Simian Immunodeficiency Virus and Human Immunodeficiency Virus Type 1 Nef Proteins Show Distinct Patterns and Mechanisms of Src Kinase Activation

Abstract

The nef gene from human and simian immunodeficiency viruses (HIV and SIV) regulates cell function and viral replication, possibly through binding of the nef product to cellular proteins, including Src family tyrosine kinases. We show here that the Nef protein encoded by SIVmac239 interacts with and also activates the human Src kinases Lck and Hck. This is in direct contrast to the inhibitory effect of HIV type 1 (HIV-1) Nef on Lck catalytic activity. Unexpectedly, however, the interaction of SIV Nef with human Lck or Hck is not mediated via its consensus proline motif, which is known to mediate HIV-1 Nef binding to Src homology 3 (SH3) domains, and various experimental analyses failed to show significant interaction of SIV Nef with the SH3 domain of either kinase. Instead, SIV Nef can bind Lck and Hck SH2 domains, and its N-terminal 50 amino acid residues are sufficient for Src kinase binding and activation. Our results provide evidence for multiple mechanisms by which Nef binds to and regulates Src kinases.

The nef gene is unique to primate lentiviruses (human immunodeficiency type 1 [HIV-1], HIV-2, and simian immunodeficiency virus [SIV]) and encodes a myristoylated membrane-associated protein of 25 to 34 kDa (13, 21). Nef is essential for high-level virus replication and full pathogenic potential during SIV infection of adult rhesus macaques, and HIV-1 infection of reconstituted hu-Scid mice (23, 24). The mechanisms by which Nef acts as a positive factor during infection are unclear but are most likely related to both viral and cellular factors. Nef has been shown to enhance virion infectivity and reverse transcriptase activity (1, 32, 36). At the cellular level, Nef reduces the level of cell surface receptors including CD4 (14, 21), interleukin-2 receptor (19), and major histocompatibility complex (MHC) class I (37), interferes with T-cell signalling (6, 9, 22, 28, 40), and impairs specific cytokine production (8, 30). These observations imply that nef has a role in perturbing the cell activation pathway(s), which presumably influences viral replication in the host and possibly cause dysfunction of cells in the immune system. This hypothesis is substantiated by the observations that Nef interacts with several cellular proteins including multiple members of the Src kinase family and modulates their catalytic activities (9, 11, 17, 35).

HIV-1 Nef interacts with the Src family kinases Lck and Hck via the Src homology 3 (SH3) domain of each kinase and a highly conserved polyproline type II (PPII) helix-structured proline motif within Nef, the disruption of which affects HIV replication and infectivity as well as MHC class I down-regulation (15, 16, 18, 35, 41). Binding of HIV-1 Nef to Hck causes a dramatic increase in Hck catalytic activity (7, 33). This effect was considered a consequence of Nef displacement of the SH3 domain of the kinase from a PPII helix chain linking the SH2 and the catalytic domains in an inactive form, causing a conformational change in the amino-terminal lobe of the catalytic domain which enhances phosphotransfer (33, 38). Such displacement was proposed as a mechanism by which the catalytic activity of all Src family kinase members may be regulated. However, binding of Nef to the SH3 domain of Lck results in inhibition of Lck catalytic activity, suggesting that SH3 regulation of Src kinase activity may differ among family members (9, 18).

The Nef protein encoded by SIV shares striking functional similarities with its HIV-1 counterpart (5, 22, 37, 39). SIV Nef was also found to associate with Src, and this interaction correlates with both the efficiency of viral replication and the severity of disease following experimental infection of macaques with the acutely pathogenic strain SIVpbj14 (11). SIV Nef contains a consensus PPII proline motif, and based on this sequence similarity we hypothesized that SIV Nef may interact with and regulate Src family kinases via their SH3 domains. We have now investigated the mechanism(s) by which SIV Nef interacts with and modulates Src family kinases.

SIV Nef can bind directly to Src family kinases.

We compared the binding of SIV and HIV-1 Nef to the Src family kinases Lck and Hck. These kinases are expressed in T lymphocytes and monocytes, respectively, both cell types being targeted by HIV-1 as well as SIV. Direct binding to Lck of a purified glutathione S-transferase (GST)–Nef fusion protein, corresponding to the nef genes from HIV-1(NL43) and SIVmac239, was investigated by using a previously described quantitative microtiter plate binding assay with immobilized pure Lck (100 nM; Upstate Biotechnology, Lake Placid, N.Y.) (18). For the expression of GST-SIV Nef, plasmid p239SpE3′ containing the 3′ half of SIVmac239 open proviral genome, obtained through the AIDS Research and Reference Reagent Program Division of AIDS, National Institute of Allergy and Infectious Diseases, National Institutes of Health, from R. Desrosiers, J. Gibbs, and D. Regier, was used as the template for the PCR amplification of SIVmac239 nef. To obtain the full-length nef sequence to be cloned into pGEX 4T-1, the nef gene was amplified by PCR by using the 5′ primer 5′-CGGGATCCATGGGTGGAGCTATTTCCATGAGG and the 3′ primer 5′-ATAAGAATGCGGCCGCTCAGCCATGTTAAGAAGGCCTCTTGC, which introduced the unique BamHI (5′ primer) and NotI (3′ primer) restriction sites. To construct GST-SIV Nef, the PCR fragment was digested with BamHI and NotI and cloned into pGEX 4T-1 digested with the same enzymes. All vectors were sequenced to ascertain precise nef sequence and to ensure that the nef sequences were in frame with the sequence coding for GST. The expression and purification of GST–HIV-1 Nef and GST-SIV Nef were performed as described previously (3, 19a). Both the GST-SIV and GST-HIV-1 Nef proteins, but not the GST control protein, bound to Lck in a concentration-dependent manner, plateauing above 100 nM, indicating an equimolar interaction (Fig. 1A). HIV-1 Nef has been reported to interact with multiple members of the Src family kinases via its highly conserved proline motif but with different affinities (9, 11, 12, 17, 35). As SIV Nef also possesses the highly conserved PPII minimal proline motif (Fig. 2), we investigated whether similarly to HIV-1 Nef, the interaction of SIV-Nef with Lck was dependent on this motif. A GST-SIV Nef fusion protein which contained alanine residues in place of the proline residues at positions 104 and 107 [GST-SIV Nef (AxxA)] was included in the microtiter plate binding assay. To mutate the proline residues occurring at amino acid residues 104 and 107 of SIV Nef to alanine residues, mutagenic primers (5′ primer 5′-GGTATCAGTGAGGGCAAAAGTTGCCCTAAGAACAATG and 3′ primer 5′-CAT TGT TCT TAGGGCAAC T TT TGCCC TCAC TGATACC) were used for site-directed mutagenesis using a QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, Calif.) according to the manufacturer’s instructions. The mutant form of nef was completely sequenced to verify the presence of the mutation and the absence of any other changes. The expression in Escherichia coli and purification of Nef derived from the nef gene of SIVmac239 were performed essentially as described previously (3). GST-SIV Nef (AxxA) bound to Lck as efficiently as the parental SIV Nef protein, suggesting that the proline motif is not essential for SIV Nef interaction with Lck (Fig. 1A). In contrast, the corresponding GST–HIV-1 Nef (AxxA) did not bind to immobilized Lck, supporting our previous findings (19a) that the P72 and P75 residues of HIV-1 Nef are essential for its interaction with Lck (Fig. 1A).

FIG. 1.

Binding of SIV-Nef or HIV-1 Nef to full-length Lck and Hck. (A) Purified Lck at 100 nM was coated onto the wells of 96-well polystyrene microtiter plates. After coating and blocking with gelatin, the wells were incubated with increasing amounts of either GST–HIV-1 Nef, GST–SIV Nef, GST–HIV-1 Nef (AxxA), GST-SIV Nef (AxxA), or GST as a control (0 to 300 nM). Binding of GST-Nef was detected with an anti-GST antibody or as a control an irrelevant antibody at the same immunoglobulin concentration as anti-GST, followed by sequential incubation with a biotin-conjugated anti-rabbit immunoglobulin, streptavidin-conjugated horseradish peroxidase (HRP) and o-phenylenediamine substrate solution (18). After incubation, the optical density was measured at 450 and 630 nm. Results are graphed after subtraction of background binding obtained with the irrelevant antibody control. (B) Purified Lck was reacted with purified GST, GST–HIV-1 Nef, GST–HIV-1 Nef (AxxA), GST–SIV Nef, or GST-SIV Nef (AxxA), which were expressed and purified as described above, coprecipitated by using glutathione-Sepharose beads, separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), transferred to nitrocellulose, and immunoblotted with polyclonal antibodies specific for Lck. This was followed by incubation with HRP-conjugated anti-rabbit immunoglobulin and development using enhanced chemiluminescence (ECL) substrate. (C) Cellular protein extracts from monocyte-derived macrophages, purified as pre- viously described (19a), were reacted with GST, GST–HIV-1 Nef, GST–HIV-1 Nef (AxxA), GST–SIV Nef, or GST-SIV Nef (AxxA), which were expressed and purified as described above, coprecipitated by using glutathione-Sepharose beads, separated by SDS-PAGE, transferred to nitrocellulose, and immunoblotted with polyclonal antibodies specific for Hck. This was followed by incubation with HRP-conjugated anti-rabbit immunoglobulin and development using ECL substrate. (D) Association of SIV Nef and Hck during intracellular expression. 293 cells were grown to 60 to 80% confluency before being transfected as described previously (34) with cytomegalovirus (Cmv)-driven expression vectors for SIV nef or SIV nef (AxxA) together with a pBabe retroviral vector expressing Hck or puromycin resistance alone (kindly donated by K. Saksela). At 48 h after transfection, the cells were harvested from culture and washed twice in phosphate-buffered saline, and cytoplasmic extracts prepared as described previously (17). Anti-Hck and control immunoglobulin used at the same concentration as anti-Hck were used to prepare immunoprecipitates (IP), which were then separated by SDS-PAGE, transferred to nitrocellulose, and immunoblotted with SIV Nef monoclonal antibody 17.2. con Ab, control antibody.

FIG. 2.

Nef-SH3 interface. Secondary structure and structurally based sequence alignment of Nef core regions from primate lentiviruses. The secondary structure shown is that of HIV-1 Nef. Structurally related residues of HIV-1, HIV-2, and SIV Nef (Los Alamos National Laboratory database) are shown in boldface for conserved and nonconserved residues. The asterisk indicates the strictly conserved R77 residue involved in the formation of an extensive network of interactions with other components of the Nef structure (in the αB helix) and with the SH3 domain (2, 26). cons., consensus.

In parallel experiments, purified Lck was reacted in solution with the recombinant GST-Nef fusion proteins followed by immobilization on glutathione-Sepharose 4B beads. The precipitates were analyzed by specific immunoblotting for Lck. Lck was coprecipitated by both GST–SIV Nef and GST–HIV-1 Nef but not by GST alone (Fig. 1B). Again the interaction of HIV-1 Nef with Lck was dependent on the proline motif within Nef, as GST–HIV-1 Nef (AxxA) did not coprecipitate Lck (Fig. 1B). In sharp contrast, the binding of Lck to SIV-Nef appeared independent of the corresponding proline motif, as GST-SIV Nef (AxxA) coprecipitated similar amounts of Lck as GST-SIV Nef. Similarly, both GST-SIV Nef and GST–HIV-1 Nef, but not GST alone, coprecipitated with Hck, as determined by Hck immunoblotting, when the recombinant GST fusion proteins were reacted with cytoplasmic extracts prepared from monocytes (Fig. 1C). The HIV-1 Nef interaction with Hck displayed a proline motif dependency identical to that observed with the Nef-Lck interaction (Fig. 1C). Once more, SIV-Nef interaction with Hck was independent of its proline motif (Fig. 1C).

HIV-1 and SIV Nef interaction with Lck and HIV-1 Nef interaction with Hck have been identified when each protein is expressed during transient transfection of cells with the appropriate vectors or viral infection of cells with HIV-1 (4, 7, 9, 19a). To determine that the GST-SIV Nef-Hck coprecipitation studies described above reflect protein-protein interactions which occur when the proteins are expressed together in a cellular environment, we expressed Hck, SIV Nef, and SIV Nef (AxxA) during transient transfection of 293 cells. For the generation of the SIV nef or SIV nef(AxxA), plasmid p239SpE3′ containing the 3′ half of the SIVmac239 open proviral genome was used as the template for the PCR amplification of SIVmac239 nef. To obtain the full-length nef sequence to be cloned into pCMV, the nef gene was amplified by PCR by using the 5′ primer 5′CGGGATCCCCACCATGGGTGGAGCTATTTCCATGAGG and the 3′ primer 5′-ATAAGAATGCGGCCGCTCAGCCATGTTAAGAAGGCCTCTTGC, which introduced the unique BamHI (5′ primer) and NotI (3′ primer) restriction sites. pCMV SIVnef was constructed by using pEGFP-N1 (Clontech, San Diego, Calif.) previously digested with BamHI and NotI to remove the coding sequence for enhanced green fluorescent protein and subcloning the PCR DNA fragments digested with the same enzymes into the vector to generate pCMV SIVnef. pCMV SIVnef(AxxA) was generated as described above. Anti-Hck immunoprecipitates derived from cells cotransfected with vectors expressing the Hck or Nef protein were subjected to immunoblot analysis with anti-SIV Nef antibodies. Hck immunoprecipitates derived from 293 cells transfected with Hck and SIV Nef constructs specifically contained SIV Nef (Fig. 1D), as well as the expected band of approximately 60 kDa when immunoblotted with further antibodies specific to Hck (data not shown). SIV Nef immunoblotting of the immunoprecipitates obtained with the control antibody verified the specificity of the Hck-SIV Nef interaction (Fig. 1D). Consistent with the GST coprecipitation studies described above, the interaction of SIV Nef with Hck in transfected 293 cells was not dependent on its proline motif, as SIV Nef (AxxA) was efficiently coprecipitated with Hck (Fig. 1D).

Therefore, Nef from both HIV-1 and SIV can interact with the Src kinases Lck and Hck, as verified by GST fusion protein coprecipitation studies and also during cellular expression. However, SIV Nef and HIV-1 Nef display different requirements for the proline motif in binding to the Src kinases. Furthermore, Lck binding to Nef does not require other virion or cellular proteins since binding was observed with purified Lck and HIV-1 Nef or SIV Nef proteins only.

SIV-Nef can induce phosphotransferase activity of Src family kinases.

Next, the functional consequences of SIV Nef binding to Src family kinases Lck and Hck were investigated. Lck and Hck were immunoprecipitated from the Jurkat leukemic CD4⁺ T-cell line and primary monocytes, respectively, using specific rabbit antibodies. Immunoprecipitated Lck and Hck were incubated with increasing amounts of either Nef or GST protein, added to the peptide p34^cdc2[Lys 19 (6–20)NH₂] or the control peptide p34^cdc2[Lys 19, Phe 15 (6–20)NH₂] and subjected to kinase assay as previously described (18). Both SIV Nef and HIV-1 Nef efficiently and specifically activated cell-derived Hck phosphotransferase activity in a dose-dependent manner, up to 400% above the basal level, compared to the control GST protein (Fig. 3A). Surprisingly, SIV Nef induced a dose-dependent increase in Lck activity (up to 300% above the basal level), with half-maximum activation reached at 21.1 nM (Fig. 3A). This result directly contrasted with the dose-dependent inhibition of Lck activity caused by HIV-1 Nef (Fig. 3A) (9, 18). Similar results were obtained for purified Lck (Fig. 3B), indicating that cellular cofactors of Lck were not required for regulation of Lck activity by SIV and HIV-1 Nef. As expected, mutation of the proline motif within HIV-1 Nef abrogated its ability to modulate Lck and Hck catalytic activities, supporting our observations that this motif is essential for HIV-1 Nef interaction with Src kinases (Fig. 3). In contrast, however, alteration of the corresponding proline motif within SIV-Nef did not affect its ability to augment both Hck and Lck kinase activities (Fig. 3).

FIG. 3.

Differential effects of SIV Nef and HIV-1 Nef on Lck and Hck kinase activity. (A) The kinase activity of Lck or Hck derived by immunoprecipitation from Jurkat cells or monocytes, respectively, was measured by using p34^cdc2 peptide as the substrate (18, 19a). Lck and Hck precipitates were preincubated with increasing amounts of GST, GST–HIV-1 Nef, GST–HIV-1 Nef (AxxA), GST–SIV Nef, or GST-SIV Nef (AxxA) protein, added to the peptide p34^cdc2[Lys 19 (6–20)NH₂] or the control peptide p34^cdc2[Lys 19, Phe 15 (6–20)NH₂], and subjected to kinase assay. Incorporation of [³²P]ATP was measured by scintillation counting. Results are expressed as a percentage of untreated kinase activity after subtraction of the control peptide from the test samples. (B) As for panel A except that recombinant purified Lck was used (100 nM) instead of Lck derived from Jurkat cells by immunoprecipitation.

Differential binding of HIV-1 and SIV-Nef to SH3 domains.

Since augmentation of Hck activity by HIV-1 Nef occurs through SH3 binding (33), we compared the interaction of SIV Nef and HIV-1 Nef with the regulatory domains of Lck and Hck in our quantitative microtiter plate binding assay. In contrast to HIV-1 Nef, SIV Nef did not bind to immobilized Lck and Hck SH3 domains in detectable levels (Fig. 4). Similarly, the HIV-1 Nef (AxxA) mutant failed to detectably interact with Lck and Hck SH3 domains in this assay (Fig. 4). We also compared HIV-1 and SIV Nef binding to Src family SH3 domains in coprecipitation studies using recombinant GST fusion proteins (9, 12). Jurkat cells were transfected with HIV-1 and SIV nef plasmid constructs. Cell lysates were prepared and reacted with GST-Lck-SH3 and GST-Hck-SH3 recombinant proteins. GST-Lck-SH3 (Fig. 5A, lane 4) and GST-Hck-SH3 (Fig. 5B, lane 4) specifically precipitated HIV-1 Nef from the transfected Jurkat cells. As previously described, GST-Hck-SH3 bound greater amounts of HIV-1 Nef than Lck-SH3 (9, 12) (although this is not reflected in the solid-phase binding assays, where we see no appreciable difference between Lck and Hck binding to Nef). In contrast, neither GST-Lck-SH3 nor GST-Hck-SH3 coprecipitated detectable levels of SIV Nef. These data suggest that SIV Nef has a greatly reduced binding capacity for the SH3 domain of Lck and Hck compared with HIV-1 Nef and that the SH3 domain of each kinase is unlikely to be a major determinant of SIV Nef-Src kinase interaction (Fig. 5A and B, lanes 4). These results were unexpected since the proline motif in HIV-1 Nef, which binds to various Src kinases via their SH3 domains, is well conserved in various Nef variants, including that of SIVmac239 (Fig. 2). However, our results are consistent with both Lck and Hck binding and activation by SIV Nef independently of its proline motif, as shown above.

FIG. 4.

GST fusion proteins of HIV-1 but not SIV Nef bind Hck and Lck SH3 domains directly. The SH3 domain of Lck or Hck was coated at a concentration of 100 nM onto the wells of 96-well polystyrene microtiter plates. Binding of either GST–HIV-1 Nef or GST–HIV-1 Nef (AxxA), GST-SIV Nef, or GST-SIV Nef (AxxA) was determined as described for Fig. 1A.

FIG. 5.

Binding of HIV-1 Nef and SIV Nef to the SH2 domains of Src family kinases. (A) Jurkat cells were transiently transfected with pcDNA3-nef, corresponding to HIV-1(BRU) nef, or SIVmac239 nef or were mock transfected. Cell lysates were then prepared and reacted with 10 μg of the indicated GST protein as previously described (9). Glutathione-Sepharose affinity-purified precipitates were analyzed by SDS-PAGE, transferred to nitrocellulose, reacted with antibodies specific for either HIV-1 Nef (HIV-1 Nef monoclonal antibody MAT0020; Transgene, Strasbourg, France) or SIV Nef (monoclonal antibody 17.2), incubated with HRP-conjugated anti-rabbit immunoglobulin, and developed using ECL substrate. A fraction (1/20) of the cell lysates was loaded directly on gels to visualize nef expression in the transfected cells (TL). (B) As for panel A, using the GST-Hck SH2 and SH3 recombinant proteins. (C) Purified GST-Lck SH2 and SH3 fusion proteins were reacted with purified HIV-1 Nef and SIV Nef produced in bacteria. Following coprecipitation using glutathione-Sepharose beads, the proteins were electrophoresed, transferred to nitrocellulose, and immunoblotted with antibodies specific for HIV-1 Nef or SIV Nef. Sizes in panels B and C are indicated in kilodaltons. (D) Purified recombinant SIV Nef was preincubated with cytoplasmic extracts derived from monocytes, prepared as described previously (19a), before incubation with GST-Hck SH2, GST-Hck SH3, or GST as a control. Following incubation, the GST fusion proteins and any associated proteins were coprecipitated by using glutathione-Sepharose beads. The coprecipitates were separated by SDS-PAGE and transferred to nitrocellulose filters. The filters were then immunoblotted with anti-SIV Nef followed by incubation with HRP-conjugated anti-mouse immunoglobulin and development using ECL substrate.

SH2 domain binding by HIV-1 and SIV Nef.

SIV-Nef is phosphorylated on tyrosine residues upon coexpression with Src, and a putative SH2 binding motif specific for Src family SH2 domains was identified in the amino terminus of SIV Nef (11, 29). We have previously reported that the SH2 domain of Lck can bind HIV-1 Nef and cooperates synergistically with the SH3 domain of Lck for binding to cell-derived HIV-1 Nef (9, 12). Strikingly, cell-derived SIV Nef also interacted with GST fusion proteins containing the SH2 or the SH2 and SH3 domains of Lck (Fig. 5A, lanes 3 and 5). However, in contrast to HIV-1 Nef, no or low cooperation was found between Lck SH2 and SH3 domains for SIV Nef binding, which further supports the minor contribution of Lck SH3 to allow binding of SIV Nef to Lck.

Interestingly, recombinant purified HIV-1 and SIV Nef proteins derived from E. coli failed to interact with the GST-Lck SH2 fusion protein (Fig. 5C), suggesting the role of a third cellular partner which may act as an intermediate docking protein for Nef-Lck SH2 binding or posttranslational modification of Nef by such events as phosphorylation. Furthermore, the Hck SH2 fusion protein specifically coprecipitated SIV Nef which had been preincubated with cytoplasmic extracts derived from purified monocytes (Fig. 5D). Neither HIV-1 nor SIV Nef expressed during transfection of Jurkat cells interacted with Hck SH2, which also suggests the role of a third cellular partner, present or active only in monocytes (Fig. 5B). Together, these results show that SIV Nef can interact with the SH2 domain of Src family kinases, yet this interaction involves at least one additional cellular partner, which differs for Lck and Hck SH2 binding.

The amino terminus of SIV Nef is sufficient for binding to Lck and Hck and augments their catalytic activities.

The amino termini of HIV-1 and SIV-Nef proteins have been reported to represent binding sites for Lck, in addition to the proline motif. We investigated whether the N-terminal 50 amino acid residues of SIV Nef (GST–SIV Nef 1–50) can also support binding to Hck and whether this region of SIV Nef is sufficient for augmenting Hck and Lck catalytic activities. For the expression of GST–SIV Nef 1–50, plasmid p239SpE3′ was used as the template for the PCR amplification of SIVmac239 nef. To obtain the sequence encoding the first 50 amino acid residues of SIV Nef to be cloned into pGEX 4T-1, the nef gene was amplified by PCR by using the 5′ primer 5′-CGGGATCCATGGGTGGAGCTATTTCCATGAGG and the 3′ primer 5′ - ATAAGAATGCGGCCGC TCACAAGCCC T TGTCTAATCC, which introduced the unique BamHI (5′ primers) and NotI (3′ primer) restriction sites. The PCR fragment was digested with BamHI and NotI and cloned into pGEX 4T-1 digested with the same enzymes. For the expression of HIV-1 Nef 1–57, which contains the first 57 amino acid residues of HIV-1 Nef, the molecular clone NL4-3 was used as the template for the PCR amplification of HIV-1 nef. To obtain the sequence corresponding to amino acid residues 1 to 57 of Nef to be cloned into pGEX 4T-1, the nef gene was amplified by PCR by using the 5′ primer 5′-GCTCCGGATCCATGGGTGGCAAGTGGTCAAAAAG and the 3′ primer 5′-ATAAGAATGCGGCCGC TCACCAGGCACAAGCAGCAT TG T TAGC, which introduced the unique BamHI (5′ primer) and NotI (3′ primer) restriction sites. To construct GST–HIV-1 Nef 1–57, the PCR fragment was digested with BamHI and NotI and cloned into pGEX 4T-1 digested with the same enzymes. All vectors were sequenced to ascertain precise nef sequence and in the case of nef sequences cloned into pGEX 4T-1 to ensure that the nef sequences were in frame with the sequence coding for GST. The proteins were expressed and purified as described above. When GST–SIV Nef 1–50 was reacted with cytoplasmic extracts derived from Jurkat cells or monocytes and immobilized on glutathione-Sepharose beads, Lck and Hck, respectively, were detected by immunoblotting in the coprecipitates (Fig. 6A). GST alone did not coprecipitate either Lck or Hck (Fig. 6A). These data indicate that the N terminus of SIV Nef can direct the recruitment of Lck and Hck.

FIG. 6.

The N terminus of SIV Nef binds to and activates Lck and Hck. (A) Cell extracts from purified monocytes were reacted with GST or GST–SIV Nef 1–50, containing the first 50 amino acid residues from SIV/mac239 Nef. Following incubation with the cell lysates, the GST recombinant proteins were coprecipitated by using glutathione-Sepharose beads. The coprecipitates were then separated by SDS-PAGE, transferred to nitrocellulose, and immunoblotted with polyclonal antibodies specific for Hck, followed by incubation with HRP-conjugated anti-rabbit immunoglobulin and development using ECL substrate. (B) Purified GST-HIV-1 Nef, GST–HIV Nef 1–57, GST-SIV Nef, GST–SIV Nef 1–50, or GST as a control was coated onto the wells of 96-well polystyrene microtiter plates (100 nM). After coating and blocking with gelatin, the wells were incubated with increasing amounts of Lck (0 to 300 nM). Binding of Lck was detected with an anti-Lck antibody or as a control an irrelevant antibody at the same immunoglobulin concentration as anti-Lck. Detection of binding was performed essentially as described above except that a biotin-conjugated anti-rabbit immunogblobulin (Amersham) was used. Results are graphed after subtraction of background binding obtained with the irrelevant antibody control. (C) The kinase activity of Lck or Hck derived by immunoprecipitation from Jurkat cells or monocytes, respectively, was measured by using p34^cdc2 peptide as described above. Lck and Hck precipitates were preincubated with GST, GST–HIV-1 Nef, GST–HIV-1 Nef 1–57, GST-SIV Nef, or GST–SIV Nef 1–50 and then added to the peptide p34^cdc2[Lys 19 (6–20)NH₂] or p34^cdc2[Lys 19, Phe 15 (6–20)NH₂; control peptide] followed by kinase assay. Incorporation of [³²P]ATP was measured by scintillation counting. Results are expressed as a percentage of untreated kinase activity after subtraction of the control peptide from the test samples.

In parallel experiments, GST–SIV Nef 1–50 and a GST fusion N-terminal fragment of HIV-1 Nef corresponding to amino acid residues 1 to 57 were coated onto the wells of polystyrene microtiter plates, and the binding of purified Lck was assessed in the solid-phase binding assay described above. Full-length Lck bound directly to immobilized purified GST–SIV Nef 1–50 protein in a concentration-dependent manner (Fig. 6B). Lck did not bind to GST alone, verifying the specificity of the interaction. In contrast to these findings, immobilized GST–HIV-1 Nef fragment corresponding to amino acid residues 1 to 57 did not support binding of Lck. Hence, while the N-terminal region of HIV-1 Nef does not bind directly to Lck, the N-terminal region of SIV Nef supports direct association of this kinase. Inclusion of the N-terminal fragment of SIV Nef into the assays which measure the catalytic activities of Lck and Hck showed that the N-terminal region of SIV Nef was able to augment the catalytic activities of both kinases, albeit not as efficiently as its full-length counterpart (Fig. 6C). GST-SIV Nef 1–50 specifically augmented the cell-derived Lck and Hck phosphotransferase activity in a dose-dependent manner, up to 280 and 340%, respectively, above basal levels, compared to the control GST protein (Fig. 6C). In contrast, GST–HIV-1 Nef 1–57 did not affect either Lck or Hck catalytic activity (Fig. 6C).

We have shown in vitro that both SIV Nef and HIV-1 Nef can bind directly to both Lck and Hck. In the case of at least Lck, the interaction between SIV Nef and Lck is direct. However, our study demonstrates fundamental differences between the Nef proteins of HIV-1 and SIV. Unlike the case for its HIV-1 counterpart, significant binding of SIV-Nef to the Src family kinases was independent of its the proline motif which is highly conserved among HIV-1 and SIV isolates. Instead, the N-terminal region of SIV Nef supported significant binding to both Lck and Hck, corroborating published findings that the N termini of both HIV-1 and SIV Nef are involved in Src family kinase interaction (4). However, unlike the N terminus of HIV-1 Nef, the N-terminal region of SIV Nef supported direct interaction with at least Lck, further illustrating the distinct characteristics of HIV-1 and SIV Nef-Src family kinase interactions. Although both HIV-1 Nef and SIV Nef augment Hck catalytic activity, they do so by discordant mechanisms: HIV-1 Nef via its proline motif binding to SH3 domain and SIV Nef independently of its proline motif through an apparent SH3-independent mechanism. Further, binding of HIV-1 Nef to Lck inhibits its catalytic activity, while SIV Nef stimulates Lck.

That HIV-1 Nef and SIV Nef differ in Src kinase SH3 binding is surprising since both proteins have a conserved proline motif and are generally thought to possess similar functional properties, including an ability to enhance viral replication and infectivity (1, 39), to modulate cell surface receptors (5, 37), and to alter cellular signal transduction pathways (22). Interestingly, however, SIV Nef appears to augment viral infectivity and down-regulate MHC class I molecules less efficiently than HIV-1 Nef (37, 39). Furthermore, while mutation of the proline motif within Nef disrupted the ability of HIV-1 Nef to bind Hck (7) and to alter T-cell signalling (22), the corresponding mutation in SIV Nef did not alter this function (22). Structural studies have allowed the identification of residues in the folded HIV-1 Nef molecule required for high-affinity binding to Fyn SH3 (2, 26). Some of these residues present in the αB helix are altered in both SIV and HIV-2 nef variants (Fig. 2), thereby providing an attractive explanation for altered binding to Hck and Lck SH3 domains. This observation supports a role for elements additional to the PPII helix in the Nef-SH3 interaction. Recently, Lang et al. reported that the proline motif of SIVmac239 Nef was dispensable for viral propagation in vivo (25). Together with the results reported herein, these data suggest alternative mechanisms for Src family kinase binding. SIV- and HIV-1 Nef proteins differ chiefly at the amino terminus, a flexible portion of the molecule in solution (20) which was not resolved by crystallography (2, 26). This portion of both SIV and HIV-1 Nef proteins was reported to mediate indirect interactions with Lck and an unidentified serine kinase (4). This region may play a more prominent role in Src family kinase binding by SIV Nef than HIV-1 Nef. Similarly, in the yeast two-hybrid system, SIV Nef and HIV-1 Nef appear to interact differently with the μ1 and μ2 subunits of adapter protein complexes regulating cellular receptors endocytosis (27). Collectively and together with the present results, these data argue that although HIV-1 Nef and SIV Nef may use similar cellular targets to achieve similar functions, the two proteins have evolved different mechanisms of binding.

SH2 ligation by a specific phosphotyrosine-containing peptide can increase phosphotransferase activity (33). SH2 binding by Nef may similarly be responsible for kinase activation and may account for the opposite effects of HIV-1 Nef and SIV Nef on Lck. Although recombinant SIV Nef and HIV-1 Nef did not bind the SH2 domain of Lck unless cell extracts were added, both proteins profoundly altered Lck activity, albeit differentially. It is possible that as SIV Nef activates Lck, it may also act as a substrate for this kinase and promote SH2 binding, while HIV-1 Nef, which inhibits Lck kinase activity, would not be expected to be a substrate for this kinase. In support of this hypothesis, SIV Nef was found to be phosphorylated on tyrosine residues when expressed in COS cells together with cotransfected Src kinases (11, 29), whereas we have previously reported that HIV-1 Nef can interact with Lck SH2 in a phosphotyrosine-independent manner (12). Furthermore, SIV Nef bound to the SH2 domain only after it was treated with cell lysates, suggesting that modification of Nef or another protein is required for the interaction. At least in the case of Lck, only SIV Nef and the kinase are required for interaction, thus favoring the first hypothesis that modification of Nef by the kinase itself mediates interaction. As the N-terminal region of SIV Nef, which contains two tyrosine residues, binds to and regulates both Lck and Hck, this region of Nef, when phosphorylated, may represent the SH2 interactive domain leading to activation of the kinases. The present data showing activation of Lck by SIV Nef markedly contrasts with the dramatic inhibitory effect observed with its HIV-1 counterpart (9, 18, 19a). These data illustrate the potential for opposite effects by HIV-1 and SIV Nef proteins on T-cell signalling events, particularly those emanating from the T-cell receptor, and suggest that the two proteins may have different roles in the pathogenesis of HIV-1 and SIV infection. These differences have been borne out in recent work describing activation of the T-cell receptor pathway by the aggressive strain of SIV, SIVpbj14 (11, 29), while this pathway is inhibited by HIV-1 Nef (8–10, 18, 30).

Precedence exists for closely related proteins having different specificities for Src family kinase binding and modulation. Indeed, despite their close relatedness, the middle-T proteins of mouse and hamster polyomaviruses display such differences (31). Although the mouse polyomavirus mT antigen (MomT) binds Src, Yes, and Fyn equally, it dramatically increases Src and Yes activity but does not increase Fyn activity. While the kinase domain of Fyn is sufficient for association with MomT through an undefined region, the hamster mT antigen requires the Fyn SH2 domain for binding but does not affect Fyn activity. That HIV-1 and SIV have evolved different strategies to target Src kinases represents a novel example of the genome plasticity of retroviruses to subvert host cell proteins to their own replicative advantage.