(Received for publication, November 7, 1996)
From the Cold Spring Harbor Laboratory, Cold Spring Harbor, New York 11742
We demonstrate that the interferon-induced, double-stranded (ds) RNA-activated kinase, PKR, is able to bind to and phosphorylate the human immunodeficiency virus type 1 (HIV-1) trans-activating protein, Tat. Furthermore, Tat can inhibit the activation and activity of the kinase. Phosphorylation of Tat by PKR is dependent on the prior activation of PKR by dsRNA and occurs on serine and threonine residues adjacent to the basic region important for TAR RNA binding and Tat function. Activated PKR efficiently phosphorylates both the two-exon form of Tat (Tat-86) and the single exon form (Tat-72). Mutagenesis indicates that the interaction between PKR and Tat requires the RNA-binding region of Tat. Tat competes with eukaryotic initiation factor 2, a well-characterized substrate of PKR, for phosphorylation by activated PKR. Tat also inhibits the autophosphorylation of PKR by dsRNA. This biochemical evidence of an intimate relationship between Tat, an important regulator of HIV transcription, and PKR, a pleiotropic cellular regulator, may provide insights into HIV-1 pathogenesis and, more generally, virus/host interactions.
The human immunodeficiency virus type 1 (HIV-1)1 tat gene product
trans-activates viral gene expression and is essential for HIV-1
replication (1-3). Tat strongly activates transcription from the HIV-1
long terminal repeat by binding to the Tat-responsive region (TAR), an
RNA stem-loop structure located at the 5 end of HIV transcripts (4).
Although the precise mechanism by which Tat exerts its effect is not
yet known, it has been established that Tat regulates transcription at
the level of initiation and elongation (5, 6). The Tat protein exists
in two forms, which in the HXB2 viral isolate consists of 72 and 86 amino acids. The 86-amino acid protein (Tat-86) is encoded by two
exons, whereas the 72-residue protein (Tat-72), which is identical
except for lacking 14 residues from the C terminus, is the product of
the first tat exon. The shorter form is sufficient for
trans-activation (7). The second exon of Tat has been proposed to play
a role in activation of integrated long terminal repeats, regulation of
MHC class 1 gene promoter activity, and TAR-independent
trans-activation (8-10). Mutational analysis of Tat has revealed three
major regions that are important for function (Fig. 1);
these include the N terminus, the cysteine rich-region which is
important for metal binding, and a charged, arginine-rich region
important for nucleolar localization and for binding to the cis-acting
TAR element (11-13).
One of the primary cellular responses to viral infection is the
production of interferon (14). The RNA-dependent protein kinase PKR, also referred to as DAI, P1 kinase, and p68 kinase, is a
serine/threonine kinase that is induced by interferon and activated in
the presence of dsRNA. PKR exerts a well-established regulatory effect
on initiation of protein synthesis. Activation of PKR by dsRNA, or
polyanions such as heparin and some structured single-stranded RNAs
(15), is accompanied by autophosphorylation (16). Following its
activation, PKR in turn catalyzes the phosphorylation of the subunit of eukaryotic initiation factor 2 (eIF2) on a serine residue at
amino acid position 51 (17, 18). Phosphorylation of eIF2 results in the
sequestration of a second initiation factor, the guanosine nucleotide
exchange factor eIF2B, leading to the inhibition of protein synthesis
(18). This mode of translational shut down provides a mechanism of host
defense and, as such, is detrimental to the viral life cycle. Many
viruses have developed strategies to circumvent the action of PKR
activation. The mechanisms by which viruses prevent the action of PKR
vary as follows: adenovirus (19), vaccinia virus (20), and influenza
virus (21) directly reduce PKR activity via different means, whereas
poliovirus infection leads to PKR degradation (22). PKR appears to be
down-regulated by HIV-1 (23), but the mechanism remains to be
elucidated. PKR has also been implicated in oncogenic transformation
and tumorigenesis (24, 25) as well as differentiation (26) and
apoptosis (27). The substrate specificity of PKR has been shown to
extend beyond eIF2 to include I
B (28, 29) an inhibitor of the
transcriptional activator NF
B, as well as histone H2A (30), and a
90-kDa protein found in rabbit reticulocytes (31) which can be
phosphorylated by PKR in vitro.
Tat binds a variety of cellular factors including a putative ATPase and
DNA helicase (32), and a 36-kDa nuclear factor (33), as well as the
transcription factors TFIID (34) and Sp1 (35). Several observations
prompted us to evaluate the possibility of an interaction between Tat
and PKR. First, both have been demonstrated to interact stably with TAR
(14, 36). Moreover, the TAR RNA binding protein can interact with PKR,
preventing the activation of PKR (37, 38). Second, recent reports
indicate that Tat binds a novel cellular kinase (39, 40), and
Tat-mediated transcription is sensitive to the kinase inhibitor
5,6-dichloro-1--D-ribofuranosyl benzimidazole (41).
Third, PKR has been shown to activate NF-
B by phosphorylation of its
inhibitor, I
B (28, 29). Previous data indicated that NF-
B plays a
role in Tat-regulated transcription from the HIV-1 long terminal repeat
(42). Finally, the stable expression of Tat in HeLa cells treated with
interferon is associated with reduced levels of PKR (23).
An interaction between Tat and the cellular kinase PKR would present an attractive regulatory mechanism. In this report, we describe the ability of PKR (activated in the presence of dsRNA) and of protein kinase C (PKC) to phosphorylate purified Tat-72 in vitro. We show that activated PKR is able to phosphorylate a series of Tat proteins expressed as glutathione S-transferase (GST) fusions. Binding studies indicate a correlation between phosphorylation and the ability of Tat to bind PKR. Furthermore, Tat competes with eIF2 as a PKR substrate, and preincubation with Tat prevents the activation of PKR by dsRNA. The significance of these interactions as they pertain to viral regulation and mechanisms by which HIV-1 is able to avoid the antiviral effect of interferon are discussed.
The region of Tat encoded by the first exon (amino acids 1-72) of the HIV-1 HXB2 isolate was overexpressed in Escherichia coli (43) and purified to greater than 90% homogeneity as assessed by silver staining, by C4 reverse phase high pressure liquid chromatography (HPLC) as described previously (44).
Expression of GST-Tat Fusion ProteinsPlasmids expressing
glutathione S-transferase (GST) fusions with wild-type Tat-1
86 or Tat-1 72 from the HIV-1 HXB2 isolate were generously donated by
A. Rice. Wild-type Tat-2 130 (ROD isolate), mutant Tat-1, and Tat-2
fusion proteins (Tat-86p18IS, Tat-86C22G, Tat-72p18IS, Tat-72C22G,
Tat-48, Tat-48
p18IS, Tat-48
C22G, Tat-86
2/36, Tat-84
,
Tat-99 8184A, and Tat-99
8/47) were obtained from the NIH AIDS
Research Reagent Program, NIAID (Rockville, MD). GST-Tat constructs
were expressed and purified as described previously (45). Competent
E. coli, strain BL21 (Stratagene, La Jolla, CA), was
transformed with either pGEX2T or pGEX2TK vectors (Pharmacia Biotech
Inc.) containing either wild-type or mutant forms of Tat (see Fig. 1).
A GST construct lacking Tat protein was used as a control. Luria
Bertani (LB) broth cultures (50 ml), supplemented with ampicillin
(Sigma) 50 µg/ml, were incubated overnight with shaking at 37 °C.
Overnight cultures were diluted 1:10, and the incubation was allowed to
continue for a further 3 h. Fusion protein expression was induced
by the addition of isopropylthio-
-galactosidase (Sigma) to a final
concentration of 0.1 mM, and growth was continued for a
further 1.5 h. Bacteria were pelleted by centrifugation at 5,000 rpm for 10 min at 4 °C and resuspended in 4 ml of EBC buffer (50 mM Tris, pH 8.0, 120 mM NaCl, 0.5% Nonidet
P-40) containing 5 mM DTT, 2 mg/ml lysozyme, 1 µg/ml
leupeptin, 2 µg/ml aprotinin, and 50 µg/ml phenylmethylsulfonyl
fluoride. After incubation for 15 min on ice and sonication three times
for 30 s on ice, the lysed bacteria were centrifuged at 12,000 rpm
for 15 min at 4 °C. Supernatants were transferred to fresh tubes and
stored at
70 °C.
Bacterial extracts (50 µl) were mixed with 250 µl of EBC buffer and 25 µl of glutathione S-Sepharose beads (equilibrated in EBC buffer). Samples were incubated on a rocking platform for 30 min at 4 °C, centrifuged for 10 s at 12000 rpm in a microcentrifuge, and the supernatant discarded. Pelleted complexes were washed twice with 500 µl of EBC buffer containing DTT (5 mM) and SDS (0.075%). Precipitates were resuspended in 20 ml of thrombin cleavage buffer (50 mM Tris-HCl, pH 7.6, 20 mM KCl, 1 mM DTT). Suspensions were centrifuged at 3000 rpm for 3 min at room temperature. Supernatants were discarded, and pellets were resuspended in 100-µl volumes of cleavage buffer. Two units of human thrombin (Sigma) were added, and the reactions were allowed to proceed for 3 h at room temperature. Following the incubations, reactions were centrifuged in a microcentrifuge at 3000 rpm for 3 min and the supernatants collected. The concentration of Tat protein obtained following thrombin digestion was estimated by running samples of the supernatant on a 15% polyacrylamide gel containing SDS, staining the gel with Coomassie Blue, and comparing the band intensities with those of standard proteins of known concentrations.
Kinase AssaysReactions (20 µl) containing 2.5 µCi of
[-32P]ATP (ICN Biomedical Inc., Costa Mesa, CA) and
0.5 µl of PKR (approximately 5 ng) purified to the mono-S stage (46)
were conducted as described previously (47) in the presence of dsRNA
derived from reovirus (a gift from A. Shatkin). Kinase reactions (20 µl total volume) containing PKC (10 ng) were carried out as described
by the manufacturer (Upstate Biotechnology Inc., Lake Placid, NY).
Kinase reactions containing the
-insulin receptor kinase (a gift
from A. Flint) were performed as described by Villalba et
al. (48). Mammalian and yeast kinases involved in the Ras signal
transduction pathway (STE 20, STE 11, MEK, BYR and ERK) were a gift
from A. Polverino; activation assays were performed as described by
Polverino et al. (49). In each case the enzyme was added
last to the reaction components assembled on ice. Phosphorylation was
visualized using SDS-polyacrylamide gel electrophoresis and
autoradiography at
70 °C with an intensifier screen.
Kinase assays in the
presence of Tat were performed by the addition of 10 µl from a
20-µl PKR activation assay (described above) to 50 ng of purified
Tat-72 or Tat from thrombin-cleaved GST-Tat protein. Substrate
competition assays containing eIF2 and Tat-72 were carried out by
mixing 50 ng of purified eIF2 (a gift from J. Hershey) with increasing
concentrations of purified Tat-72 (up to 500 ng). Concurrently, 50 ng
of purified Tat was mixed with increasing concentrations of purified
eIF2 (up to 500 ng). To each reaction 10 µl volumes from an kinase
assay containing activated PKR was added. Reactions were incubated for
20 min at 30 °C and stopped by the addition of Laemmli sample
buffer. Phosphorylated proteins were resolved in a 20% polyacrylamide
gel containing SDS. Dried gels were exposed at 70 °C to x-ray film
(Eastman Kodak Co.) in the presence of an intensifier screen.
Phosphate-labeled Tat-72 was excised from the gel and processed as described by Beemon and Hunter (50). Labeled protein was digested "in-gel" for 20 h at 30 °C (51) with either trypsin or Achromobacter protease I (final concentration of 1 µg/reaction) and fractionated by HPLC. Fractions containing phosphate-labeled derivatives were subjected to Edman degradation using the modification of Russo et al. (52). Following acid hydrolysis, one-dimensional phosphoamino acid analysis was performed as described by Cooper et al. (53), using electrophoresis on glass thin layer chromatography plates (J. T. Baker, Inc.) in Buffer 3.5 (10:100:1890, pyridine:glacial acetic acid:H2O), for 30 min at 1000 V.
Binding of PKR to TatAliquots (100 µl) containing recombinant GST-Tat protein were mixed with 100-µl volumes of glutathione S-Sepharose beads (equilibrated in EBC buffer) and incubated for 30 min at 4 °C. Complexes were sedimented by centrifugation in a microcentrifuge for 20 s at 12,000 rpm. The pellets were washed five times in EBC buffer containing 5 mM DTT and 0.075% SDS. Activated PKR (10 µl from a 20-µl kinase assay) was added to the GST-Tat, glutathione S-Sepharose bead complexes, and incubated for 20 min at 30 °C. 32P-Labeled PKR that remained bound to GST-Tat was eluted by boiling in sample buffer and resolved in a 15% polyacrylamide gel containing SDS.
The requirement for dsRNA in
the activation and autophosphorylation of PKR is well established (14).
In addition to its role in translational control via the
phosphorylation of the subunit of eIF2 (17, 18), activated PKR has
been shown to phosphorylate histone 2A, I
B, and a 90-kDa protein
found in rabbit reticulocytes (28, 30, 31). To determine whether PKR is able to phosphorylate the HIV-1 trans-activating protein Tat, in
vitro kinase assays were performed using combinations of dsRNA, PKR, and Tat-72 (Fig. 2). The intense phosphate-labeled
band in lane 1 corresponds to the autophosphorylated
(activated) form of PKR; faintly labeled bands correspond to
degradation products of activated PKR. The reaction displayed in
lane 4 was similar to that of lane 1, but was
incubated for a further 20 min in the presence of purified Tat-72. It
contains an additional labeled band with a mobility similar to that of
Tat, suggesting that PKR previously activated in the presence of dsRNA
is subsequently able to phosphorylate highly purified Tat-72.
Phosphorylation of Tat by PKR increased linearly between 0.25 and 25 µg/ml, but the intensity of phosphorylation was reduced at higher Tat
concentrations (data not shown). No phosphorylation of either PKR or
Tat occurred in the absence of either dsRNA (lanes 2 and
5) or PKR (lanes 3 and 6).
Additional tests were conducted to rule out the possibility that these
observations resulted from contamination of either the Tat or PKR
preparations. Initial attempts to confirm that the
32P-labeled protein is indeed Tat, by using antibodies
directed against Tat in immunoprecipitation or immunoblotting
experiments, were unsuccessful. Three different anti-Tat antibody
preparations all failed to react with the phosphoprotein. However,
Tat-72 purified by reverse phase chromatography using a C18 column (6;
data not shown), and both Tat-72 and Tat-86 isolated from GST-Tat
fusion proteins (see Fig. 4), also served as substrates for activated PKR. The latter differ from one another in electrophoretic mobility, thereby eliminating potential contaminants from consideration. Two
other HIV-1 encoded regulatory proteins, Rev and Nef, were not
phosphorylated detectably by PKR (data not shown). A variety of mono-S
fractions containing PKR (46), as well as PKR purified by an
immunoaffinity column containing monoclonal antibody against PKR (54),
all phosphorylated Tat-72 in a dsRNA-dependent manner (data
not shown). Therefore, Tat-72 is a substrate for activated PKR.
Specificity of Tat Phosphorylation
To further assess the
specificity of the interaction between Tat and PKR, several kinases
were tested for their ability to phosphorylate Tat. Activated kinases
were incubated in the presence of 50 ng of purified Tat-72. The results
shown in Fig. 3 (lanes 8, 10, 12, 14, 16, and
18) suggest that neither the -insulin receptor kinase nor
a series of kinases (ST20, ST11, MEK, Byr, and Erk) required for
Ras-regulated signal transduction was able to phosphorylate the
purified form of Tat (lanes 7-18). Confirmation that these
kinases were activated prior to incubation with Tat was achieved by
observing the autophosphorylation of PKR, PKC, BIRK, ST20, and Erk (not
shown). Of the kinases tested, only PKR and PKC phosphorylated Tat
(lanes 1 and 5). As expected, activated PKR and
PKC also phosphorylated histones, histone 2A only in the case of PKR
(lanes 3 and 6).
Several radiolabeled nucleotides were tested to further address the
specificity of phosphorylation of Tat by activated PKR. Both
[-32P]ATP and [
-32P]GTP were utilized
in Tat phosphorylation, but [
-32P]ATP,
[
-32P]UTP, and [
-32P]CTP failed to
phosphorylate Tat in a PKR-dependent manner. However, when
Tat was present at high concentrations (greater than approximately 10 µg/ml), it was nonspecifically labeled by all of these nucleotides even in the absence of PKR or dsRNA. This nonspecifically labeled Tat
comigrated with the unphosphorylated marker Tat, unlike specifically labeled Tat which migrated slightly slower than its precursor (Fig. 7,
lane 8). Since the nonspecific labeling was reduced by addition of excess unlabeled ATP, we tentatively attributed it to an
affinity of the highly basic Tat protein for nucleotides, rather than
to a covalent modification.
Phosphorylation of Tat-86 and Mutant Tat Derivatives by Activated PKR
To address the sequence requirements for Tat phosphorylation,
several mutant Tat proteins were expressed as GST-Tat fusions and
selectively purified by absorption to glutathione
S-Sepharose beads. Tat proteins were isolated from their GST
parent vector by cleavage with thrombin, and equivalent concentrations
of the released recombinant Tat proteins were incubated in the presence of activated PKR (Fig. 4). Cleavage of the GST-Tat
fusion with thrombin was necessary as the intact Sepharose-bound fusion
proteins were not labeled by PKR. Control reactions containing
activated PKR, without and with reverse-phase HPLC purified Tat, are
shown in lanes 1 and 2. PKR phosphorylated the
HPLC purified Tat (lane 2) and the recombinant single exon
form of Tat (Tat-72) released from GST (lane 7). The
full-length, two-exon form of Tat (Tat-86) was also phosphorylated
(lane 4). Furthermore, activated PKR phosphorylated mutant
forms of Tat-86, namely Tat-86p18IS and Tat-86C22G (lanes 5 and 6), and the corresponding mutant forms of Tat-72,
Tat-72p18IS and Tat-72C22G (lanes 7-9). However, the 48
Tat truncation and its mutant derivatives 48
p18IS and 48
C22G
(Fig. 1) were not labeled in the presence of activated PKR (lanes
10-12). As expected, no proteins in the Tat size range were
labeled when activated PKR was incubated with the product of the GST
vector alone (lane 3). From these results it appears that
residues 49-72, containing the basic region, are important for
phosphorylation by PKR. Mutations in the activation domain did not
affect Tat phosphorylation, and in several repeats of this experiment
Tat-86 and Tat-72 were labeled to an equivalent extent.
There are 5 serine residues and 4 threonine residues in Tat-72. To determine which
amino acids on Tat-72 are phosphorylated in vitro by PKR,
labeled Tat was digested with trypsin, and the products were resolved
by HPLC. A single predominant peak resulted (Fig.
5A). One-dimensional phosphoamino acid analysis on material from the HPLC peak showed that the label is associated with both serine
and threonine residues (Fig. 5B). The same material was subjected to sequential Edman degradation, and the release of radioactive derivatives was monitored. Peaks of radioactivity were
observed at cycles 6, 8, and 12, indicative of phosphorylation sites at
these distances following an arginine or lysine residue. The only place
that such a pattern occurs in the Tat-72 molecule is at serine 62, threonine 64, and serine 68 (Fig. 5C). These residues follow
a run of basic residues at positions 49-58; it seems that digestion
took place preferentially after arginine 56, in accordance with
previous observations (55). Digestion with an alternative protease in
place of trypsin corroborated these assignments.
Achromobacter protease I cleaves specifically after lysine
residues, yielding peaks at cycles 11, 13, and 17 following a lysine
residue at position 51 (Fig. 5B). These results indicate
that the phosphorylation sites are clustered in the C terminus
immediately after the basic region of Tat.
Correlation Between the Binding and Phosphorylation of Tat by Activated PKR
To investigate the sequence requirements for the
binding of Tat to PKR, similar amounts of wild-type and mutant GST-Tat
proteins were bound to glutathione S-Sepharose beads and
then incubated with 32P-labeled PKR. Fig. 6
shows that GST alone is unable to bind PKR. GST-Tat 86, GST-Tat 72, and
their mutant variants, p18IS and C22G, all bound PKR, whereas the 48
Tat truncation and its associated mutants failed to bind PKR. In
addition the Tat construct Tat 86
2/36, an N-terminal deletion of
residues 2-36 from GST-Tat 86, successfully bound activated PKR. These
data suggest that the Tat sequence contained between amino acids 49 and
72 is important for binding PKR, but the N-terminal and C-terminal
residues 2-36 and 73-86 are dispensable. In similar experiments with
HIV-2 Tat (see Fig. 1), the intact molecule bound labeled PKR with an
efficiency comparable with that of HIV-1 Tat (GST-Tat 130; Fig. 6).
Removal of residues from the N terminus of HIV-2 Tat had little effect on PKR binding (GST-Tat 99
8/47), whereas a truncation in the basic
region (GST 84
) or the mutation of four consecutive arginine residues in this region to alanines (GST-Tat 99 8184A) abolished binding. These data emphasize the importance of the Tat basic region for PKR binding and raised the possibility that the interaction of the two proteins might be mediated by an RNA bridge. No support for
this idea was obtained from experiments with RNases, however. RNase
treatment of activated PKR, or of the bacterial extract containing
GST-Tat, or of the GST-Tat·PKR complex, had no discernible effect on
the interaction. Both single-stranded RNA (RNase A) and double-stranded
RNA (RNase III) -specific RNases were tested, singly and in combination
(data not shown).
Substrate Competition Between Tat and eIF-2
PKR regulates
protein synthesis by phosphorylating eIF2 on serine 51 of its subunit (17, 18). The ability of Tat to serve as a substrate for PKR
raised the possibility that Tat might compete with eIF2 for
phosphorylation by activated PKR. To address this hypothesis, activated
PKR was incubated with Tat and increasing concentrations of purified
eIF2 (Fig. 7, lanes 1-4). The results suggest that at high concentrations, eIF2 reduces Tat phosphorylation by PKR (lane 4). Conversely, increasing the concentration of
Tat protein, while maintaining a fixed concentration of eIF2, resulted in a marked reduction in eIF2 phosphorylation (lanes 5-8).
These data suggest that Tat and eIF2 can compete as substrates for
phosphorylation by autophosphorylated PKR. Considering the relative
molecular masses of Tat and eIF2 (about 10 and 125 kDa, respectively),
they appear to serve as substrates for PKR and as competitors on a comparable molar basis.
Experiments to this juncture have shown that Tat is not only a
substrate for PKR but that it is also able to inhibit eIF2 phosphorylation by this kinase. The activation of PKR is closely associated with its autophosphorylation (14). To determine whether Tat
can inhibit this reaction as well, Tat was added to kinase reactions
either before or after the introduction of dsRNA (Fig. 8). In the absence of Tat, PKR was autophosphorylated in a
dsRNA-dependent fashion (lanes 1 and
2). Preincubation of PKR with increasing concentrations of
purified Tat-72 prior to the addition of dsRNA eliminated PKR
autophosphorylation (lanes 3 and 4). As in
previous experiments (Figs. 2, 3, 4, 5), when the addition of Tat was delayed until after PKR activation by dsRNA had occurred, both Tat and PKR were
phosphorylated (lane 5). These observations indicate that
Tat-72 can inhibit the activation of PKR by dsRNA in vitro, as well as its activity in phosphorylating eIF2.
The molecular mechanisms governing the control of HIV-1 gene expression are complex and not fully understood. In the present study our aim has been to further characterize the properties of the HIV-1 regulatory protein, Tat, specifically with regard to its relationship with the interferon-induced dsRNA-activated protein kinase PKR. Several reports link these two regulatory proteins. First, both of them can bind to TAR RNA, the structured RNA segment that is the target for Tat trans-activation (56, 57). Second, TAR RNA modulates the activity of PKR, although there is disagreement as to whether this RNA activates the kinase (57-59) or blocks its activation (36, 60). Third, Tat stimulates the translation of TAR-containing RNAs in vitro (61). Fourth, a synthetic Tat peptide that binds TAR RNA was reported to inhibit PKR activation (62). Finally, Tat appears to down-regulate PKR in cells infected with HIV-1 or stably expressing Tat (23). We therefore set out to test the hypothesis that there is a direct interaction between Tat and PKR. Three lines of evidence support this hypothesis.
The results of in vitro kinase assays demonstrated that
dsRNA-activated PKR phosphorylates Tat purified according to a number of different protocols. Both the one-exon and two-exon forms of Tat
(Tat-72 and Tat-86, respectively) are substrates for the kinase. Mutations that prevent Tat from trans-activating HIV-1 transcription (P18IS and C22G) do not affect Tat phosphorylation, but a mutation removing residues 49-72 (48) eliminates phosphate labeling by activated PKR. Without exception, the phosphorylation of Tat proteins by PKR was shown to be dependent on the prior activation of PKR by
dsRNA. Tat-72 could also serve as an inhibitor of PKR, both by blocking
its ability to autophosphorylate in response to dsRNA (and hence to
become activated for phosphorylation of substrates) and by competing
with its natural substrate eIF2. In the competition assay, the affinity
for Tat and eIF2 seemed to be comparable in molar terms. Binding
studies showed that PKR can form a complex with either Tat-72 or Tat-86
and that the Tat sequences required for this interaction include
residues 49-72. Similarly, the binding of PKR to HIV-2 Tat was
dependent on the integrity of the basic region. Deletions in the N
terminus of both Tat proteins led to increased PKR binding, suggesting
that a conformational change may render the interaction site more
available to PKR. While this report was in preparation, McMillan and
co-workers (63) reported the interaction of Tat with PKR. In their
study, Tat-86 was phosphorylated but labeling of Tat-72 was not,
although its ability to inhibit PKR was observed. Although we cannot
provide an explanation for this discrepancy, we speculate that it might
be attributable to the different sources of Tat employed (McMillan
et al. (63) used synthetic Tat-72).
PKR phosphorylates two serine and one threonine residues immediately
adjacent to the basic region of Tat, consistent with the known
preference of PKR for serines in the context of a basic amino acid
environment (17, 18, 64). This suggested that Tat phosphorylation might
influence its ability to trans-activate via its interaction with TAR,
or via effects on its cellular localization, or both. When tested in a
Tat-dependent transcription assay (65), phosphorylation by
PKR elicited no discernible effect on
trans-activation2; however, this negative
result is difficult to interpret as only a fraction of the Tat was
modified. On the other hand, in this study we demonstrate the
phosphorylation of Tat by the mitogen-activated kinase PKC. PKC also
phosphorylates Nef (66), although PKR did not. Several researchers have
proposed that PKC plays a role in Tat-mediated trans-activation
(67-69). Depletion of PKC in Jurkat and 293 cells resulted in a
reduction in Tat trans-activation, and a PKC mutant lacking a
functional ATP-binding site failed to support trans-activation (69). It
is also possible that PKC influences HIV-1 transcription via the
phosphorylation of IB (70) or another unidentified PKC substrate
(71). Since both PKR and PKC phosphorylate Tat and I
B in
vitro, it will be important to learn whether they modify the same
sites on these proteins and whether they mediate similar responses
in vivo. In this connection, we note that a number of
laboratories have reported their inability to detect Tat
phosphorylation in vivo (67, 68). Our failure to observe
reaction of antibodies with PKR-phosphorylated Tat suggests that these
negative results may be due to the inability of anti-Tat antibodies to
recognize the modified form of Tat. Indeed, phosphorylation of the Tat
protein of HIV-2 has been observed both in vivo and in
vitro (39).
The ability of Tat to inhibit PKR also has potential biological
implications. Dominant negative forms of PKR can cause malignant transformation of 3T3 cells (24, 25). Therefore it is conceivable that
down-regulation of the kinase by Tat could contribute to the
deregulation of growth control seen in Tat-treated cells derived from
Kaposi's sarcoma lesions of HIV-1-infected individuals (72, 73).
Furthermore, PKR plays a recognized role in the interferon-induced antiviral response (14). It has been reported that HIV-1 replication is
sensitive to interferon and that Tat confers partial resistance to the
effect of interferon (3). Our data suggest that Tat may exert its
effect via its ability to inhibit the activation of the
interferon-induced protein kinase PKR, thereby reducing the
PKR-mediated phosphorylation of eIF2, or of IB, or other substrates
of this kinase. Moreover, the ability of Tat to form a complex with PKR
provides a possible mechanism for the repression of PKR levels that
have been observed in HeLa cells stably expressing Tat or in T-cells
infected with HIV-1 (23).
Although both PKR and Tat are both RNA-binding proteins and can
interact with TAR RNA, the PKR/Tat interaction documented here is
evidently independent of TAR RNA and of RNA in general. In that both
Tat and eIF2 are substrates for PKR and interact with similar
affinities, it would appear that Tat resembles the vaccinia virus PKR
inhibitor K3L (74). This early protein displays sequence homology to
the subunit of eIF2 in the region of its phosphorylation site;
lacking an appropriately placed serine residue, it acts as a
pseudo-substrate for PKR. Like Tat, K3L down-regulates PKR by directly
binding to the kinase and preventing its activation (i.e.
its autophosphorylation) and inhibiting its activity (74). This
mechanism is distinct from the competition that can occur between Tat
and PKR for binding to TAR RNA but not dsRNA (59, 61, 62).
In conclusion, we have demonstrated that both the virally activated
protein kinase PKR and the mitogen-activated kinase, PKC, are able to
phosphorylate the HIV-1 trans-activation protein Tat. Protein
phosphorylation constitutes an important mechanism for the regulation
of intracellular events (75). Although the consequences of Tat
phosphorylation with respect to the regulation of HIV-1 gene expression
remain to be established, the ability of Tat to control the activity of
PKR may have far-reaching implications. These potentially include
effects on both transcription and translation, mediated by the
phosphorylation of IB and eIF2, respectively, and suggest a
mechanism whereby Tat may contribute to the ability of HIV to evade the
action of PKR. The regulation of Tat by a kinase may provide the
trigger whereby latent integrated virus is activated and elicits
exciting possibilities for targeted intervention.
We thank Nora Poppito for technical help.