The HIV nef protein associates with protein kinase C theta.

Expression of the human immunodeficiency virus (HIV) Nef protein has been linked to both decreased cell surface expression of CD4 and an impairment of signal transduction. The recently reported association of Nef with an unidentified serine kinase provides a clue as to how Nef might exert its effects. Considering the key role of protein kinase C (PKC) in T cell activation, we investigated the possibility that Nef interacts with PKC. Our results, using two approaches for detecting interactions between Nef and PKC isozymes in Jurkat cells, show that Nef interacts preferentially with θPKC. The interaction of Nef and θPKC is independent of calcium, enhanced by phospholipid activators of PKC and not affected by a PKC pseudosubstrate peptide. Phorbol 12-myristate 13-acetate and phytohemagglutinin stimulation of Jurkat cells expressing Nef fails to produce the usual translocation of θPKC from the cytosol to the particulate fraction; translocation of βPKC and εPKC was unaffected. Indeed, there appears to be a net loss of θPKC in Nef-expressing cells following stimulation. The loss of θPKC, which may be a result of inhibition of its binding to RACKs due to Nef binding, could contribute to the various impairments of T cell function associated with HIV infection and Nef expression.

Expression of the human immunodeficiency virus (HIV) Nef protein has been linked to both decreased cell surface expression of CD4 and an impairment of signal transduction. The recently reported association of Nef with an unidentified serine kinase provides a clue as to how Nef might exert its effects. Considering the key role of protein kinase C (PKC) in T cell activation, we investigated the possibility that Nef interacts with PKC. Our results, using two approaches for detecting interactions between Nef and PKC isozymes in Jurkat cells, show that Nef interacts preferentially with PKC. The interaction of Nef and PKC is independent of calcium, enhanced by phospholipid activators of PKC and not affected by a PKC pseudosubstrate peptide. Phorbol 12myristate 13-acetate and phytohemagglutinin stimulation of Jurkat cells expressing Nef fails to produce the usual translocation of PKC from the cytosol to the particulate fraction; translocation of ␤PKC and ⑀PKC was unaffected. Indeed, there appears to be a net loss of PKC in Nef-expressing cells following stimulation. The loss of PKC, which may be a result of inhibition of its binding to RACKs due to Nef binding, could contribute to the various impairments of T cell function associated with HIV infection and Nef expression.
Nef is one of seven accessory proteins encoded by the human and simian immunodeficiency viruses (HIV and SIV). The 25-kDa nonmyristylated and the 27-kDa myristylated forms occur in the cytoplasm, the nucleus, and the plasma, Golgi, and perinuclear membranes (1)(2)(3)(4)(5)(6)(7). The effects of Nef on virus replication, latency, and host cell functions and survival are unclear (8). Nef has been reported to down-regulate the level of CD4 on the surface of infected T cells (9,10) and to block T cell receptor-mediated induction of interleukin-2 in Jurkat cells, possibly via the inhibition of NFB and AP1 activation (11)(12)(13). Several reports suggest that Nef may impair signal transduction by associating with a serine/threonine kinase (14,15) or disrupt calcium mobilization (16,17) or inhibit protein tyrosine phosphorylation and the subsequent activation of transcription factors (17).
Activation of protein kinase C (PKC) 1 is required for antigenmediated T cell activation, although the cellular targets for its action in the signaling pathway are unknown (18 -20). Moreover, which of multiple PKC isozymes in T cells participates in the signaling pathway has not been determined. The multiple functions attributed to different PKC isozymes as well as the specific intracellular localization characteristic of individual isozymes suggests that endogenous anchoring proteins or receptors for activated C-kinase (RACKs) exist for each isozyme (21)(22)(23). Because the expression of PKC is limited primarily to lymphocytes and skeletal muscle and PKC is among the most abundant of the PKC isozymes in lymphocytes (24 -26), this isozyme is a plausible candidate for involvement in the T cell receptor signaling pathway. However, the intracellular substrates and RACKs that PKC interacts with have yet to be determined. In this work, we have explored the possibility that Nef interacts with PKC and more specifically with PKC. Our results suggest that Nef may interfere with the interaction of PKC with its endogenous anchoring proteins and that this inhibition may result in the net loss of the isozyme.

Co-purification of Nef Binding Proteins and PKC Kinase Assay-
The production and purification of the GST and GST-Nef fusion proteins in Escherichia coli followed the manufacturer's protocol (Pharmacia Biotech Inc.). The GST-Nef fusion protein used the SR␣nef-1 plasmid fused to the pGEX GST gene. Approximately 30 g of GST or GST-Nef fusion protein were bound to 30 l of glutathione-agarose beads in phosphatebuffered saline solution. The labeled beads were then incubated with the cytosolic extract from approximately 5 ϫ 10 6 Jurkat cells with phospholipids (50 g/ml phosphatidylserine and 0.8 g/ml diacylglycerol; Avanti Polar Lipids) and 1 mM calcium for 30 min at 4°C. The beads were then washed three times with cold phosphate-buffered saline. The Jurkat cell extract was prepared by spinning down J25 cells (11), washing with cold phosphate-buffered saline, and resuspended in homogenization buffer (20 mM Tris-HCl, 1 mM EGTA, 1 mM EDTA, 5 g/ml soybean trypsin inhibitor, 5 g/ml phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, pH 7.5). The cells were then sonicated and centrifuged at 14,000 ϫ g and 4°C for 10 min. The supernatant was then precleared with 30 l of GST-labeled glutathione beads for 30 min at 4°C. To detect kinase activity, the washed beads were used in a PKC kinase assay similar to that previously described (27), except that the reaction was carried out at room temperature for 5 min, and the mixture was spotted onto P81 phosphocellulose squares, which were then washed with water and methanol before counting. In the assay, the labeled beads were incubated with a solution containing 20 mM Tris-HCl (pH 7.5), 20 mM MgCl 2 , 20 M ATP, 12 mM ␤-mercaptoethanol, 5 Ci/ml [␥-32 P]ATP, 15 M substrate peptide, plus or minus 50 g/ml phosphotidylserine and 0.8 g/ml diacylglycerol, plus or minus 20 mM CaCl 2 or EGTA, and plus or minus 50 g/ml pseudosubstrate peptide (28). A peptide based upon the PKC pseudosubstrate sequence (25) with the alanine replaced with a serine was used as the substrate (Protein and Nucleic Acid Facility, Stanford).
Western Blot Analysis of Nef Binding Proteins-For the Western blot analysis of PKC isozymes bound to the GST fusion proteins, samples were extracted using glutathione beads prelabeled with GST or GST-Nef fusion proteins as described above. SDS-polyacrylamide gel electrophoresis sample buffer was added to the washed beads to extract the bound proteins for Western analysis as described previously (29). Extract from 1 ϫ 10 6 cells was also run along with the purified proteins for comparison. The nitrocellulose membranes were probed with antibodies for ␣, ␤, ⑀, , and PKC isozymes (the ␣ and ␤ monoclonal antibodies were obtained from Seikagaku America; the ⑀ and polyclonal antibodies were obtained from Life Technologies, Inc./BRL; and the monoclonal antibody was obtained from Transduction Laboratories). The monoclonal antibodies were used at a 1:1000 dilution, whereas the polyclonal antibodies were diluted 1:300. Previous experiments found that these were the most abundant PKC isozymes in these cells. The Western blots were developed using chemiluminescence (ECL; Amersham Corp.).
Co-immunoprecipitations from Jurkat Cells-Immunoprecipitations were performed on the 0.1% Triton X-100 extracts of approximately 5 ϫ 10 7 Jurkat cells that had or had not been exposed to PMA and PHA for 5 min. Cells were stimulated with 100 ng/ml PMA (Sigma) and 2 g/ml PHA (Sigma). The Nef-1-expressing cell line (133) and the J25 parent cell line were used for the immunoprecipitations. Construction of cell lines and their culture are described in Luria et al. (11). Following stimulation, the cell extracts were prepared as above, except for the addition of 0.1% Triton. 6 -8 l of antibodies to Nef (11), PKC (a generous gift from Isakov and Altman (19)) or rabbit preimmune serum were incubated with the extracts for 2 h. The antibody-antigen complex was then precipitated with protein A-Sepharose (Pharmacia). Alternatively, the antibodies were biotinylated with SS-NHS-Biotin (Pierce) in order to prevent the immunoglobulin from disassociating with the other proteins in the extraction step. The antibody-antigen-biotin complex was then precipitated with avidin coupled to agarose (Pierce). The antigen complex was extracted with SDS-polyacrylamide gel electrophoresis sample buffer. Precipitated proteins were detected by Western analysis.
PKC Translocation Assays-To determine the translocation of PKC isozymes (30,31), equal amounts of protein (as determined by Bradford assays; Bio-Rad) from the cytosolic and Triton soluble particulate fractions of unstimulated and stimulated cells were loaded onto 10% SDSpolyacrylamide gels for Western analysis for ␤, ⑀, , and PKC isozymes. The cellular fractions were produced from cell homogenates by a 30-min 100,000 ϫ g centrifugation followed by a 0.1% Triton extraction of the pellet and recentrifugation. The supernatant from the final centrifugation was used as the Triton soluble particulate fraction. The larger molecular weight protein detected by the antibody to PKC, which translocates with stimulation, may be ␣ or ␤PKC as reported by others (32). The Triton soluble particulate fraction does not contain the nuclei or other cellular components not extracted with 0.1% Triton. Western analysis of the Triton nonextracted pellet from unstimulated or stimu-lated cells revealed little PKC remaining in the pellet (results not shown).

RESULTS AND DISCUSSION
To test for an interaction between Nef and PKC, Nef was expressed as a GST fusion protein in E. coli, purified by adsorption on glutathione-agarose beads, and incubated with the cytosolic fraction of Jurkat cells. PKC associated with the GST-Nef fusion protein was assayed by measuring the phosphorylation of a specific PKC substrate peptide (Fig. 1). PKC activity was readily detected in the proteins extracted with GST-Nef but not with GST. Phosphorylation occurred only in the presence of PKC activators and was inhibited by a specific PKC inhibitor, the PKC pseudosubstrate peptide (28). Maximal phosphorylation activity occurred in the presence of phosphatidylserine, diacylglycerol, and EGTA, suggesting that a calcium-independent PKC isozyme binds Nef. A decrease in kinase activity in the presence of calcium was not observed in all experiments and may be the result of increased degradation of the enzyme due to the calcium in some of the experiments.
The PKC family consists of at least eleven isozymes (33), five of which (␣, ␤, ⑀, , and ) were clearly detected by Western blot analysis in Jurkat cells ( Fig. 2A, cytosol). To identify which PKC isozyme bound to Nef, the purified GST-Nef and associated proteins recovered after incubation with Jurkat cell cytosol were electrophoresed, and the protein blots were analyzed with antisera specific for the five PKC isozymes (Fig. 2A). Of the five PKC isozymes detected in Jurkat cells, only PKC bound to the GST-Nef fusion protein; lower amounts of the PKC isozymes bound to GST alone due to nonspecific interactions. Further evidence supporting our conclusion that PKC binds Nef is the observation that the addition of increasing amounts of purified, unbound Nef proportionally lowered the binding of PKC to the fusion protein (Fig. 2B). In addition, treatment of the GST-Nef fusion protein with thrombin, which cleaves Nef from GST, eliminated the binding of PKC to the glutathione beads (results not shown).
We have noted that under our conditions only a fraction of

FIG. 1. A kinase activity from Jurkat cell extracts with characteristics of PKC bound to a GST-Nef fusion protein.
The kinase assay was performed using glutathione beads that had been prelabeled with the fusion proteins and used to purify binding proteins from Jurkat cell cytosolic extracts. The kinase activity was determined in the presence of phospholipids (phosphotidylserine and diacylglycerol, PS/ DAG) and calcium (Ca), phospholipids and EGTA, EGTA alone, or phospholipids, calcium, and the PKC pseudosubstrate inhibitory peptide. A peptide based upon the PKC pseudosubstrate sequence (25) with the alanine replaced with a serine was used as the substrate. The PKC pseudosubstrate inhibitory peptide is a highly specific inhibitor of PKC that is not isozyme-specific. The results are expressed as the averages of duplicate assays and are from one of three experiments with similar results.

FIG. 2. Western blot analysis of PKC isozymes from Jurkat cell cytosolic extracts that bind to a GST-Nef fusion protein.
Glutathione beads were incubated with GST alone or with the GST-Nef fusion protein and were used to purify binding proteins from Jurkat cell extracts as described (A). For comparison, 20% of the cytosol prepared from the Jurkat cells was run in parallel on the Western blot (cytosol) for each isozyme probed (approximate molecular masses: ␣PKC, 82 kDa; ␤PKC, 82 kDa; PKC, 80 kDa; ⑀PKC, 90 kDa; PKC, 69 kDa). The results are from one of four independent experiments with identical results. Preincubation of the Jurkat cell extract with increasing amounts of Nef, purified from E. coli (11), decreased the binding of PKC to the GST-Nef bound to the beads (B; approximate molecular mass of detected protein is 80 kDa). For comparison, approximately 30 g of GST-Nef fusion protein was bound to the beads as determined by Western blot analysis of extracted proteins.
the cytosolic PKC (about 5%) is bound to the GST-Nef, even though the amount of GST-Nef is in molar excess over the amount of PKC. In addition, PKC activators phosphatidylserine and diacylglycerol were present in sufficient concentrations to fully activate the PKC. Moreover, when the cytosolic fraction, which had already been reacted with GST-Nef, was reincubated with fresh GST-Nef, there was no additional binding of PKC. This suggests that PKC in the cytosolic fraction is heterogeneous with respect to its ability to bind Nef.
Co-incubation of PKC with phosphatidylserine, diacylglycerol, and some isozymes with calcium as well results in a conformational change in the protein that allows for substrate phosphorylation and binding to RACKs or other PKC binding proteins (23,28). Therefore, we sought to determine if PKC needs to be activated to bind Nef. To this end GST-Nef, immobilized on glutathione beads, was incubated with Jurkat cell cytosol in the presence or the absence of PKC activators. The recovered binding proteins were analyzed by Western blot analysis (Fig. 3). Binding of PKC was maximal in the presence of phosphatidylserine, diacylglycerol, and EGTA, consistent with the results obtained with the PKC kinase assay (Fig. 1). Nef has been reported to be a PKC substrate (34). However, the interaction between PKC and Nef was not via the PKC substrate binding site. Similar to PKC binding to other proteins that have been shown to interact with PKC such as RACKs (21,22), a PKC peudosubstrate peptide did not inhibit Nef binding to PKC (Fig. 3).
Another approach to determine the association of Nef with PKC relied on co-immunoprecipitation. Antibodies directed against Nef immunoprecipitated both Nef and PKC from extracts of Jurkat cells that express Nef (Fig. 4A). By contrast, ␤PKC was not co-precipitated with anti-Nef antibodies (results not shown), nor did preimmune serum or protein A-Sepharose beads alone precipitate Nef or PKC (Fig. 4A). PMA and PHA stimulation of the Nef-expressing cells did not affect the amount of PKC that co-precipitated with Nef. Considering that the cells are grown in media containing growth factors, it is possible that there is a sufficient amount of activated PKC to associate with the endogenously produced Nef; such a possibility is inferred from the observation that some particulate fraction-associated PKC is usually present in Nef-expressing Jurkat cells even before stimulation.
Treatment of the Jurkat cell extracts with various preparations of antibodies to PKC immunoprecipitated PKC but failed to co-precipitate detectable levels of Nef using protein A-Sepharose (Fig. 4A). However, Nef could be co-immunoprecipitated with PKC using biotinylated antibodies to PKC and avidin-agarose (Fig. 4B). Using biotinylated antibodies removed the background signal resulting from the immunoglobulin light chain that migrated at the same molecular weight as Nef on SDS-polyacrylamide gel electrophoresis. This method greatly improved the sensitivity of the Western for immunoprecipitated Nef but also lowered the amount of antigen that was precipitated. In particular, biotinylated antibodies to Nef did not consistently co-precipitate PKC (Fig. 4B), possibly because of interference by the biotin on the binding of the antibodies to the complex.
In vivo activation of PKC generally results in the translocation of the isozymes from the cytosol to the particulate fraction with no net change in the total amount of the enzymes (30,31). However, in Jurkat cells expressing Nef, the total amount of PKC declined after 5 min of PMA and PHA stimulation (Fig.  5A); by comparison the total level of ␤PKC remained unchanged after stimulation. To examine this isozyme-specific difference further, the translocation of four PKC isozymes was compared in PMA-and PHA-activated control and Nef-expressing Jurkat cells (Fig. 5B). Translocation of ␤, ⑀, and PKC isozymes was observed after 5 min of PMA and PHA stimula- FIG. 3. PKC binding to the GST-Nef fusion protein requires phospholipids and is not inhibited by a PKC pseudosubstrate peptide. GST or GST-Nef fusion proteins immobilized onto glutathione beads were incubated with Jurkat cell cytosol in the presence of phosphotidylserine and diacylglycerol (PS/DAG), calcium (Ca), EGTA, or the PKC pseudosubstrate peptide. PKC binding was determined as above (approximate molecular mass of detected protein is 80 kDa).
FIG. 4. PKC and Nef co-immunoprecipitate from Jurkat cells expressing Nef. Immunoprecipitations were performed using extracts of unstimulated or stimulated (PMA and PHA) cells expressing Nef or control cells not expressing Nef, using polyclonal antibodies against Nef, PKC, or rabbit preimmune serum or protein A-Sepharose beads alone (A). For control precipitations, equal volumes of unstimulated and stimulated cells were pooled (pool) or extracts from an equal number of J25 Jurkat cells not expressing Nef were used. Alternatively, Nef and PKC were co-precipitated using biotinylated antibodies to PKC (B). Biotinylated antibodies to Nef and biotinylated rabbit preimmune serum were also used. Precipitated proteins were detected by Western analysis with antibodies against PKC and Nef (approximate molecular masses: PKC, 80 kDa; Nef, 27 kDa). tion in the control Jurkat cells; translocation of the ␤ and ⑀ isozymes was virtually unchanged in the Jurkat cells expressing Nef (PKC localization is unaffected by PMA, whereas the higher molecular weight cross-reactive proteins may be other PKC isozymes (32)). However, PKC translocation was disrupted in the cells expressing Nef; stimulation resulted in the loss of PKC from the cytosolic fraction without a concomitant increase in the level of PKC in the particulate fraction. Furthermore, as pointed out earlier, even without stimulation, there was a greater proportion of PKC in the particulate fraction of Nef-expressing cells than in control cells (Fig. 5B).
The loss of PKC following PMA and PHA treatment of Nef-expressing cells may be a result of the isozyme's interaction with Nef. This could result from the failure of PKC to bind to its endogenous anchoring protein(s) or RACKs (21-23) following activation of the PKC. Activated PKC has been shown to be more sensitive to proteolysis (36). Inhibition of binding of activated PKC to its endogenous binding proteins or RACKs may leave the PKC more susceptible to degradation. Indeed, an inhibition of translocation and a net loss of PKC following activation was observed in oocytes in which translocation inhibitors (either a purified RACK protein or peptides based upon sites of interaction between PKC and RACKs) were introduced by microinjection (35,37,38). Nef may act as an inhibitor of PKC translocation in lymphocytes in a similar manner. Quite possibly, the loss of PKC or its inappropriate binding to its targets could account for the various phenotypic impairments of T cell function associated with Nef and HIV infection. Further work is required to determine the molecular basis for the loss of PKC in cells expressing Nef and how this disruption in signal transduction affects T cell activation.
FIG. 5. Nef expression in Jurkat cells correlates with the specific loss of PKC following PMA and PHA stimulation. The total level of PKC declined in Nef-expressing cells following PMA and PHA stimulation, whereas total ␤PKC levels remain unchanged (A; approximate molecular masses: PKC, 80 kDa; ␤PKC, 82 kDa; Nef, 27 kDa). Triton extracts of total cell homogenates were prepared of the stimulated and unstimulated Nef-expressing cells and subjected to Western analysis. The levels of Nef in the two cell extracts were also determined to demonstrate that equal amounts of protein were loaded in both lanes. Control Jurkat cells or Nef-expressing cells were stimulated with PMA and PHA and fractionated into cytosol and Triton-soluble particulate fractions to further analyze the loss of PKC (B). Samples containing equal amounts of protein were analyzed by Western blot to determine the relative levels of , ␤, ⑀, and PKC before and after stimulation in the two cellular fractions (approximate molecular masses: PKC, 80 kDa; ␤PKC, 82 kDa; ⑀PKC, 90 kDa; PKC, 69 kDa, lower band). The results are from one of six independent experiments with similar results.