(Received for publication, December 6, 1996, and in revised form, February 6, 1997)
From the MRC Retrovirus Research Laboratory,
Department of Veterinary Pathology, University of Glasgow, Glasgow G61
1QH, Scotland, United Kingdom and § University Clinical
Biochemistry, University of Southampton School of Medicine, Southampton
General Hospital, Southampton SO16 6YD, United Kingdom
Stable human cell lines expressing the human immunodeficiency virus type I (HIV-I) Nef protein from inducible promoters were used to analyze the phosphorylation status of Nef in vivo. Nef phosphorylation in both HeLa and Jurkat cells was stimulated by phorbol ester treatment. Phosphoamino acid analysis revealed a predominance of phosphoserine with a small proportion of phosphothreonine. Treatment of cells with selective protein kinase inhibitors revealed that Nef phosphorylation was markedly reduced by bisindolylmaleimide, an inhibitor of protein kinase C, but was unaffected by inhibitors of mitogen-activated protein kinase kinase or cAMP-dependent kinase. These data implicate protein kinase C in Nef phosphorylation in vivo, and thus confirm and extend earlier in vitro data. Phosphorylation of a nonmyristoylated Nef mutant was impaired, suggesting that membrane targeting of Nef was required for phosphorylation. This was expected given that activated protein kinase C translocates from the cytosol to the plasma membrane. However, analysis of the subcellular localization of phosphorylated wild-type Nef revealed that both the cytosolic and membrane-associated pools of Nef were phosphorylated to an equivalent extent. Thus the significance of myristoylation for Nef function may be in influencing protein conformation, although these data could be explained by a transient and dynamic interaction between myristoylated Nef and the plasma membrane.
The HIV-I1 Nef protein is a 206-amino acid polypeptide co-translationally modified at the N terminus by the addition of a myristate residue (1). This modification is, at least in part, responsible for the association of a proportion of Nef with cellular membranes (2, 3). Nef has been shown to have a number of biological functions in vitro that may be pertinent to its role in disease. In particular the presence of Nef in HIV-I-infected cells enhances the production of infectious virions (4, 5), and this has been recently shown to involve the incorporation of Nef into the virus particle and its subsequent cleavage by the viral protease (6, 7). However the precise biochemical mechanisms underpinning the enhanced infectivity of Nef+ virions remains to be defined. Nef expression also results in the down-modulation of CD4 from the cell surface (8, 9), and this function is apparently independent of its role in enhancing viral infectivity (10). Nef-mediated down-modulation of CD4 results from increased rates of CD4 endocytosis and lysosomal targeting/degradation (11-13), although again the precise biochemical mechanisms remain obscure. Finally Nef has been reported to have effects on signal transduction pathways both in lymphocytes and other cell types (14-17). In this context effects of Nef on induction of interleukin-2 synthesis have been observed by some workers (18-20), but there is little agreement in this area as to the precise effects of Nef and the potential role of this aspect of Nef function in viral replication is currently unclear.
In an attempt to unravel the biochemical mechanisms of Nef function, a
number of groups have analyzed the cellular proteins with which Nef
interacts, both in vitro and in vivo. A
significant proportion of these Nef-interacting proteins have been
identified as protein kinases of both the tyrosine and serine/threonine
families. A proline-rich motif (amino acids 70-79) has been shown to
interact with the SH3 domains of the Src family tyrosine kinases Hck
(21, 22) and Lck (20, 23, 24). In addition a peptide corresponding to
this motif inhibited the in vitro binding of Nef to MAP
kinase (20, 24). Binding of Nef to both Lck and MAP kinase has been shown to inhibit kinase activity (23, 24). Nef has also been shown to
associate with a 65K serine/threonine kinase that has been shown to be
a member of the p21-associated kinase family (25, 26), and recently an
association between Nef and PKC has been reported (27). However,
none of these associations has been shown to result in the
phosphorylation of Nef by the interacting kinase. An early study
demonstrated that Nef was phosphorylated in recombinant vaccinia
virus-infected BHK21 cells and that this phosphorylation could be
stimulated by treatment of cells with phorbol ester (8). This study
also showed that partially purified bacterially expressed Nef could be
phosphorylated in vitro by purified PKC. We (28) and others
(29) have subsequently extended the latter observation, demonstrating
that bacterially expressed glutathione S-transferase-Nef
fusion proteins purified to homogeneity on glutathione-agarose beads
could be phosphorylated on serine and threonine residues, both by
purified PKC and PKC present in lysates from mammalian cells. We now
demonstrate that Nef expressed in human HeLa and Jurkat cells is
phosphorylated in vivo. This phosphorylation was stimulated
by phorbol ester treatment of cells and was inhibited by a selective
PKC inhibitor, but not by inhibitors of MAP kinase kinase or
cAMP-dependent kinase. Phosphorylation of Nef did not
demonstrate an absolute dependence on association of Nef with
cytoplasmic membranes as a nonmyristoylated mutant was also
phosphorylated, albeit with reduced efficiency. Additionally, phosphorylation did not affect the observed distribution of Nef between
cytosolic and membrane fractions, as both populations of Nef were
equally efficiently phosphorylated.
The construction of HeLa and Jurkat cells expressing Nef from tetracycline- or heavy metal-responsive promoters has been described previously (30). Construction of the nonmyristoylated Nef mutant has been described (31). Cells were maintained in DMEM (HeLa) or RPMI 1640 (Jurkat) containing 10% fetal calf serum, supplemented with 1 µg/ml tetracycline in the case of the HeLa cells to repress the tet-responsive promoter (32).
ChemicalsAll chemicals and reagents were purchased from
Sigma, with the exception of
4-(2-aminoethyl)benzenesulfonylfluoride-HCl (AEBSF), bisindolylmaleimide-HCl (BIM), and 2-amino-3
-methoxyflavone (PD98059), which were purchased from Calbiochem. The myristoylated PKI
inhibitor peptide was a kind gift from Roger Clegg (Hannah Research Institute, Ayr, Scotland).
For phosphorylation assays HeLa cells were seeded in 50-mm dishes at 5 × 105 cells/dish in the absence of tetracycline. After 40 h, monolayers were washed once in DMEM without sodium phosphate but supplemented with 1% dialyzed fetal calf serum. Cells were then incubated for 4 h at 37 °C in 2 ml of labeling medium containing 200 µCi/ml [32P]orthophosphate (Amersham Corp.) to equilibrate the intracellular ATP pools with labeled phosphate, prior to treatment with 100 ng/ml phorbol 12-myristate 13-acetate (PMA) for 30 min. For experiments involving kinase inhibitors, the inhibitors were added at the same time as the label and were thus present throughout the labeling period. Monolayers were washed twice with ice-cold phosphate-buffered saline and harvested into 1 ml of phosphate-buffered saline containing 10 mM EDTA. Cell pellets were lysed in 500 µl of 10 mM PIPES-NaOH, pH 7.2, 120 mM KCl, 30 mM NaCl, 5 mM MgCl2, 1% Triton X-100, 10% glycerol containing 100 nM okadiac acid, 10 mM sodium fluoride, 2 µg/ml aprotinin, 1 µg/ml leupeptin, 1 µg/ml pepstatin A, and 0.2 mM AEBSF for 30 min at 4 °C. Jurkat cells were treated similarly except that 2.5 × 107 cells were incubated for 24 h in the presence of 100 µM ZnCl2 prior to labeling. Unlabeled lysates (for immunoblotting) were processed in parallel.
Labeling with [35S]methionine was carried out for 4 h in DMEM without methionine but supplemented with 1% dialyzed fetal calf serum and 100 µCi/ml [35S]methionine (Tran35S-label >1000Ci/mmol, ICN). Labeling with 3H-labeled myristic acid was carried out for 16 h in complete DMEM supplemented with 1% dialyzed fetal calf serum and 200 µCi/ml 3H-labeled myristic acid (40-60 Ci/mmol, Amersham Corp.). Myristic acid (supplied in ethanol) was dried down under vacuum and resuspended in Me2SO to a final volume of 1% of the labeling medium.
Immunoprecipitation and ImmunoblottingLysates were
adjusted to 0.5 M KCl and precleared by the addition of 20 µl of protein G-Sepharose beads and incubation on a rotating platform
for 2 h at 4 °C. Nef was immunoprecipitated overnight at
4 °C by the addition of a murine monoclonal antibody specific for
the N-terminal 7 amino acids of Nef (3 µl of ascitic fluid). 10 µl
of protein G-Sepharose beads were added (in 50 µl of lysis buffer)
and incubated for a further 4 h at 4 °C on a rotating platform.
Beads were washed three times in lysis buffer containing 0.5 M KCl and once in lysis buffer prior to addition of 20 µl
of 1 × boiling buffer (0.8% SDS, 8% glycerol, 2%
-mercaptoethanol, 25 mM Tris-HCl, pH 6.8), analysis by
15% SDS-PAGE, and autoradiography. For immunoblotting gels were
transferred to polyvinylidene difluoride membrane (Millipore Immobilon
P) using a Bio-Rad semidry blotting apparatus. After blocking in TBS-T
(25 mM Tris, 137 mM NaCl, 0.1% Tween-20)
containing 10% (w/v) dried skimmed milk, membranes were sequentially
probed with a sheep polyclonal Nef serum (1:10000) and donkey
anti-sheep horseradish peroxidase (Sigma), prior to visualization by
enhanced chemoluminescence (Amersham Corp.).
Cells were harvested as described above and fractionated by hypotonic lysis and Dounce homogenization essentially as described previously (31). Briefly, cells were resuspended in 10 mM PIPES-NaOH, pH 7.2, 0.5 mM MgCl2, swollen on ice for 10 min, and homogenized by 30 strokes with a tight fitting Dounce homogenizer. The suspension was adjusted to 120 mM KCl, 30 mM NaCl, and spun at 500 × g for 5 min at 4 °C. The supernatant from that spin was then spun at 100,000 × g for 30 min at 4 °C. The resulting clarified supernatant was termed cytosol, and the pellet (membrane fraction) was washed in lysis buffer without Triton X-100 and resuspended in lysis buffer. All buffers contained protease and phosphatase inhibitors as described above.
Phosphoamino Acid AnalysisImmunoprecipitates were separated by 15% SDS-PAGE and transferred to polyvinylidene difluoride membrane. The area of the membrane corresponding to the phosphorylated Nef band was identified following overnight autoradiography, excised, and hydrolyzed in 5.7 N HCl at 110 °C for 1 h. The sample was lyophilized and analyzed by two-dimensional thin layer electrophoresis on cellulose plates (Eastman Kodak Co.), in the first dimension at pH 1.9 and in the second dimension at pH 3.5. Unlabeled phosphoamino acid standards (0.5 µg each) were visualized by ninhydrin staining.
We have previously described the construction of stable
cell lines expressing Nef from inducible promoters (30). Stable HeLa
cell lines were generated using a derivative of the tetracycline responsive system originally developed by Gossen and Bujard (32) in
which the components of the system were cloned into the pREP series of
episomal vectors. This system proved to be unsuitable for regulated Nef
expression in Jurkat cells as expression was constitutive, so in these
cells Nef was expressed from a modified metallothionein promoter (33).
Fig. 1 demonstrates the inducibility of Nef expression
in these cells. HeLa cells were incubated with and without tetracycline
(1 µg/ml), and Jurkat cells were treated with 100 µM
ZnCl2 for 24 h prior to harvesting and analysis of Nef
expression by immunoblotting. In both cases Nef expression can be
induced at least 10-fold. In all the phosphorylation experiments described subsequently Nef expression was induced either by growth in
the absence of tetracycline (HeLa cells) or in the presence of 100 µM ZnCl2 (Jurkat cells).
Phosphorylation of Nef Expressed in Human HeLa or Jurkat T Cells Is Stimulated by Phorbol Ester Treatment
Our previous results
indicated that Nef could be phosphorylated in vitro by
kinases present in extracts of HeLa or Jurkat cells (28). To extend
this result and to examine whether Nef could be phosphorylated under
physiological conditions, HeLa and Jurkat cells expressing Nef were
metabolically labeled with [32P]orthophosphate and
lysates were immunoprecipitated with a Nef-specific murine monoclonal
antibody. Fig. 2a demonstrates that in HeLa cell lysates a 28-kDa phosphorylated band corresponding to Nef was
immunoprecipitated by the Nef monoclonal antibody from cells expressing
Nef but not from control cells stably transfected with an empty vector.
Phosphorylation of this species was stimulated at least 8-fold (as
judged by densitometry of autoradiographs) by treatment of cells with
100 ng/ml PMA for 30 min. Fig. 2b demonstrates that
phosphorylation of Nef expressed in Jurkat T cells was also stimulated
by PMA treatment of the cells. Fig. 2c shows an immunoblot of immunoprecipitates from unlabeled lysates, demonstrating that similar levels of Nef were present in lysates from cells, independently of PMA treatment. Thus the increase in phosphate labeling apparent in
the presence of PMA is not due to an increase in Nef expression levels,
but rather by phosphorylation of preexisting Nef molecules. The high
molecular weight phosphorylated proteins present in each lane of Fig.
2, a and b, represent cellular phosphoproteins
that bound nonspecifically to protein G-Sepharose beads. Identical profiles of these proteins were seen in immunoprecipitates from control
and Nef-expressing cells (for example compare Fig. 2a, lanes 1 and 3) and were also seen in control
precipitations carried out using antibodies to unrelated proteins or
protein G-Sepharose beads alone (data not shown).
Nef Phosphorylation Occurs Predominantly on Serine Residues
In vitro data demonstrated that Nef
phosphorylation by kinases present in cell extracts occurred
predominantly on serine residues; however, phosphothreonine could also
be detected at much lower abundance (28). Our preliminary analysis of
Nef immunoprecipitated from [32P]orthophosphate-labeled
HeLa-Nef cells revealed that Nef phosphorylation was entirely
alkali-labile and thus consisted of phosphoserine and/or
phosphothreonine (data not shown). To define the identity of the
phosphoamino acids precisely, Nef immunoprecipitates from PMA-treated
HeLa-Nef cells were transferred to polyvinylidene difluoride membranes,
and the radiolabeled bands corresponding to Nef were excised and
analyzed by hydrolysis in 5.7 N HCl followed by
two-dimensional thin-layer electrophoresis. Fig. 3
demonstrates that, in common with the earlier in vitro data,
Nef was phosphorylated principally on serine, although phosphothreonine
was also detected. Identical results were obtained after phosphoamino
acid analysis of phosphorylated Nef expressed in Jurkat cells (data not
shown).
Inhibition of Nef Phosphorylation by a Selective Inhibitor of PKC
Both the conventional and novel groups of protein kinase C
isoforms are activated by binding to phorbol esters (34). The observation in Fig. 2 that phosphorylation of Nef was stimulated by PMA
treatment suggested that PKC may either directly phosphorylate Nef or
phosphorylate and activate a distinct kinase that subsequently phosphorylates Nef. A number of other lines of evidence point to PKC as
a direct effector of Nef phosphorylation; in vitro data from
a number of laboratories has indicated that purified PKC can
phosphorylate purified Nef (8, 28, 29), and in addition our previous
data demonstrated that only kinase inhibitors selective for PKC
suppressed Nef phosphorylation in vitro by kinases present in cell extracts (28). To investigate whether PKC plays a role in Nef
phosphorylation in vivo HeLa and Jurkat cells expressing Nef
were treated with a number of selective kinase inhibitors, prior to PMA
stimulation and analysis of Nef phosphorylation. As shown in Fig.
4 treatment of cells with a cell-soluble inhibitor of
PKC, BIM, dramatically reduced the levels of PMA-induced Nef phosphorylation in comparison to the level of phosphorylation detectable in the absence of inhibitor. Neither PD98059, a synthetic inhibitor of MAP kinase kinase (35), or a myristoylated peptide inhibitor of cAMP-dependent kinase, corresponding to
residues 5-24 of the heat-stable inhibitor PKI (36) had any
significant effect. These data confirm the relevance of our previous
in vitro data and attest to a role for PKC in Nef
phosphorylation in vivo.
Nef Phosphorylation Is Enhanced by Myristoylation
As PKC is
known to translocate from the cytosol to the plasma membrane once
activated by phorbol ester binding, it was conceivable that
phosphorylation would require membrane localization of Nef. Myristoylation is absolutely required for the stable association of Nef
with a cytoplasmic membrane fraction, therefore HeLa and Jurkat cells
expressing a nonmyristoylated Nef mutant (Gly2 Ser)
were utilized to address this question. To confirm that this mutant was
indeed nonmyristoylated, HeLa cells were metabolically labeled with
either [3H]myristate or [35S]methionine and
immunoprecipitated using the N-terminal-specific Nef monoclonal
antibody. Fig. 5a shows that this Nef mutant
could be labeled with [35S]methionine, but failed to
incorporate [3H]myristate. In contrast wild-type Nef was
myristoylated as demonstrated by the incorporation of
[3H]myristate. The intracellular localization of the
wild-type and nonmyristoylated Nef proteins was analyzed by subcellular
fractionation of HeLa cells expressing each protein. Cytosolic and
membrane-associated fractions of these cells were analyzed by
immunoblotting with a Nef specific sheep polyclonal serum (Fig.
5b). The data confirm that a proportion of wild-type Nef
molecules was stably associated with the membrane fraction; however, it
is clear that more than 50% of myristoylated Nef molecules remained
cytosolic. In comparison the nonmyristoylated Nef mutant failed to
associate with the membrane fraction and was exclusively cytosolic.
Identical results were obtained for expression and localization of the
wild-type and nonmyristoylated Nef mutant in Jurkat cells (data not
shown).
HeLa and Jurkat cells expressing wild-type and nonmyristoylated Nef
were then labeled with [32P]orthophosphate and Nef
phosphorylation was analyzed by immunoprecipitation. Fig.
6a demonstrates that a phosphorylated species
corresponding to the nonmyristoylated Nef mutant could be detected;
however, in comparison with wild-type Nef the extent of phosphorylation was clearly greatly reduced. Unlabeled extracts immunoprecipitated in
parallel and analyzed by immunoblotting revealed that comparable amounts of the two Nef species were present in immunoprecipitates (Fig.
6b), although it should be noted that in the HeLa cell lines the nonmyristoylated mutant was expressed at a lower level than the
wild-type (see Fig. 5a). Thus it is clear that, although
phosphorylation does not demonstrate an absolute requirement for
myristoylation, the latter modification greatly enhances the ability of
Nef to function as a phosphorylation substrate.
It has been reported that the distribution of some myristoylated
proteins such as MARCKS (37) and HIV-I p17Gag (38) is regulated by
phosphorylation. These proteins are stably anchored into the plasma
membrane by a combination of myristoylation and electrostatic
interactions between basic amino acids and acidic phospholipids.
Phosphorylation results in their dissociation from the membrane by
electrostatic repulsion. To investigate whether this situation might
also apply to Nef, HeLa cells expressing wild-type Nef were labeled
with either [32P]orthophosphate or
[35S]methionine and fractionated into cytosolic and
cytoplasmic membrane fractions prior to immunoprecipitation. The
results of this experiment are shown in Fig. 7; it can
clearly be seen that the majority of both
[32P]orthophosphate (Fig. 7a) or
[35S]methionine (Fig. 7b) labeled Nef was
present in the soluble fraction. Thus it appears that both the
cytosolic and membrane-associated populations of Nef were
phosphorylated to an equivalent extent, indicating that phosphorylation
of Nef does not grossly influence its subcellular localization.
Furthermore these data imply that stable membrane association of
myristoylated Nef is not required for phosphorylation, with the proviso
that the interaction between Nef and the membrane may be transient and
therefore dynamic.
In this study we report that the HIV-I Nef protein can be phosphorylated in vivo by a cellular serine/threonine protein kinase, and we present evidence to suggest that this phosphorylation is mediated by members of the PKC family. The evidence for the latter statement is 2-fold. First, Nef phosphorylation is stimulated by treatment of Nef expressing cells with the phorbol ester PMA, a well characterized activator of PKC (Fig. 2a). Second, Nef phosphorylation was reduced by a selective PKC inhibitor, BIM, but not by inhibitors of cAMP-dependent kinase or MAP kinase kinase (Fig. 4). Although previous in vitro data both from our laboratory and others have indicated that purified PKC can directly phosphorylate recombinant Nef in the absence of other cellular factors (8, 28), it cannot be ruled out that in vivo PKC acts indirectly to induce Nef phosphorylation by activating a distinct protein kinase. In this regard it should be noted that BIM treatment did not completely abrogate phosphorylation of Nef, but merely reduced it to levels comparable to those observed in unstimulated cells (in the absence of PMA). One explanation for this observation is that the constitutive phosphorylation of Nef is mediated by a distinct kinase activity that is itself stimulated by PKC.
Activation of PKC by phorbol ester binding is concomitant with its translocation from the cytosol to the plasma membrane where it inserts into the lipid bilayer (34). The majority of PKC substrates are thus transmembrane glycoproteins (e.g. epidermal growth factor, transferrin, and insulin receptors) or proteins associated with the cytoplasmic face of the plasma membrane (e.g. MARCKS, c-Src, Raf-1). The observation that Nef phosphorylation is enhanced by myristoylation (which is absolutely required for membrane association of Nef) could therefore be construed as further evidence that Nef is directly phosphorylated by PKC. However, two lines of evidence suggest that Nef phosphorylation was not absolutely dependent upon stable association with cytoplasmic membranes. First, the nonmyristoylated mutant was phosphorylated, albeit at a much lower level than wild-type (Fig. 6a). Second, wild-type Nef present both in the cytosolic and membrane fractions was phosphorylated to an equivalent extent (Fig. 7). There are two simple explanations for these data. First, by analogy with the catalytic subunit of cAMP-dependent kinase (40), myristoylation might influence the tertiary structure of the protein such that a PKC site(s) was exposed. In support of this conjecture structural studies of bacterially expressed Nef have indicated that the nonmyristoylated N-terminal domain lacked any defined structure (41, 42). Phosphorylation might therefore be occurring independently of membrane association, and the enhancement by myristoylation would merely reflect altered conformation of the protein. Second, the interaction between Nef and the membrane might be dynamic; however, in contrast to proteins such as MARCKS and c-Src (43) the data presented here suggest that association of Nef with the membrane is not dynamically regulated by phosphorylation.
The data in Fig. 3 demonstrate that both phosphoserine and phosphothreonine could be detected in immunoprecipitates of phosphorylated Nef, although the major component was phosphoserine. Two studies have reported tyrosine phosphorylation of both HIV-I (23) and simian immunodeficiency virus Nef (44); however, in these experiments described here we found no evidence for the presence of phosphotyrosine in phosphorylated Nef species. In repeated experiments phosphoserine was invariably detected; however, phosphothreonine was not always present. It is possible therefore that phosphothreonine might represent a contaminating cellular protein, as in some experiments a phosphorylated protein co-migrating with Nef was observed following prolonged exposures of immunoprecipitations with the Nef-specific monoclonal antibody from control cells (data not shown). The identity of the phosphorylated serines/threonines in Nef remains to be established. Residues phosphorylated by PKC tend to be preceded and/or followed by arginine or lysine residues (45), there are a number of serine residues and one threonine in the BH10 Nef allele used in this study (46, 47) that fulfil these criteria and therefore constitute potential PKC phosphorylation sites. It is important to note that an early study suggested that Nef was phosphorylated by PKC exclusively on threonine 15 (8); however, the BH10 allele of Nef used in this study has a substitution of alanine for threonine 15, so this residue is clearly not absolutely required for PKC phosphorylation of Nef in vivo. Experiments are presently underway to identify the sites of phosphorylation by site-directed mutagenesis and phosphopeptide mapping.
Our previous data have demonstrated that the nef allele used in this study (BH10) encodes a protein functionally active in both HeLa and Jurkat cells, as judged by a number of criteria including enhancement of viral infectivity (30), CD4 down-modulation (30), and inhibition of AP-1 transcription factor induction.2 It is likely, therefore, that phosphorylation plays a role in the function of Nef. That role remains elusive; however, one immediate functional consequence is that phosphorylation will influence the interactions between Nef and cellular proteins. There are many pertinent examples of phosphorylation-dependent protein-protein interactions, for example MARCKS binding to calmodulin is inhibited by PKC phosphorylation (37), and in some instances phosphoserine has been implicated in phosphotyrosine-independent interactions with SH2 domains (48, 49). Further insights into the functional consequences of Nef phosphorylation must await the results of future experiments with mutated genes.
We thank Professor Jim Neil for critical reading of this manuscript.