Here, we show that the ability of Nef to associate with activated PAK2 is important for its ability to enhance HIV replication in freshly isolated T cells. Mutations at positions 89 and 191, which disrupt PAK2 association, rendered Nef defective for the ability to enhance cellular activation and viral replication in freshly isolated T cells, but not the ability to enhance viral infectivity or downmodulate CD4 and MHC Class I . As expected and consistent with other reports [62, 63], by targeted siRNA knockdown we show that PAK2 is important not only for Nef-mediated enhancement of T cell activation but also for activation of T cells in the absence of Nef (Figure 5). We also show that Nef-mediated enhancement of T cell activation is abrogated in the presence of a dominant negative PAK2 mutant (Figure 6). The ability of wild-type or mutant Nefs to enhance T cell activation correlated with their ability to associate with PAK2 in co-precipitation assays in Jurkat and 293T cells (Figure 6A and 7). Together, these data are consistent with a model in which enhancement of T cell activation and HIV replication occur through related mechanisms involving Nef association with PAK2, most likely within a multiprotein complex.
Previous studies suggest that Nef enhances HIV replication by reducing the threshold of cellular activation [19, 24]. Consistent with this model, we found that Nef-dependent enhancement of viral replication in T cells stimulated after infection is highly dependent on the strength of activating stimuli. The greatest differences in levels of replication between Nef+ and Nef- viruses were observed in cultures sub-maximally stimulated with 1 μg/ml PHA-P for 72 h (Figure 2A). The strength of the activating stimulus correlated inversely with Nef-mediated enhancement of HIV-1 replication (Additional File 1, Figure S1). Reducing the concentration of α-CD3/CD28 beads (Additional File 2, Figure S2), or shortening the duration of stimulation (Additional File 3, Figure S3 and Figure 3, Figure 4, and ), is required to detect Nef-mediated enhancement of cellular activation. These findings imply that sub-maximal stimulation, regardless of whether PHA-P or α-CD3/CD28 beads are used, is required to detect Nef-mediated enhancement of viral replication and T cell activation. Excluding IL-2 from the cell culture media resulted in only 1.6% viable cells. Thus, we cannot exclude the possibility that IL-2 alone increased the number of cells permissive for infection . Importantly, these findings demonstrate that Nef-mediated enhancement of replication and activation can be masked when strong stimuli are used to induce cellular activation.
Nef may enhance HIV replication via several distinct mechanisms that are not mutually exclusive. Nef enhances viral replication in freshly isolated PBMC, which contains a small fraction of activated or partially activated T cells [6, 7, 9]. Nef expression can enhance the promoter activity of the LTR  and specific host cell genes [24, 59, 65] via mechanisms that may involve upregulation of tat-SF1, U1 SNRNP, and IRF-2 mRNAs  and enhancement of NFAT activity [20, 23], thereby increasing the amount of viral particles produced per cell. As we and others have demonstrated, Nef interaction with cell signaling machinery in unstimulated T cells may render them permissive for high levels of productive infection upon subsequent stimulation by reducing the threshold required for cellular activation. As such, the inverse correlation between Nef-mediated enhancement of replication and the strength of cellular stimulation (Additional File 1, Figure S1B) is likely to reflect both an increase in the number of permissive cells and increased p24 production per cell. We found that activation of bystander CD4 and CD8 T cells is enhanced in the presence of Nef expression cells. Bystander activation may be important for HIV replication in vivo, as this would be expected to increase the pool of target cells permissive for productive infection and to contribute to activation induced cell death of CD4 and CD8 T cells associated with HIV infection. Therefore, Nef-mediated enhancement of T cell activation may positively influence viral replication and pathogenesis via several distinct mechanisms.
The role of PAK2 in Nef-mediated enhancement of T cell activation is unclear because PAK2 activation is dependent on several cellular factors. The results of experiments using PAK2 siRNA knock-down and a dominant negative PAK2 mutant (Figure 5 and 6) imply that PAK2 itself, and/or molecules that interact with PAK2, are important for Nef-mediated enhancement of T cell activation. For example, activation of PAK2 is dependent on binding of CDC42-GTP to the PAK2 CRIB (Cdc42- and Rac-interactive binding) domain. CDC42 binds PAK2 only in its GTP bound state, which occurs after guanine nucleotide exchange factors (GEF), such as Vav, induce exchange of GDP for GTP. Nef may therefore interact with upstream signaling molecules such as SFKs, Vav1, or CDC42 in order to associate with activated PAK2. Identifying binding partners of Nef important for PAK2-association may help to identify other host cell factors necessary for Nef-mediated enhancement of T cell activation and viral replication.
Previous reports evaluating the contribution of Nef-PAK2 association to HIV replication reached conclusions that differ from our own. One study demonstrated that siRNA knockdown of PAK2 was not important for infection of HeLa and Jurkat cells . However, freshly isolated then activated primary T cells are a more relevant cell-type for studies of Nef function. Furthermore, HeLa and Jurkat cells do not require activation by external stimuli to become permissive for HIV replication. A second study used experimental conditions that may mask Nef-mediated enhancement of T cell activation . Schindler et al.  incubated Jurkat NFAT-Luc reporter cells with PHA for 16 h. However, we found that extending the time of stimulation with α-CD3/CD28 coated beads from 4 to 8 h abrogated Nef-mediated enhancement of NFAT-Luc activity (Figure 3 and Additional File 3, Figure S3) and extending the time of stimulation with 1 μg/ml PHA-P from 18 to 24 h abrogated Nef-mediated enhancement of CD25 upregulation in Jurkat E6.1 cells (Additional File 4, Figure S4). Thus, our identification of specific assay conditions in which Nef expression enhances T cell activation and HIV replication sheds light on potential explanations for different results among studies that examined the importance of PAK2 for Nef-mediated enhancement of HIV replication.
Significant controversy surrounds the issue of whether or not Nef-PAK2 association contributes to SIV pathogenesis. In two studies, rhesus macaques infected with SIV mutated in the PxxP domain of Nef did not develop high viral loads until reversion of the mutations occurred [45, 67]. In contrast, one study demonstrated that SIVmac239 containing the same PxxP mutation began to develop high viral loads 4 days prior to detection of reversion to wild-type . The inconsistency of these findings may relate to certain limitations of the SIV model. Because of rapid disease progression, SIVmac239 infection of macaques may not be an accurate model for the chronic phase of HIV infection. PAK2 activation may be more important during chronic infection, when immune activation is lower and infected T cells are more likely to be resting, than during acute infection and late stage disease, when T cells have higher levels of activation. Indeed, lymphoid-derived Nefs in late stages of disease acquire rare, non-conservative mutations at positions critical for PAK2-association potentially abrogating the ability of Nef to associate with PAK2 . Accordingly, Nef-PAK2 association may be important for pathogenesis in vivo in the same context for which it is important for enhancing replication in vitro: when infected T cells are resting.
An important finding in our study relates to technical issues that may help to explain major discrepancies between results from different groups regarding HIV replication in freshly isolated PBMC. Three technical points were critical for reproducibility of a strong Nef-dependent phenotype in freshly isolated PBMC: 1) using PBMC isolated from fresh blood instead of cryopreserved PBMC; 2) using highly purified PBMC free of platelet contamination; and 3) using pooled human A/B serum determined to be free of endotoxin. Our preliminary studies comparing different methods for this assay suggested that PBMC derived from cryopreserved rather than fresh samples, as well as PBMC containing even low levels of contaminating platelets, are more "activated" than PBMC obtained from fresh blood or isolated free of platelet contamination. Thus, assays that use cryopreserved PBMC or are conducted in the presence of contaminating platelets are not representative of results in "resting" PBMC. Platelets can release RANTES and other soluble factors that can activate PBMC, so platelet-free PBMC preparations are critical for assays that depend on a "resting" PBMC phenotype. We observed that fetal bovine serum (FBS) enhances survival of activated PBMC compared to pooled human A/B serum, thereby increasing the number of activated cells and reducing the percentage of resting cells. Thus, stimulation in the presence of FBS may result in stronger cell stimulation than delivering the same stimulus in the presence of human A/B serum. We have shown that the ability of Nef to enhance replication in freshly isolated PBMC is highly dependent on the strength and duration of stimulation. Therefore, confounding factors that enhance the strength of stimulation may mask certain Nef phenotypes. For example, the presence of endotoxin in some serum preparations may alter the cytokine profile of PBMC. In summary, using freshly isolated rather than cryopreserved PBMC and reducing PBMC activation by factors such as platelet contamination and serum endotoxin, and determining the strength and duration of stimulation required for sub-maximal stimulation are critical in order to achieve consistent effects of Nef on HIV-1 replication in resting T cells from a given donor.