Truncated CNP1 proteins demonstrate anti-HIV-1 activity

To quantify the abundance of CNP variants in monocytes and CD4⁺ T cells, the primary targets of HIV-1, we amplified mRNA variants of CNPs using designed upstream primers located within their respective exon 1 and a common downstream primer situated in exon 2 (Fig S1). In resting human monocytes and CD4⁺ T cells, the mRNA levels of the CNP2 variant were ∼200 times higher than those of CNP1. After treatment with IFN-α and pseudotyped HIV-1 particles, both CNP1 and CNP2 transcript levels increased in these cells, with CNP2 mRNA levels remained 40–90 times higher than those of CNP1 (Fig 1A and B). Despite the higher levels of CNP2 mRNA, Western blot analysis revealed that CNP1 was the predominant CNP protein isoform in these cells, irrespective of IFN-α or pseudotyped HIV-1 particle treatment (Fig 1C).

To assess the antiviral potency of both CNP isoforms, we transfected 293T cells with CNP1 or CNP2 mammalian expression plasmids along with HIV-1-based lentiviral pseudotyped particle assemble vectors. Subsequently, we evaluated the levels of viral particulate capsid (p24) in purified supernatants, and the intracellular levels of CNP, p24, and Gag precursor p55 in transfected 293T cells using Western blot. Simultaneously, the infectious virion yield was monitored by luciferase activity in 293T cells at 48 h postinfection. As depicted in Fig 1E, luciferase assays demonstrated that both CNP1 and CNP2 plasmids resulted in more than a 1,000-fold reduction in luciferase activity. Both CNP1 and CNP2 plasmids significantly reduced p24 levels released in the culture media, accompanied by an increase in intracellular levels of p55 gag precursors in transfected 293T cells (Fig 1D). These findings suggest that the short protein isoform CNP1 can inhibit HIV-1 virus assembly.

Posttranslational processing of CNP2 N20aa is prerequisite for antiviral capacity

CNP1 proteins could be translated from an alternative start site in CNP2 mRNA or generated by cleaving N20aa of CNP2 proteins (O’Neill et al, 1997; Lee et al, 2006). We further investigated the effects of translation and posttranslational processing on the antiviral activity of CNPs. To selectively impede the alternative translation of CNP1 proteins from CNP2 mRNA, we generated a CNP2-M21L mutant (Fig S2A). Compared with the CNP2 construct, the CNP2-M21L mutant reduced the expression levels of CNP1 protein and impeded the ability to block HIV-1 particle assembly (Fig 1F and G).

Next, we explored whether posttranslational processing of CNP2 proteins was crucial for antiviral activity. To inhibit the cleavage of N20aa, we constructed a Flag-CNP2 construct in which a Flag tag sequence was added to the N-terminal of CNP2 cDNA, and similarly, a Flag-CNP1 construct. We transfected the plasmids encoding CNP1, CNP2, Flag-CNP1 or Flag-CNP2 cDNA into 293T cells along with HIV-1 pseudotype particle assemble vectors. The Flag-CNP2 plasmids transfected into 293T cells exhibited reduced levels of CNP1 proteins and increased levels of CNP2 (Fig 1D), suggesting that the process forming truncated CNP1 from CNP2 was impaired. Meanwhile, in comparison with CNP2 constructs, Flag-CNP2 constructs exhibited reduced antiviral capacity (Fig 1D and E). As a control, Flag-CNP1 and CNP1 constructs showed comparable antiviral capacity (Fig 1D and E). Therefore, the posttranslational processing of CNP2 N20aa to form CNP1 is prerequisite for anti-HIV-1 activity.

Posttranslational processing of CNP2 N20aa occurs in the cytoplasmic matrix

A previous study posited that the N20aa of CNP2 served as a MTS, regulating CNP2 translocation and protein processing within the mitochondria. Consequently, the mitochondrial targeting feature was deemed necessary for the antiviral activity of CNP. To scrutinize this hypothesis, we established stably transduced 293T cell lines expressing CNP1 and CNP2 based on the pLVX-TetOne lentiviral vector (Fig 2A). In both CNP1 and CNP2 stably transduced 293T cells, Western blot analysis revealed the presence of CNP1 proteins in the mitochondrial fractions (Fig 2B). Confocal imaging further demonstrated the co-localization of CNP1 with the mitochondrial marker Tom20 (Fig 2C). To explore the intrinsic property of mitochondrial targeting in CNP1, we deleted 30 aa and 40 aa from the N-terminal of CNP2 (Δ30 CNP2 and Δ40 CNP2, respectively) (Figs 2A and S2A). These mutant proteins also translocated into the mitochondria (Fig 2D), suggesting that CNP1 inherently possesses mitochondrial targeting properties. Therefore, we questioned the assumption that N20aa of CNP2 functioned as MTS and was cleaved in the mitochondria.

It has been reported that phosphorylation at amino acids Ser9 and Ser22 hindered the cleavage process of CNP2 and its targeting to mitochondria. Therefore, we generated stably transduced 293T cell lines with a CNP2-S9/22A mutant construct in which phosphorylation of Ser9 and Ser22 was blocked, or with a CNP2-S9/22D mutant to mimic CNP2 phosphorylation at Ser9 and Ser22 (Fig S2A). Contrary to the previous report, Western blot analysis revealed the presence of CNP2 and CNP1 proteins in the mitochondrial fraction in 293T cells stably transduced with these mutants (Fig S2B). Confocal imaging also showed the colocalization of CNP signals with Tom20 signals (Fig S2C).

Next, we sought an alternative approach to assess whether the N20aa of CNP2 functioned as an MTS. To achieve this, we engineered CNP2¹⁻²⁰-GFP, CNP2¹⁻³⁰-GFP, and CNP2¹⁻⁴⁰-GFP mutants encoding fusion proteins of N-terminal 20, 30 or 40 amino acids of CNP2 fused with GFP as previous research (Fig S2D) (Backes et al, 2018). We constructed a COX8A-MTS-GFP mutant encoding a fusion protein of human COX8A MTS and GFP as a positive control (Figs S2D and 2E) (Van Kuilenburg et al, 1988). Western blot analysis and confocal imaging confirmed that COX8A MTS directed GFP translocation into the mitochondria (Fig 2F and G). In contrast, GFP signals from CNP2¹⁻²⁰-GFP, CNP2¹⁻³⁰-GFP, and CNP2¹⁻⁴⁰-GFP were found in the cytosolic fraction but not in the mitochondrial fractions (Fig 2F). Confocal microscopy also revealed that GFP signals did not co-locate with the mitochondria marker in CNP2¹⁻²⁰-GFP, CNP2¹⁻³⁰-GFP or CNP2¹⁻⁴⁰-GFP stably transduced 293T cells (Fig 2G). Therefore, the N20aa of CNP2 did not function as MTS to mediate GFP translocation into the mitochondria.

We observed that the molecular weight of CNP2¹⁻²⁰-GFP was comparable with that of GFP (Fig 2E and F), GFP signals of CNP2¹⁻²⁰-GFP, CNP2¹⁻³⁰-GFP, and CNP2¹⁻⁴⁰-GFP was found in the cytosolic fraction (Fig 2F), suggesting that the N20aa of CNP2¹⁻²⁰-GFP, CNP2¹⁻³⁰-GFP, and CNP2¹⁻⁴⁰-GFP was cleaved outside the mitochondria. In addition, GFP signals were not observed in the endoplasmic reticulum, trans-Golgi, cis-Golgi, cytoskeleton, and lysosomes in CNP2¹⁻²⁰-GFP stably transduced 293T cells (Fig S3A). When we used colchicine to block the assembly of microtubule proteins or Brefeldin A to disrupt the transport of proteins from the endoplasmic reticulum to the Golgi apparatus, protein processing of CNP2 and CNP2¹⁻²⁰ GFP was not affected by colchicine or Brefeldin A (Fig S3B and C). These data suggest that the assembly of cytoskeletal components and the transport of proteins from the endoplasmic reticulum to the Golgi apparatus are not involved in the processing N20aa of CNP2. Moreover, by preventing the cleavage of N20aa through the addition of the Flag tag, Flag-CNP2 retains localization to the cell membrane and mitochondria (Fig S3D and E), and no localization to the endoplasmic reticulum (Fig S3F). Taken together, posttranslational processing of CNP2 N20aa might occur in the cytoplasmic matrix.

CNP1 interacts with HIV-1 Gag proteins on the cell membrane

Previous reports indicate that CNP blocks HIV-1 particle assembly through Gag-CNP interaction at budding sites on the plasma membrane (Wilson et al, 2012). To further investigate CNP1–Gag interaction, we generated 293T cell lines with stable expression of HIV-1 Gag proteins and dox-induced expression of CNP variants (Figs 3A and S4A). High-resolution 3D reconstruction confocal imaging confirmed the distribution of CNP1 protein on the cell membrane and in the mitochondria in the absence of HIV-1 Gag (Fig 3B and Videos 1 and 2). In the absence of CNP expression, HIV-1 Gag colocalized with the cell membrane protein CD81 rather than with the mitochondria protein Tom 20 (Fig 3C and Videos 3 and 4). Upon induction of CNP1 expression by dox, CNP1 protein co-localized with Gag protein on the cell membrane protein CD81 (Fig 3D and E, Videos 5 and 6), but not with Tom 20 (Fig S4B). Subsequently, we constructed a CNP2 D72E mutant to abolished CNP binding to Gag (Fig S4A) (Wilson et al, 2012). This mutant localized to the cell membrane (Fig S4D), but lost its anti-lentiviral assembling potency (Fig 3F and G). Thus, the primary interaction between CNP1 and HIV-1 Gag occurs on the cell membrane.

To further confirm this notion, we engineered a CNP2-C418A mutant (CNP2-CAAXM) to interfere with prenylation modifications of the CAAX box at the C-terminus and subsequently impede CNP1 protein translocation to the plasma membrane (Fig S4A). Consistent with the previous studies, Western blot showed that CNP1 proteins were enriched in the membrane fractions, whereas CNP2-C418A mutants were mainly in the cytosolic fractions (Fig S4E). High-resolution 3D reconstruction confocal imaging showed that the C418A mutant remained in the mitochondria (Fig 3H) and failed to co-localize with Gag protein on the cell membrane (Figs 3I and S4C, Videos 7 and 8). Consistent with these alterations, CNP2-C418A mutant showed reduced anti-lentiviral assembling potency (Fig 3F and G).

N20aa may hinder the interaction between CNP2 and HIV-1 Gag

We observed that cleavage N20aa of FLAG-CNP2 protein was impaired by a Flag tag (Fig 1D). However, Flag-CNP2 protein still localized on the cell membrane (Fig S3D) but displayed a reduced anti-lentiviral assembling potency (Fig 1D and E). We speculated that N20aa segment may hinder the interaction between CNP2 and Gag. To test this hypothesis, we employed molecular docking calculations using the MOE software to analyze the binding interactions between CNP and Gag proteins. Protein-protein docking analysis revealed a higher binding energy for CNP2-Gag docking compared with CNP1-Gag (Fig 4A), and N20aa segment increased the spatial distance between of CNP2 protein and Gag protein (Fig 4B).

To validate the docking calculations of CNP2-Gag protein interactions, we performed a fluorescence complementation assay in 293T cells as described previously (Kamiyama et al, 2016; Zhang et al, 2020). We co-expressed split super-folder GFP11-CNP1 (sfGFP11-CNP1) with Gag containing a split super-folder GFP (1–10) tag (Gag-sfGFP10) (Figs 4C and S5A). Western blot assay demonstrated that sfGFP11 at CNP2 N-terminus prevents the cleavage of the N20aa (Fig S5B). Confocal microscopy assay confirmed that sfGFP11-CNP2, sfGFP11-CNP1 and sfGFP11-CNP1-D72E still localized on the cell membrane (Fig S5C). The GFP levels in sfGFP11-CNP2 decreased by ∼30% compared with sfGFP11-CNP1 and were comparable with sfGFP11-CNP1-D72E (Fig 4D and E). These results indicate that N20aa hinders the interaction between CNP2 and Gag proteins on the cell membrane.

Cell lines and cell culture

293T cells were cultured in DMEM supplemented with 10% FBS and 1% penicillin and streptomycin (PS). Human CD4⁺ T cells and monocytes were cultured in Roswell Park Memorial Institute 1640 Medium (RPMI-1640) supplemented with 10% heat-inactivated FBS and 1% PS. All cells were gifts from members of the Beijing Key Laboratory of Emerging infectious Diseases.

Western blot

For Western blot analysis, cells were lysed in lysis buffer (20 mM Tris–HCl, 150 mM NaCl, 1 mM EDTA, 1% TritonX-100, and protease inhibitors, pH 7.5) at 4°C for 30 min. Equal amounts of cell lysate (25 μg) were separated on 4–12% or 10% acrylamide gels (Thermo Fisher Scientific) and transferred onto 0.22 μm PVDF membranes (Millipore). After blocking with blocking buffer (TBST supplemented with 5% nonfat dry milk), membranes were probed with specific antibodies: mouse anti-CNP (ab6319, 1:2,000; Abcam), goat anti-HIV1 p24 (ab53841, 1:2,000; Abcam), rabbit anti-Tom20 (ab186735, 1:1,000; Abcam), rabbit anti-GAPDH (5174S, 1:1,000; Cell Signaling Technology), rabbit anti-Flag (14793S, 1:1,000; Cell Signaling Technology), and rabbit anti-β-tubulin (ab6046, 1:1,000; Abcam). Subsequently, blots were probed with species-specific IRdye secondary antibodies, and visualization was performed using an Odyssey infrared imaging system (Li-COR).

CNP transcript variants quantification by real-time RT–PCR

To quantify the relative abundance of CNP transcript variants, real-time RT–PCR was employed. Primers CNP1-F (to amplify CNP1 mRNA, GenBank accession number: NM_001330216.2): GGAGAGCTTCAGACAAGCTTCC, CNP2-F (to amplify CNP2 mRNA, GenBank accession number: NM_033133.4), CTCCGCGCAGGCGGGCGGCC, and commonly used downstream CNP-R, AGAGCGTCTTGCACTCTAGC were used to discriminate between different transcript variants. GAPDH-F, GAATGGGCAGCCGTTAGGAA, and GAPDH-R, GATCTCGCTCCTGGAAGATG were used to amplify GAPDH as a control (Fig 1A). RNA and protein were isolated from healthy donor human monocytes and CD4⁺ T cells before and after stimulation with IFN-α (500 IU/ml) or pseudo-typed HIV-1 particles for 24 or 48 h. The relative amount of mRNA before and after IFN-α treatment was semi-quantified by real-time PCR. CNP1 and CNP2 proteins before and after IFN-α treatment were determined by Western blot.

Construction of CNP isoforms and mutants

The CNP2 cDNA fragment was amplified from the cDNA pool of reverse-transcribed total RNA from CD4⁺ T cells and cloned into pcDNA5. PCR-based mutagenesis was employed to construct CNP1 and various motif mutants. CNP2-S9/22A, with phosphorylation sites at Ser9 and Ser22 mutated to non-phosphorylated Ala, and CNP2-S9/22D, with Ser replaced by phosphomimic Asp, were constructed. CNP2-M21L, where Met21 was mutated to Leu to abolish CNP1 translation by direct translation at the second ATG, was also generated. The C-terminal CAAX box of the prenylation site CTII was mutated to ATII based on CNP2 to construct CNP2-CAAXM. CNP2-D72E, with aspartic acid at position 72 involved in CNP-Gag interaction mutated to Glu, was created. C-terminal CAAX box of prenylation site CTII was mutated to ATII to construct CNP2-CAAXM.

Antiviral potency test of different CNP mutants

The antiviral potency of these mutants was assessed by measuring pseudotyped viral particle production and subsequent luciferase assay. 293T cells were transfected with 1 μg CNP mutant constructs along with the third pseudotyped HIV-1 particle assembly system, which included four plasmids named pLenti6-Luc (1.33 μg), pMDLg/pRRE (0.67 μg), pRSV-Rev (0.5 μg), and pLP/VSVG (0.5 μg) in a six-well plate using TurboFect Transfection Reagent (Thermo Fisher Scientific). 48 h post-transfection, the virion-containing supernatant (1,000 μl) was layered onto 400 μl of 20% sucrose in PBS, and viral particles in the supernatant were collected by centrifuging at 20,000g for 120 min at 4°C. CNP, p55 Gag precursor, and p24 in the viral particle and transfected 293T cell were also tested by Western blot.

Pseudotyped HIV-1 production in the supernatant was determined by Luciferase assay. 293T cells were seeded in a 96-well black-walled plate 24 h in advance and infected with virion-containing supernatant (100 μl/well) for 4–6 h and then replaced with DMEM complete medium (200 μl/well). After 48 h of infection, luciferase activities were monitored by luminometry in a TopCounter (Perkin Elmer). A two-tailed unpaired t test was used to analyze differences between CNPs and control using GraphPad Prism 9 software.

Generation of HIV-1 Gag stable expression cell lines

The native codon Gag in pMDLg/pRRE vector was optimized and cloned into pCMV3 (pCMV3-synGag). 293T cells expressing HIV-1 Gag (293T Gag⁺) were generated by transduction with pCMV3-synGag vectors, hygromycin selection, and maintenance of single-cell clones in DMEM supplemented with hygromycin. This cell line was routinely tested for the subcellular localization of Gag.

Generation of doxycycline-inducible expression system in 293T cells

To establish a doxycycline-inducible expression system in 293T cells, the CNP2 WT and mutant cDNA fragments were amplified by PCR using primers derived from the aforementioned CNP cDNA in pcDNA5. Subsequently, the PCR products were cloned into the transfer plasmid pLVX-TetOne (Takara), which was digested with BamHI and EcoRI restriction enzymes. This process resulted in the creation of DNA constructs, including Dox-CNP2, Dox-CNP2 S9/22A, Dox-CNP2-S9/22D, Dox-CNP2-M21L, Dox-CNP1, Dox-CNP2 D72E, Dox-CNP2-CAAXM, and Dox-CNP2-CAAXM, using the One Step Cloning Kit (Vazyme Biotechnology). Truncated CNP mutants, named CNP2 Δ30 and CNP2 Δ40, lacking various lengths at the N-terminus, were also constructed into the pLVX-TetOne vector using the same protocol. In addition, the sequence encoding the human COX8A mitochondrial targeting signal fused with GFP protein (COX8A-MTS-GFP) was constructed into the pLVX-TetOne vector. This was achieved by amplifying the human COX8A MTS sequence, the first 25 amino acids at the N-terminus, from the cDNA pool of reverse-transcribed total RNA from CD4⁺ T cells. This sequence was then fused to the N-terminal of the GFP encoding sequence by PCR to generate the COX8A-MTS-GFP sequence. Similarly, CNP N-terminal fusions with GFP protein, named CNP^1–20 GFP, CNP^1–30 GFP, and CNP^1–40 GFP, were also constructed into the pLVX-TetOne vector using the same protocol.

To produce lentivirus particles, 5 × 10⁵ 293T cells were seeded into a six-well plate 24 h before transfection. Subsequently, the cells were transfected with 2 μg of the abovementioned doxycycline-inducible expressing transfer plasmids, along with 1 μg pMDLg/pRRE, 0.5 μg pRSV-Rev, and 0.5 μg pLP/VSVG, using TurboFect Transfection Reagent (Thermo Fisher Scientific). After 8 h of transfection, the cells were replenished with 6 ml DMEM complete medium. Supernatants were harvested after 48 h of transfection. Cell-free supernatants were obtained by centrifugation for 10 min at 5,000g, and lentivirus particles were concentrated by centrifugation through a 20% sucrose cushion at 20,000g for 120 min at 4°C. Virus particles were resuspended in DMEM medium.

To generate doxycycline-inducible expressing 293T cells, these viruses were used to infect 293T cells or Gag stable expressing 293T cells. Stable transduced cell lines were screened in DMEM complete medium complemented with 1 μg/ml puromycin (InvivoGen). To test the doxycycline-inducible expression of the proteins of interest, the cell lines were seeded into a six-well plate to adhere overnight and the DMEM medium was replaced with medium supplemented with50 ng/ml doxycycline (Takara) for 12 h. The induced cells were then examined by Western blot using an anti-CNP antibody (Abcam) or anti-GFP antibody (Beyotime Biotechnology).

Separation of mitochondrial and cytosolic components

Doxycycline-inducible expressing 293T cell lines were seeded into a six-well plate and supplemented with 50 ng/ml doxycycline (Takara). 12 h post-induction, mitochondrial and cytosolic fractions were separated by differential centrifugation using the Cell Mitochondria Isolation Kit (Beyotime Biotechnology). In brief, induced cells were trypsinized, washed with cold PBS, and homogenized in homogenizing buffer. The homogenate was spun at 600g for 10 min, and mitochondria in the supernate were pelleted at 11,000g for 10 min, and then resuspended in mitochondria storage buffer. The homogenate without mitochondria was centrifuged at 12,000g for 10 min, and the supernate was labeled as the cytosolic fraction. Protein concentration was determined by the BCA Protein Assay Kit (Thermo Fisher Scientific). Both mitochondria and cytosolic fractions were subjected to Western blot. An equal relative amount of mitochondria and cytosolic protein was loaded to maintain the original mitochondria/cytosolic ratio within the transfected cell between different transfections.

Confocal microscopy assay

Doxycycline-inducible expressing 293T cell lines were seeded on slide covers and induced with 50 ng/ml doxycycline for 12 h. The induced cells were fixed with 1% PFA in PBS for 30 min at 4°C and permeabilized with 0.1% Triton X-100 in PBS for 5 min at RT. CNP or GFP distribution was detected by immune-fluorescence staining and confocal imaging. CNP was detected using mouse anti-CNP (ab6319, 1:500; Abcam) and Alexa Fluor 488-conjugated goat anti-mouse antibody (ab150113, 1:1,000; Abcam); Mitochondrial marker Tom20 was detected using rabbit anti-Tom20 (ab186735, 1:500; Abcam) and Alexa Fluor 555-conjugated goat anti-rabbit (ab150114, 1:1,000; Abcam); nuclei were stained with DAPI (ab285390, 1:1,000; Abcam). The confocal microscopy assay was performed with the confocal laser scanning microscopy platform Zeiss LSM 880 with Airyscan under a ×100 objective, and pictures were analyzed by ZEN blue 3.2 (Zeiss) software.

Multiplex tyramide signal amplification (TSA)

Doxycycline-inducible 293T cells expressing stable HIV-1 Gag protein were seeded on slide covers and induced with 50 ng/ml doxycycline for 12 h. After fixation with 1% PFA in PBS for 30 min at 4°C and permeabilization with 0.1% Triton X-100 in PBS for 5 min at RT, multiplex fluorescence labeling was executed using TSA-dendron-fluorophores (NEON 5-color All-round Discovery Kit for ICC, NECC550; Histova Biotechnology). Briefly, endogenous peroxidase was quenched with 3% H₂O₂ for 30 min at RT, followed by treatment with a blocking reagent for 30 min at 37°C in a humidified chamber. Primary antibodies were incubated for 90 min in a humidified chamber at 37°C, followed by washing with 0.1% Tween-20 in PBS three times. HRP-conjugated secondary antibodies were then incubated for 1 h in a humidified chamber at 37°C and subsequently washed three times with 0.1% Tween-20 in PBS. TSA visualization was performed in a reaction buffer for 30–60 s. The primary and secondary antibodies were thoroughly eliminated by Abcracker buffer for 30 min in a humidified chamber at 37°C. Each antigen was labeled with distinct fluorophores in a serial fashion. The multiplex antibody panels applied in this study included rabbit anti-CD81 (ab244297, 1:250; Abcam) and GAR-HRP (PV-6001; Zsbio) labeled by NEON420 (NECC550, 1:500; Histova Biotechnology); mouse anti-CNP (ab6319, 1:500; Abcam) and GAM-HRP (PV-6002; Zsbio) labeled by NEON670 (NECC550, 1:500; Histova Biotechnology); goat anti-HIV1 p24 (ab53841, 1:500; Abcam) and RAG-HRP (PV-9003; Zsbio) labeled by NEON520 (NECC550, 1:500; Histova Biotechnology); rabbit anti-Tom20 (ab186735, 1:500; Abcam) and GAR-HRP (PV-6001; Zsbio) labeled by NEON600 (NECC550, 1:500; Histova Biotechnology). After sequential detection of all antibodies, the cells were stained with SN470 (NECC550, 1:500; Histova Biotechnology) to label the nucleus at RT for 10 min. For 3D super-resolution cell imaging, multiplex TSA-stained cell slides were imaged using the confocal laser scanning microscopy platform Zeiss LSM 880 with Airyscan under a ×100 objective. Reconstruction and image analysis of the 3D images were performed using ZEN blue 3.2 software (Zeiss).

Extraction of transmembrane proteins and cytosolic protein fraction

Transmembrane proteins were extracted using the ProteoExtract Transmembrane protein extraction kit (TB477; Novagen). Doxycycline-inducible CNP-expressing 293T cell lines, capable of stable expression of HIV-1 Gag protein, were seeded in six-well plates and induced with 50 ng/ml doxycycline for 12 h. A total of 0.1 ml Extraction Buffer 2A was prepared by mixing 50 μl Extraction Buffer 2 and 50 μl TM-PEK Reagent A. The 293T cells were washed twice with cold PBS, transferred to a 1.5-ml conical tube, and collected by centrifugation at 800g for 5 min at 4°C. Cells were resuspended in 0.1 ml Extraction Buffer 1, and 5 µl Protease Inhibitor Cocktail Set III was added. After incubation for 10 min at 4°C with gentle agitation, the cells were collected by centrifugation at 1,000g for 5 min at 4°C. The supernatant was carefully removed and referred to as the cytosolic protein fraction. The pellet was resuspended in 0.1 ml Extraction Buffer 2A, and 5 µl of Protease Inhibitor Cocktail Set III was added. After incubation for 1 h at 4°C with gentle agitation, the mixture was centrifuged at 16,000g for 15 min at 4°C. The supernatant, enriched in integral membrane proteins, was collected.