For functional analyses, we generated untagged and AU1-tagged forms of the wildtype and S52A HIV-1 NL4-3 Vpus and verified their expression by Western blot analysis (Fig. 1A). Down-modulation of CD4 from the cell surface was measured by flow cytometric analysis of Jurkat T cells transiently transfected with vectors co-expressing Vpu and GFP via an internal ribosomal entry site (IRES). Transport of CD4 to the cell surface was measured by co-transfection of 293T cells with CD4 and constructs expressing GFP alone or together with Vpu. Wildtype Vpu caused about 2-fold reduced levels of CD4 expression on Jurkat T cells and efficiently blocked the transport of newly synthesized CD4 to the surface of 293T cells (Fig. 1B, C). In contrast, the S52A Vpu was inactive in both assays (Fig. 1B, C).

It has been shown that Vpu reduces the total levels of cellular tetherin, and it has been suggested that this effect may be important for its capability to promote virus release 9101819. To test whether the S52A change affects tetherin degradation by Vpu, we generated an N-terminally eCFP-tagged version of tetherin. Confocal microscopy showed that the fusion protein had a subcellular localization comparable to endogenous tetherin and inhibited viral particle release (data not shown). Degradation of total cellular tetherin was measured by co-transfection of eCFP-tetherin with the various Vpu/GFP constructs. Expression of wildtype Vpu resulted in about 50% reduction in the number of tetherin expressing cells, whereas the S52A Vpu degraded tetherin in only about 20% of cells (Fig. 1D). It has been shown that Vpu is phosphorylated by CK-II 12, but the importance of an active CK-II for the ability of Vpu to degrade tetherin is not known. Therefore, we measured Vpu-mediated tetherin degradation in the presence of different CK-II inhibitors (Fig. 1E). Tyrphostin inhibited degradation of tetherin by Vpu already at 25 μM whereas Cay10577 and DRB did so in a dose-dependent manner, demonstrating the importance of CK-II activity for the degrading effects of Vpu on tetherin (Fig. 1E). These results show that mutation of S52A is sufficient to entirely disrupt the effect of Vpu on CD4 and establish at a single cell level that an intact CK-II phosphorylation site as well as active CK-II are important for degradation of tetherin by Vpu.

Vpu S52A still degraded tetherin to some extent in cells co-transfected with Vpu and tetherin expression plasmids (Figures 1D and 1E). Therefore, we speculated that Vpu S52A might be able to enhance HIV-1 release at low levels of tetherin expression. We co-transfected 293T cells with WT, Vpu-defective, and Vpu S52A expressing proviral constructs and different amounts of tetherin ranging from 100 ng (1:50; ratio transfected tetherin:provirus) to 10 ng (1:500); and we measured cellular as well as released p24 by a quantitative Western blot two days later (Fig. 2A and 2B). As expected, 293T cells expressing very low 10 ng (1:500) levels of tetherin released p24 independently of functional Vpu expression. However, transfection of 20 ng (1:250) tetherin already reduced virus release of Vpu-defective HIV-1 by about 50%. At these levels of tetherin expression the S52A Vpu enhanced p24 release as efficiently as the wildtype Vpu protein. In contrast, virus release of the mutant was suppressed by more than one order of magnitude at higher levels of tetherin expression (Fig. 2A, B). Of note, we did not detect any p24 in the supernatant of cells expressing Vpu-defective HIV-1 when tetherin was transfected at a ratio of 1:50. As a control, we measured virion content by ELISA in supernatants of transfected cells before the virus was pelleted. These analyses demonstrated that results obtained by ELISA correlated highly significantly (R = 0.9159; p < 0.0001) with the quantitative WB results (Additional file 1). In sum, the S52A change severely attenuates the ability of Vpu to enhance HIV-1 release with increasing levels of tetherin expression.

Previously, it was reported that macrophages and primary T-cells, the main HIV-1 target cells in vivo, express different amounts of endogenous tetherin 20. Prompted by our results, we speculated that Vpu with a mutated CK-II site might not be able to counteract high levels of tetherin expression found in macrophages, but may replicate efficiently in T-cells that express low levels of tetherin. Since it is known that macrophages exert phenotypically high donor variations, we first aimed to investigate the levels of endogenous tetherin in macrophages from various donors in comparison to autologous T-cells (Fig. 2C). Western blot analysis revealed multiple bands, which is in agreement with previous findings showing that tetherin is glycosylated and can multimerize 101820. Untransfected 293T cells that allow efficient release of HIV-1 particles in the presence and absence of Vpu did not express detectable levels of tetherin (Fig. 2B). Of note, macrophages expressed markedly higher levels of tetherin than PHA-stimulated or unstimulated PBL (Fig. 2C). Thus, Vpu S52A might be differentially active in the enhancement of particle release from primary T-cells and monocyte-derived macrophages (MDM) because it is only able to counteract tetherin at low expression levels.

To investigate the effect of the S52A mutation in Vpu on HIV-1 release we constructed CXCR4(X4)- and CCR5(R5)-tropic HIV-1 NL4-3 mutants carrying this change alone or in combination with a disrupted nef gene. The latter constructs were generated because Nef is known to down-modulate CD4 and to enhance viral infectivity and replication and may thus bias possible effects of the S52A change in Vpu 212223. Western blot analyses confirmed that all proviral constructs showed the expected differences in Vpu and Nef expression (Fig. 3A). Next, we decided to assess first the release of the different HIV-1 NL4-3 variants in the well established HeLa-derived P4-CCR5 cells 2425. We transfected them with normalized quantities of proviral DNA and measured p24 content in the cell culture supernatant. Importantly, transfection efficiencies were comparable, since similar levels of Tat-dependent expression of the LTR-driven β-galactosidase gene were detected in all cell lysates (data not shown). In agreement with the previous finding that Vpu is required for effective virus release from HeLa-derived cell lines 16, the expression of wildtype Vpu resulted in about 5- to 6-fold increased levels of p24 antigen in the culture supernatant. The S52A Vpu enhanced the release of progeny virions with similar efficiency, whereas Nef had no significant effect (Fig. 3B). This result was in line with our hypothesis that Vpu S52A can overcome relatively low levels of tetherin expression, because our P4-CCR5 cells expressed tetherin in a range comparably to unstimulated PBMCs (Additional file 2). Infection of P4-CCR5 cells with virus stocks containing normalized amounts of p24 (1 ng p24) 25 showed that only changes in nef, but not in vpu, impaired viral infectivity (Fig. 3C). Most importantly, these findings demonstrated that the S52A Vpu is capable of enhancing virion release from HeLa derived P4-CCR5 cells that express relatively low levels of tetherin.

It has been demonstrated that Vpu is critical for efficient HIV-1 replication and CD4+ T-cell depletion in HLT ex vivo 2627. This system allows productive HIV-1 infection without exogenous stimulation and mimics infection of lymphatic tissues, one of the major sites of viral replication in vivo 28. To study the effect of the S52A change in Vpu on HIV-1 replication and cytopathicity, we infected HLT ex vivo with the X4 and R5 NL4-3 variants (Fig. 3A). Representative examples of replication results are presented in Figure 4A. Overall, we found that a defective vpu gene reduced the production of wildtype X4 NL4-3 by 60% and of the R5-tropic derivative by 75% (Fig. 4B). Similarly, deletion of nef reduced cumulative virus production by about 75% (Fig. 4B). In contrast, the HIV-1 S52A Vpu mutation did not significantly attenuate HIV-1 replication (Fig. 4A and 4B). Consistent with the results of previous studies 2627, wildtype X4 NL4-3 virus depleted the ex vivo infected tissues of 80% of X4-expressing CD4+ T cells, whereas the R5 HIV-1 derivative depleted 20% of R5+/CD4+ cells (Fig. 4C). Individual or combined deletions in nef and vpu significantly reduced CD4+ T-cell depletion irrespectively of the viral coreceptor tropism, whereas the S52A mutation in Vpu had no significant effect (Fig. 4C). The efficiency of viral replication correlated well with CD4+ T-cell depletion (Fig. 4D) suggesting that these differences in cytopathicity resulted from lower numbers of infected cells, rather than from direct effects of Nef or Vpu on cell killing. These data show for the first time that the CK-II phosphorylation site in Vpu is not critical for effective viral spread and CD4+ T-cell depletion in ex vivo infected lymphoid tissue.

It has previously been established that HIV-1 replication in HLT occurs mainly in both activated and non-activated CD4+ T-cells 29 that express relatively low levels of tetherin (Fig. 2C, Additional file 2). Therefore, the wildtype like phenotype of HIV-1 Vpu S52A in HLT might be due to low tetherin expression levels in the relevant HIV-1 tissue target cells. Since it is difficult to isolate a sufficient number of CD4+ T-cells from these tissues to directly assess endogenous tetherin levels, we decided to investigate if replication of the HIV-1 variants in PBL mimics the situation in HLT. As expected, HIV-1 Vpu S52A replicated as efficiently as WT HIV-1 in cultures of primary blood lymphocytes, whereas Vpu-defective HIV-1 showed attenuated and delayed replication kinetics (Additional file 3 fig. S3a). Furthermore, electroporation of Jurkat T-cells with the proviral constructs and increasing amounts of tetherin expression plasmids confirmed that in T-cells the ability of Vpu S52A to enhance HIV-1 release also decreases in a tetherin-expression dependent manner (Additional file 3 fig. S3B).

Macrophages express markedly higher levels of tetherin than PHA-stimulated or unstimulated PBL (Fig. 2C, Additional file 2). Thus, we finally wanted to challenge the hypothesis that Vpu S52A might be impaired in the enhancement of particle release from infected MDM, because it is not able to counteract high tetherin expression levels. Therefore, we investigated the replicative capacity of the different R5-tropic viruses (Fig. 3A) in MDMs. In agreement with previous reports 30313233, only the disruption of vpu but not of nef severely attenuated HIV-1 replication (Fig. 5). Most remarkably, the S52A mutation in Vpu impaired the replicative capacity of HIV-1 in macrophages as severely as the complete lack of Vpu function. Thus, Vpu S52A might be impaired in the enhancement of particle release from infected MDM, because it is not able to counteract tetherin at high expression levels.

Currently, it is not known whether Vpu modulates cell surface expression of tetherin in primary T-cells and macrophages. To address this, we generated proviral HIV-1 constructs containing wildtype or mutated vpu genes co-expressing Nef and eGFP via an IRES 2534. PBL and MDM were infected with VSV-G pseudotyped viruses and assessed for the modulation of cell surface CD4 and tetherin by FACS. In agreement with previous reports 2122, we found that inactivation of Nef more severely reduced than Vpu the ability of HIV-1 to remove CD4 from the surface of infected primary T-cells (Fig. 6A). Nevertheless, the fact that the combined deletions had the most disruptive effects on cell surface CD4 expression demonstrated that both Nef as well as Vpu are important for effective removal of CD4. Moreover, the S52A change as well as inactivation of Vpu impaired the ability of HIV-1 to down-modulate CD4 to the same extent (Fig. 6A, left). Down-modulation of cell surface tetherin from HIV-1 infected PBL was clearly dependent on Vpu expression (Fig. 6A, right). Furthermore, the levels of cell surface tetherin in infected cells expressing S52A Vpu were significantly lower than in cells infected with HIV-1 containing an entirely defective vpu gene (Fig. 6A, right). Thus, Vpu S52A down-modulates tetherin from HIV-1 infected T-cells, albeit with lower efficiency than wildtype Vpu.

Next, we assessed if our viruses allow us to investigate the modulation of cell surface expressed receptors in macrophages, and we measured the down-modulation of MHC-I as a control. Inactivation of Nef resulted in about 2.5 fold higher MHC-I surface levels compared to WT infected MDM (Fig. 6B). Surprisingly, CD4 expression levels in HIV-1 infected MDM were comparably to uninfected cells, irrespective of Vpu or Nef expression (Fig 6B, left). Moreover, Vpu as well as the S52A mutant had similar minor effects on the levels of cell surface tetherin in MDM (Fig. 6B, right). Notably, MDMs infected with Nef-defective HIV-1 expressed lower levels of tetherin (Fig. 6B, right). Thus, Nef seems to induce tetherin cell surface expression in HIV-1 infected macrophages, perhaps as a result of Nef induced release of inflammatory cytokines 35.

In summary, our experiments demonstrate that VpuS52A reduces the levels of cell surface expressed tetherin in PBL, whereas it does not in macrophages.

For functional analysis, we generated vectors co-expressing Vpu or VpuS52A and GFP from a single bicistronic RNA via an internal ribosome entry site (IRES), as initially described for the analysis of Nef function 49. Briefly HIV-1 NL4-3 Vpu was amplified with primers introducing unique XbaI and MluI restriction sites and subcloned into the pCGCG-IRES-GFP vector 50. AU1-tagged Vpu and VpuS52A variants were constructed by introducing the DTYRYI-sequence at the C-terminus together with the MluI primer. Site directed mutagenesis was utilized to introduce the S52A change in NL4-3 Vpu. The HIV-1 NL4-3 proviral constructs carrying disrupting mutations in nef, vpu or both viral genes have been previously described 27. Splice overlap extension PCR was used to introduce mutation S52A in HIV-1 NL4-3 vpu and the element was subcloned by using the unique restriction sites StuI in env and the PflmI site just downstream of the pol gene, respectively. R5-tropic HIV-1 NL4-3 variants were constructed by exchanging the V3-loop region of NL4-3 with the one from the R5-tropic 92th014.12 isolate 51 by using the unique restriction sites StuI and NheI. HIV-1 NL4-3 variants co-expressing eGFP via an IRES were constructed by subcloning of fragments containing mutations in nef or vpu in the pBR-NL4-3-IRES eGFP backbone 2534. The pECFP-tetherin construct was cloned by amplification of tetherin from a cDNA library (Spring Bioscience) introducing the single cutter restriction sites XhoI and EcoRI. The fragment was cloned in the pECFP-C1 vector backbone (Clontech). An untagged tetherin plasmid was cloned by amplification of tetherin with primers introducing XbaI and MluI sites and subcloning in the pCGCG vector 50. The IRES-GFP cassette was removed by digestion and religation with BamHI. The integrity of all PCR-derived inserts was verified by sequence analysis.

P4-CCR5, 293T and Jurkat cells were cultured as described previously 2550. P4-CCR5 and 293T cells were maintained in Dulbecco's modified Eagle's medium containing 10% heat-inactivated fetal bovine serum. The human Jurkat T-cell line was cultured in RPMI1640 medium supplemented with 10% fetal calf serum and antibiotics. PBMC were generated by Ficoll gradient centrifugation 34 and PBLs were recovered post plastic adherence of monocytes. To generate primary macrophage cultures PBMCs from healthy human donors were isolated using lymphocyte separation medium and macrophages were generated in teflon tubes (CellGenix) and cultured as described before 5253. Transfection of Jurkat T-cells was performed using the DMRIE-C reagent (Invitrogen, Gibco) following manufacturer's instructions. Furthermore, electroporation of Jurkat T-cells with proviral constructs and tetherin expression plasmids was performed with the MP-100 microporator device (PeqLab) as recommended by the manufacturer. Briefly, 4 μg of proviral constructs co-expressing GFP were electroporated with the indicated amounts of tetherin plasmid. Two days post infection GFP+ cells were determined by FACS and the amount of released p24 was quantified in the supernatants using a p24 ELISA provided by the "AIDS & Cancer virus program" (NCI, Frederick). P4-CCR5-cells were transfected using magnetic assisted transfection (IBA Tagnology) following standard protocols of the manufacturer. Briefly, 4000 P4-CCR5 cells per well were sown into 96-well plates one day prior to transfection. For transfection 0.1 μg of proviral DNA was co-incubated with 0.1 μl MaTRA-A reagent in 15 μl OMEM (optimized minimum essential media, GIBCO) per well for 30 min. Three days post transfection supernatants were harvested and analyzed for p24 antigen concentrations and β-galactosidase activity. To generate viral stocks, 293T cells were transfected with the proviral NL4-3 constructs by the calcium chloride method as already described 2534. Virus stocks and supernatants of transfected or infected cells to assess p24 release were quantified using the p24 ELISA described above. Virus infectivity was determined using P4-CCR5 cells as described 25. Briefly, 4000 cells per well were sown out in 96-well-dishes in a volume of 100 μl and infected after overnight incubation with virus stocks containing 1 ng of p24 antigen. Three days post-infection viral infectivity was detected using the Gal screen kit from TROPIX as recommended by the manufacturer. β-galactosidase activities were detected as relative light units per second (RLU/s) in a microplate reader.

CD4 and GFP reporter expression levels in Jurkat cells co-expressing Vpu and eGFP were measured as described previously for the analysis of Nef function 50. Retention of newly synthesized CD4 from the endoplasmic reticulum to the cell surface in 293T cells was measured by standard calcium chloride co-transfection of 1 μg pCDNA-CD4 plasmid with 4 μg pCG plasmid expressing Vpu, VpuS52A or GFP only. Cells were harvested and stained for FACS analysis 2 days post transfection essentially as described previously 50. pECFP-tetherin and GFP expression in 293T cells were analyzed similar to CD4 expression, but on a FACSAria equipped with a 405 nm laser. For the CK-II inhibition experiments, we used Tyrphostin AG1112 (Sigma), Cay10577 (Biozol) and DRB (EnzoLife) reconstituted in DMSO. The concentrations used did not induce cytotoxic effects as determined by FACS FSC/SSC and MTT test (data not shown). PBLs and MDMs were infected with VSVG pseudotyped virus stocks containing 50 ng p24. PBLs were analyzed by flow cytometry three days post infection for CD4 and tetherin expression as already described 34. Similarly, primary macrophage cultures were trypsinized five days post infection and stained for CD4 and MHC-I expression as before 54. Cell surface tetherin was measured by staining of PBLs or MDMs with 1:50 dilutions of the anti-tetherin/CD317 HM1.24 mAb kindly provided by Chugai Pharmaceuticals. As secondary antibody we used an 1:100 diluted Alexa633-conjugated goat anti-mouse antibody (Invitrogen).

293T cells were transfected with 5 μg of vector DNA as described above. Two days post-transfection cells were pelleted, lysates were generated and separated through 12% SDS-PAGE. Expression of Vpu in whole cellular lysates was analyzed by immunoblot using 1:500 diluted mouse anti-AU1 AB (Covance) or 1:500 dilutions of rabbit anti-Vpu sera kindly provided by U.Schubert 55. GFP was detected with a 1:2000 dilution of rabbit anti-GFP AB (Abcam). Viral envelope, GAG and Nef was visualized by 1:5000 dilutions of human-anti-HIV-1 gp120 Ab2G12, 1:5000 diluted rabbit-anti-HIV-1 p24 (provided by the AIDS & Cancer virus program) and 1:1000 dilutions of mouse-anti-HIV-1 Nef (aa151-170) (ABI). For WB-analyses of cellular tetherin we lysed 293T cells that were transfected with different amounts of tetherin as well as P4-CCR5, PBMC, PBL and MDMs and quantified total protein content using the 2D-quant kit (Amersham Biosciences). 40 μg protein of each lysate were separated on a 15%-SDS-bisacrylamide gel. For analysis of tetherin we used mouse anti-tetherin (B01P, Abnova) at a concentration of 1:500 and secondary goat anti-mouse (Jackson Immuno Research) at 1:10000 for ECL detection. To analyze enhancement of p24 release in the presence of tetherin we transfected 400.000 293T cells with proviral constructs (5 μg) and different amounts of tetherin (10, 20 and 100 ng). Two days post infection, we pelleted the filtered supernatants and lysed them as well as the producer cells with 50-100 μl RIPA buffer. Western blots were performed as described above, and band intensity was quantified using the Gene-snap software.

HIV-1 replication and cytopathicity in ex vivo-infected HLTwas determined as described previously 2627. Briefly, human tonsillar tissue removed during routine tonsillectomy was received within 5 hours of excision. The tonsils were washed thoroughly with medium containing antibiotics and sectioned into 2- to 3-mm³ blocks. These tissue blocks were placed on top of collagen sponge gels and infected with virus stocks containing 0.5 ng p24 antigen essentially as described previously 2627. Supernatants were collected at three day intervals and productive HIV-1 infection was assessed by measuring p24 antigen content. Flow cytometry was performed on cells mechanically isolated from control and infected tissue blocks and depletion of CD4+ T-cells was quantified as described previously 2627. For determination of the CD4+/CD8+-T-cell ratio, cells were stained for surface markers by using anti-CD3 fluorescein isothiocyanate (FITC), anti-CD4-allophycocyanin (APC), and anti-CD8 Tri color.

Human monocyte-derived macrophages (MDM) were isolated as described in "cell culture, virus stocks and transfection" and infected with 1 ng p24 of R5-tropic HIV-1 NL4-3 variants. To assess viral spread and replication, aliquots of the infected cell culture supernatants were taken in two to three day intervals and stored at -20°C. Viral replication was determined by RT-assay essentially as described previously 25. Similarly, PBLs were infected with 1 ng p24 of X4-tropic HIV-1 NL4-3 variants expressing GFP via an IRES. For measurement of viral spread and transmission we took aliquots of infected cells in two or three days intervals and determined the amount of GFP+ cells by FACS as before 25.

All statistical calculations were performed with a one-way analysis of variances (ANOVA) using Graph Pad Prism Version 5.0. Correlations were calculated with the linear regression module.