average response from at least 3 replicates.

^bRatio: K_{D (monomer)}/K_D(hexamer)

2.4. SPR-based competition assay with CPSF6

As mentioned above, PF-74 could disturb the CA-host factor interactions by acting on capsid proteins competitively with host factors, such as CPSF6 and nucleoporin 153. Therefore, compound d13 and d19 were selected for competitive CA hexamer binding assay with CPSF6 based on SPR technology.³⁶ As Figure 3 showed, d13 and d19 effectively inhibited the binding of CPSF6 to CA hexamer with IC₅₀ values of 92.4 nM and 33.4 nM, respectively, which are similar to PF-74 (IC₅₀ = 26.6 nM). This result indicated that these compounds could also inhibit viral replication by interfering with CA-host factor interactions.

Figure 3.

CPSF6 SPR-based competition with d13, d19 and PF-74.

2.5. Determination of the action stage

It was reported that PF-74 displayed antiretroviral potency against HIV-1 variant in both early and late stages of viral replication.^{37, 38} To determine the mechanism of action of representative compounds d13 and d19, we utilized single-round infection assay with modifications to monitor early and late-stage effects, ²¹ and the results are summarized in Table 3. At 1 μM concentration, PF-74 could thoroughly block the viral replication in the early stage. However, the inhibition of PF-74 decreased significantly in the late stage. This result indicated that PF-74 preferentially works in the early stage, as previously reported.³⁹ Similarly, d13 and d19 also possessed a dual-stage inhibition profile and exerted a better inhibitory potency in the early stage.

Table 3.

The infection of HIV-1 NL4.3 virus in early and late stages.

Compds	Infection (%)a
	Early-stage	Late-stage
d13 (5 μM)	0.09 ± 0.01	11.20 ± 8.19
d19 (5 μM)	0.06 ± 0.01	6.82 ± 4.73
PF-74 (1 μM)	0.07 ± 0.01	58.19 ± 12.54

^aInfections are an average of 3 replicates.

2.6. Molecular docking and molecular dynamic simulation (MDs) analysis

In order to research latent binding of the novel CA inhibitors with NTD-CTD interface of HIV-1 CA hexamer (PDB code: 5TSX, 5HGL), the most active compounds d19 and d13 were analyzed for molecular docking and the subsequent MDs studies, with PF-74 as the control.

The binding and conformational stability of three complexes, d19, d13 and PF-74, with capsid protein is a significant factor in supporting their inhibitory action against HIV-1 infection. As shown in Table 4, compound d19 (−5.595 kcal/mol, −5.595 kcal/mol and −63.98 kcal/mol) and d13 (−5.972 kcal/mol, −6.398 kcal/mol and −72.26 kcal/mol) showed higher docking score, XP Gscore, and the binding free energy values over the control, PF-74 (−5.068 kcal/mol, −5.068 kcal/mol and −57.72 kcal/mol). These results also correlated with anti-HIV-1 activity where EC₅₀ of d13, d19 and PF-74 were measured as 0.59 ± 0.08 μM, 0.59 ± 0.06 μM and 0.79 ± 0.09 μM, respectively. All d13, d19 and PF-74 have H-bond interactions with common binding site residues (i.e., Asn57, Gln63) of capsid protein, as depicted in Figure 4.

Table 4.

Molecular docking results

Compounds	Docking score (kcal/mol)	XP Gscore (kcal/mol)	ΔGo (kcal/mol)
d19	−5.595	−5.595	−63.98
d13	−5.972	−6.398	−72.26
PF-74	−5.068	−5.068	−57.72

Figure 4.

Binding mode and interaction profile of the (A) d19-CA complex, (B) d13-CA complex, and (C) PF-74-CA complex (PDB code: 5TSX). H-bond shown in purple color. Residues color such as green, violet, and blue represents hydrophobic, positive charged, and polar residue, respectively.

Subsequently, an extensive MD simulation (100 ns) study was performed to validate our docking results. Figures 5–7 clearly supported the stability of compounds d19, d13 and PF-74 in the binding pocket of CA (PDB code: 5TSX). In d19-5TSX complex, the protein structure became stable after 20 ns, and the average root mean square deviation (RMSD) of Cα was 10.2 Å (Figure 5a). The ligand RMSD for d19 attained stability nearly at 15 ns where abrupt changes in RMSD at 55–65 ns and 90–100 ns are mainly due to the indole moiety (Figure 5a). The average ligand RMSD for d19 was found as 2.63 ± 0.64 Å. It is also supported by a high ligand RMSF of d19 (Figure 5b). Compound d19 interacted with the binding site residues by H-bond (green), hydrophobic interaction (grey), and water bridge (blue), where Asn53, Asn57, and Lys70 interacted with d19 via H-bonds (Figures 5c, d). The Leu56, Met66, and Lys70 interacted with d19 by hydrophobic interaction. As expected, the Gly106, Thr107, Ser 109, and Gln114 interacted with d19 by water bridge interaction (Figures 5c, d). During the MD simulation, d19 remained bounded with 8–10 contacts as depicted in Figure 5e.

Figure 5.

Molecular dynamics and simulation of d19 with CA protein (PDB code: 5TSX) complex. (a) RMSD plot for Cα of CA protein in complex with d19. (b) Ligand RMSF plot of d19 in complex with capsid protein. (c) A Histogram plot shows residues interacting with d19 where H-bonds, hydrophobic interactions, and water bridges are shown in green, grey, and blue, respectively. (d) Interaction distribution (in %) during MD simulation of the d19-CA complex. The panels show the total number of specific contacts of the d19-CA complex throughout the simulation. (e) A timeline representation of the interactions and total contacts (H-bonds, hydrophobic interactions, and water bridges) obtained during the molecular dynamics simulations.

Figure 7.

Molecular dynamics and simulation of PF-74 with CA protein (PDB code: 5TSX) complex. (a) RMSD plot for Cα of 5TSX protein in complex with PF-74. (b) Ligand RMSF plot of PF-74 in complex with capsid protein. (c) A Histogram plot showing residues interacting with PF-74. (d) The percentage of interactions in molecular dynamic simulations of PF-74 complexed with 5TSX. The panels show the total number of specific contacts of the PF-74-5TSX complex throughout the simulation. (e) A timeline representation of the interactions and total contacts (H-bonds, hydrophobic interactions, and water bridges) obtained during the molecular dynamics simulations.

In the d13-CA (PDB code: 5TSX) complex, the protein structure became stable after 10 ns, and the average root mean square deviation (RMSD) of Cα was 3.61 Å (Figure 6a). The ligand RMSD for d13 attained stability nearly at 10 ns where RMSD change in upto ~3.5 ns was mainly due to N-propylmorpholinyl moiety (Figure 6a), which is also supported by a high RMSF of the ligand, d13 (Figure 6b). The average ligand RMSD for d13 was found as 3.59 ± 0.41 Å. Compound d13 interacted with the binding site residues by H-bond (green), hydrophobic interaction (grey), and water bridge (blue), where Asn53, Asn57, and Lys70 interacted with d13 through H-bonds (Figures 6c, d). The Leu56, Met66, and Lys70 interacted with d13 by hydrophobic interactions. As expected, the Gly106, Thr107, and Ser109 interacted with d13 by water bridge interactions (Figure 6c, d). During the MD simulation, d13 is always bounded with 8–10 contacts (Figure 6e). The trajectory analysis revealed that both oxygen and nitrogen atoms of N-propylmorpholinyl moiety interacted to the water molecules during simulation (Supplementary Figure S1). Other side, the methyl cyano moiety of d19 interacted to Asn53 and Gln114 through water bridge interaction (Figure 5d). The high protein RMSD of d19-HIV-1 CA hexamer complex (Figure 5a) was mainly due to residue index from 150 to 219, which was also supported by the protein RMSF (Supplementary Figure S2a), which was significantly lower in d13- HIV-1 CA hexamer complex (Figure 6a; supplementary Figure S2b). Both compounds d13 and d19 interacted to the similar residues during simulation (Figures 5c,d and 6c,d).

Figure 6.

Molecular dynamics and simulation of d13 with CA protein (PDB code: 5TSX) complex. (a) RMSD plot for Cα of CA protein in complex with d13. (b) Ligand RMSF plot of d13 in complex with capsid protein. (c) A Histogram plot shows residues interacting with d13 where H-bonds, hydrophobic interactions, and water bridges are shown in green, grey, and blue, respectively. (d) Interaction distribution (in %) during MD simulation of the d13-CA complex. The panels show the total number of specific contacts of the d13-CA complex throughout the simulation. (e) A timeline representation of the interactions and total contacts (H-bonds, hydrophobic interactions, and water bridges) obtained during the molecular dynamics simulations.

In the PF-74-CA (PDB code: 5TSX) complex, protein structure obtained stability at 40 ns of the simulation with an average Cα-RMSD of 5.7 Å (Figure 7a). The average ligand RMSD for PF-74 (Figure 7a) was of 3.4 ± 0.8 Å. The change in ligand RMSD primarily due to the indole moiety is also supported by the ligand RMSF plot (Figure 7b). Even d19 had lower ligand RMSD than PF-74, supporting the affinity and stability of d19 over PF-74. PF-74 interacted with the residues, i.e., Asn53, Asn57, and Lys70, by H-bonds and hydrophobic interactions with residues, i.e., Leu56, Met66, Lys70, and Ile73 (Figures 7c,d). During the MD simulation, PF-74 is always bounded with 6–7 contacts (Figure 7e). There is no role of salt bridge interaction for the stability of compounds d19, d13 and PF-74 in the binding pocket (Figures 5c, 6c, and 7c).

Next, the Ramachandran mapping of 5TSX residues revealed a very acceptable number of residues in favored, additional allowed, and generously allowed regions (Table 5). There were no capsid protein residues in complex with d13, d19 and PF-74 in the disallowed region (Figure 8).

Table 5.

Occurrence of residues in favored, additional-allowed, generously-allowed, and disallowed regions.

Complex	Residues in favored region	Residues in the additional allowed region	Residues in the generously allowed region	Residues in the disallowed region
d19–5TSX	88.8% (166)	10.2% (19)	1.1% (2)	0.0% (0)
d13–5TSX	86.6% (162)	10.7% (20)	2.7% (5)	0.0% (0)
PF-74–5TSX	85.6% (160)	14.4% (27)	0.0% (0)	0.0% (0)

Figure 8.

The Ramachandran plot (a) d19-CA (PDB code: 5TSX) complex; (b) d13-CA (PDB code: 5TSX) complex; (c) PF-74-CA (PDB code: 5TSX) complex.

2.7. Metabolic stability in human liver microsomes.

In order to measure the metabolic stability profiles of the newly synthesized compounds, d13 and d19 were tested in human liver microsomes (HLM). Meanwhile, PF-74, testosterone, diclofenac, and propafenone were selected as controls. As shown in Table 6, d13 and d19 displayed undesirable metabolic stability with the T_1/2 values of 0.7 and 0.9 min, respectively. Although the metabolic stability of d19 is slightly improved compared with PF-74 (T_1/2 = 0.7 min), further optimization is still needed to alleviate this problem.

Table 6.

Metabolic stability of d13, d19, and PF-74 in human liver microsomes.

Sample	HLM (Final concentration of 0.5 mg protein/mL)
	R2 a	T1/2 (min)b	CLint(mic) (μL/min/mg)c	CLint(liver) (mL/min/kg)d	Remaining (T = 60min)	Remaining (NCFe = 60min)
d13	1.0000	0.7	1853.8	1668.5	0.0%	102.4%
d19	1.0000	0.9	1485.4	1336.9	0.0%	108.3%
PF-74	1.0000	0.7	2114.6	1903.1	0.0%	106.1%
Testosterone	0.9791	13.2	104.8	94.3	4.8%	95.8%
Diclofenac	0.9995	4.5	310.7	279.7	0.0%	86.4%
Propafenone	0.9316	6.3	219.1	197.2	0.0%	103.5%

^aR²: correlation coefficient of the linear regression for the determination of kinetic constant (see raw data worksheet).

^bT_1/2: half-life.

^cCLint_(mic): 0.693/T_1/2/mg microsome protein per mL.

^dCLint_(liver) = CLint_(mic) × mg microsomal protein/g liver weight × g liver weight/kg body weight.

^eNCF: abbreviation of no co-factor. No NADPH is added to NCF samples (replaced by buffer) during the 60-minute incubation. If the NCF remaining is less than 60%, then possibly non-NADPH dependent.

3. Conclusion

Within this study, we designed and synthesized a series of novel 1,2,4-triazole phenylalanine derivatives based on the original PF-74 chemotype with the intention to target an unexplored region within the HIV-1 CA interprotomer pocket composed of aa106–109. The anti-HIV assay results indicated that numerous derivatives displayed excellent antiretroviral efficacy against HIV-1 and HIV-2 strains. Notably, compound d19 showed slightly better anti-HIV-1/2 potency than PF-74, with relatively low cytotoxicity. Furthermore, SPR binding experiments indicated that d19 prefers CA hexamer over monomer binding compared to PF-74. Nevertheless, d19 can inhibit the replication of HIV-1 NL4.3 virus in both early and late stages, with a preference for early-stage (CA hexamer) inhibition.

Additionally, d19 competes with CPSF6 in our SPR-based competition experiments, indicating a shared binding site with CPSF6 within the interprotomer pocket, as confirmed by our docking and MD simulation. Utilizing molecular docking and MD simulations, we could also confirm the experimentally derived higher affinity for the HIV-1 CA hexamer compared to PF-74.

At this stage, d19 improvement in metabolic stability compared to PF-74 is minimal; even so, d19 represents a new chemotype and, therefore, the potential for further modifications resulting in anti-HIV potency and metabolic stability enhancements.

Data availability statement

The data that support the findings of this study are available from thecorresponding author upon reasonable request.