Generation of bioluminescent HIV-1 reporter vectors.
To generate an HIV-1 reporter virus that enables quantification of the reporter protein directly from virus supernatant, we employed the previously reported strategy (4, 5, 7) of introducing the reporter gene between the MA and CA domains of Gag, a location shown to tolerate genetic insertions with minimal effects on Gag protein expression and processing (4). The NanoLuc gene was introduced between MA and CA domains of HIV-1 Gag flanked by PR cleavage sites at the N-terminus or C-terminus, or both termini of NanoLuc or not flanked by any PR cleavage site (Fig. 1A). Upon PR-mediated Gag cleavage, the resulting NanoLuc products generated from these vectors are the NanoLuc protein and MA-NanoLuc, NanoLuc-CA and MA-NanoLuc-CA fusion proteins, respectively (Fig. 1A).
To assess the effect of the NanoLuc insertion into HIV-1 Gag on Gag protein expression and PR-mediated Gag cleavage, we transfected the HIV-1 Gag-NanoLuc vectors into HEK293T cells and after 48hrs, lysed the cells and pelleted virus from the supernatant. We performed western blot analysis to evaluate the expression of HIV-1 Gag-NanoLuc fusion proteins in the cells and the pelleted virus. We observed no adverse effect on Gag expression and processing and, importantly, the expected Gag PR cleavage products were generated, including some cleavage intermediates (Fig. 1B). We also measured NanoLuc activity from the cell lysate and virus supernatant and observed that all the vectors yielded robust NanoLuc activity; i.e., up to 2.0x107 relative light units (RLUs) (Fig. 1C). Finally, we calculated virus release efficiency (VRE) from the same samples and found that release efficiency was modestly reduced (<2 fold) compared to that of WT Gag but similar to the Gag-iGFP construct (Fig. 1D). VRE is calculated as the ratio of virion-associated p24 CA to total Gag (i.e., cellular p24 CA and Pr55Gag and viral p24 CA), normalized to WT VRE, which is set to 100. Because all four Gag-NanoLuc constructs were similar in terms of Gag expression and NanoLuc activity and displayed only modest (<2-fold) differences in VRE, we chose to proceed with the pNL4-3 Gag-iNanoLuc vector for further characterization. Because pNL4-3 Gag-iNanoLuc has PR cleavage sites on both sides of the NanoLuc protein, which allows for complete processing of the Gag-iNanoLuc protein into the individual Gag domains (as is the case for WT pNL4-3), it is the most representative of the native Gag.
HIV-1 Gag-iNanoLuc enables highly sensitive measurement of virus release.
To assess the sensitivity in the detection of virus release using the HIV-1 Gag-iNanoLuc vector, we transfected HEK293T cells with either a NanoLuc expression vector, the WT HIV-1 molecular clone pNL4-3, or decreasing amounts of the HIV-1 Gag-iNanoLuc vector (i.e. 1.0, 0.5, 0.25, 0.125 and 0.0625mg). At 48hrs post-transfection, we lysed the cells and purified virions from the supernatant, analyzed Gag levels by western blot (Fig. 2A), and measured NanoLuc activity from the cell lysates and supernatants (Fig. 2B). We were able to detect NanoLuc signal under all conditions tested, including at the lowest DNA input at which virion-associated Gag was undetectable by western blot. pNL4-3 Gag-iNanoLuc vector-transfected cells produced significantly higher levels of NanoLuc activity in both the cell lysate (>10-fold) and supernatant (>1000-fold) relative to the NanoLuc expression vector control. This implies that the NanoLuc activity in the supernatant is derived from the NanoLuc protein released with the HIV-1 Gag during virus release. We also measured RT activity (Fig. 2C) and p24 protein levels (Fig. 2D) in the virus supernatant and correlated both with supernatant NanoLuc activity (Fig. 2E & F). We observed that the supernatant NanoLuc activity was positively correlated with RT activity and p24 abundance, further reinforcing the specificity of the assay.
The defect in HIV-1 Gag-iNanoLuc particle infectivity can be rescued by co-expression with WT.
We generated virus using either WT pNL4-3, the HIV-1 Gag-iGFP or the HIV-1 Gag-iNanoLuc vectors by transfecting them into HEK293T cells and collecting the supernatants containing the progeny virions at 48hrs post-transfection. We quantified the relative amounts of virus in the supernatant by RT activity (Fig. 3A). We observed that RT activity of supernatants from cells transfected with the HIV-1 Gag-iGFP and HIV-1 Gag-NanoLuc vectors was about 2-fold less than that of supernatants from cells transfected with WT pNL4-3. To test the infectivity of the virions produced from the HIV-1 Gag-NanoLuc vectors, we infected TZM-bl cells with the RT-normalized virus supernatants and measured the infectivity by quantifying the HIV-1 Tat-driven firefly luciferase activity (Fig. 3B). We observed that the HIV-1 Gag-NanoLuc viruses were approximately 10-fold less infectious than the WT virus while the Gag-iGFP virus was only about 2-fold less infectious than WT virus. We also transfected the SupT1 T-cell line with the HIV-1 Gag-NanoLuc vectors and monitored virus replication kinetics over several days and observed that replication was significantly impaired compared to the WT HIV-1 (data not shown). We generated viruses using the pNL4-3 Gag-iNanoLuc vector complemented with different ratios of the WT HIV-1 molecular clone pNL4-3 and tested their infectivity by measuring the HIV-1 Tat-driven firefly luciferase activity in TZM-bl cells. We observed that infectivity of the viruses generated with the pNL4-3 Gag-iNanoLuc vector was rescued when complemented with the WT pNL4-3 vector. The infectivity increased with increasing pNL4-3 to pNL4-3 Gag-iNanoLuc ratio; at ratios above 2:1 the infectivity was at WT HIV-1 levels (Fig. 3C). We also measured NanoLuc activity from the same infected TZM-bl cells and observed that the virus generated with the pNL4-3 Gag-iNanoLuc vector alone was able to express NanoLuc activity in the target cell despite being poorly infectious. The NanoLuc expression increased with increasing ratio of WT pNL4-3 to pNL4-3 Gag-iNanoLuc vector used in generating the Gag-iNanoLuc viruses. However, we observed that with higher ratios of WT pNL4-3 to pNL4-3 Gag-iNanoLuc i.e., 5:1 and 10:1, the NanoLuc expression decreased (Fig. 3D), presumably due to lower amounts of Gag-iNanoLuc genome transduced into target cells. We analyzed purified virus particles generated using the Gag-iNanoLuc proviral vectors either alone or complemented with the WT proviral clone by electron microscopy and we observed that the morphology of Gag-iNanoLuc virus particles was abnormal; specifically, the viral cores displayed a spherical shape as opposed to the canonical cone shape (data not shown). Consistent with the rescue of particle infectivity, normal, conical cores were observed when virus was generated by co-expressing the Gag-iNanoLuc vector with WT pNL4-3 (data not shown). The rescue of virion morphology by coexpressing Gag fusion proteins with WT Gag has also been observed previously (4).
HIV-1 Gag-iNanoLuc provides a robust tool for quantifying virus release.
To examine the utility of the HIV-1 Gag-iNanoLuc vector in functional assays for virus release, we constructed versions of the vector that lacked the PTAP motif in the p6 domain of the HIV-1 Gag protein (HIV-1 Gag-iNanoLuc-PTAP- ) and the vpu gene (HIV-1 Gag-iNanoLuc-delVpu) by cloning the iNanoLuc cassette into previously reported PTAP- and delVpu HIV-1 molecular clones (12, 13). The p6 domain of HIV-1 Gag is required for virus release (13, 14) because of its interaction with the ESCRT machinery (15-18). Vpu is also required for HIV-1 release in the presence of the restriction factor tetherin (also known as BST-2), which blocks release of virions by tethering them to the plasma membrane (19). Vpu counteracts tetherin by mechanisms involving both proteasomal and lysosomal degradation and intracellular sequestration of tetherin (20, 21) We transfected HEK293T cells with the WT, PTAP- and delVpu versions of the HIV-1 Gag-iNanoLuc vector with or without varying amounts of tetherin expression vector. At 48hrs post-transfection, we measured NanoLuc activity in the cell lysates and supernatants. We observed a 2-fold decrease in the NanoLuc activity in the supernatant of cells transfected with the PTAP- vs. the WT vector, but, as expected, no decrease in NanoLuc activity in the cell lysates. Likewise, co-transfection with a tetherin expression vector, but not an empty vector control, caused a 4- to 10-fold decrease in NanoLuc activity in the supernatant of cells transfected with the delVpu vector. The decrease in supernatant NanoLuc activity was proportional to the amount of tetherin vector transfected (Fig.4A and B). We performed western blot analysis of the cell lysates and the pelleted virions to analyze HIV-1 Gag expression. We observed that virus release measured by virion-associated p24 levels corresponded with the NanoLuc activity. Finally, we tested the utility of the HIV-1 Gag-iNanoLuc vector to detect impaired virus release induced by treatment of virus-producer cells with amphotericin B methyl ester (AME), a compound that inhibits HIV-1 particle production (22). We transfected HEK293T cells with the HIV-1 Gag-iNanoLuc vector and at 24hrs post-transfection treated the cells with either vehicle or increasing amounts (5mM or 10mM) of AME. At 24hrs post-treatment, we collected the supernatant and measured NanoLuc activity. We observed a decrease in supernatant but not cell-associated NanoLuc activity in the presence of AME but not vehicle control. Again, the decrease in supernatant NanoLuc activity corresponded with reduced virion-associated p24 measured by western blot analysis. These results demonstrate that the Gag-iNanoLuc vector provides a highly sensitive and quantitative tool for measuring the effects of Gag mutations, host cell restriction factors, and small molecule inhibitors on HIV-1 particle assembly and release.