Active cAMP-dependent Protein Kinase Incorporated within Highly Purified HIV-1 Particles Is Required for Viral Infectivity and Interacts with Viral Capsid Protein*

Host cell components, including protein kinases such as ERK-2/mitogen-activated protein kinase, incorporated within human immunodeficiency virus type 1 (HIV-1) virions play a pivotal role in the ability of HIV to infect and replicate in permissive cells. The present work provides evidence that the catalytic subunit of cAMP-dependent protein kinase (C-PKA) is packaged within HIV-1 virions as demonstrated using purified subtilisin-digested viral particles. Virus-associated C-PKA was shown to be enzymatically active and able to phosphorylate synthetic substrate in vitro . Suppression of virion-associated C-PKA activity by specific synthetic inhibitor had no apparent effect on viral precursor maturation and virus assembly. However, virus-associated C-PKA activity was demonstrated to regulate HIV-1 infectivity as assessed by single round infection assays performed by using viruses produced from cells expressing an inactive form of C-PKA. In addition, virus-associ-ated C-PKA was found to co-precipitate with and to phosphorylate the CAp24 gag protein. Altogether our results indicate that virus-associated C-PKA regulates HIV-1 infectivity, possibly by catalyzing phosphorylation of the viral CAp24 gag protein. Protein phosphorylation is (cid:1) Ci of [ (cid:3) - 32 P]ATP in the appropriate buffer containing 13.5 m M Mg 2 (cid:1) , 90 (cid:1) M ATP, and 2 (cid:1) M cAMP for 30 min at room temperature. After several washes, phosphorylated products were eluted, loaded onto 15% SDS-PAGE, transferred onto PVDF membrane, and revealed by autoradiography. Phosphorylated CAp24 gag was identified by incubating the membrane with anti-CAp24 gag goat polyclonal serum (Biogenesis).

Protein phosphorylation is one of the primary processes by which external physiological stimuli influence intracellular events in eucaryotic cells. Protein kinase activity was also reported to contribute to a cascade of events controlling the definition of infectivity for a number of retroviruses and nonretroviruses. Many purified virions, including insect viruses, as granulosis virus (1), and plant viruses, including cauliflower mosaic virus (2), have been found to display protein kinase activity associated with their viral particles. For animal viruses, association of protein kinase activity with viral particles has been frequently reported, especially for enveloped virus particles (for review see Ref. 3). Among these, some virionassociated protein kinases have been identified as virally en-coded products. As an example, herpes simplex viruses encode their own ribonucleotide reductases (ICP6 for HSV-1 and ICP10 for HSV-2), which have been assigned a serine-threonine protein kinase activity (4,5), and similar observations have been performed for the human cytomegalovirus (6), pseudorabies virus (7), and varicella-zoster virus (8). For some small genome viruses, encapsidation of cellular kinases was found to replace their lack of kinase gene. Indeed, Rhabdoviridae (9), Hepadnaviridae (10), Retroviridae including Rauscher murine leukemia virus (11), and RNA tumor viruses (12) were found to incorporate host cell protein kinases within their membrane or inside the viral core. Even though their precise contribution has not yet been identified, there is ample evidence to suggest that virus-associated protein kinases are crucial for viral infectivity, and their possible function includes both the regulation of the viral nucleic acid replication and transcription and the modification of virus structural proteins that leads to either uncoating or encapsidation of viral nucleic acids.
For the human immunodeficiency virus type 1 (HIV-1), 1 the incorporation of signaling molecules including cellular protein kinases within viral particles was also questioned. HIV-1 virions were found to contain several phosphoproteins including MAp17 gag (13), CAp24 gag (14), and recently p6 gag proteins (15). The viral matrix MAp17 gag is phosphorylated by virus-associated kinase(s) at various steps of the HIV-1 replicative cycle (16). C-terminal phosphorylation of a tyrosine residue within MAp17 gag protein during and immediately after virus production was proposed to facilitate the dissociation of viral matrix protein from the membrane phospholipids of the nascent virus particle prior to and during virus assembly (17,18). Additional phosphorylation on serine residues was next reported and was proposed to promote membrane dissociation of the reverse transcription complex from the cell membrane at the site of entry, allowing its nuclear translocation (16,19). Such modifications were found to occur in preintegration complexes isolated from target cells and in native virions, suggesting that host cell serine-threonine kinases might possibly be incorporated within HIV particles (16).
CAp24 gag protein was also proposed to be phosphorylated by a virus-associated serine-threonine kinase. Such phosphorylations of CAp24 gag protein on serine residues were reported to be required for viral infectivity, because the reverse transcription process is unable to complete in CAp24 gag mutants with mu-tated phosphoacceptor sites (14). Despite the fact that the precise function of these modification remains unknown, by similarity to other viruses (i.e. poliovirus and herpes virus), it can be hypothesized that phosphorylation of HIV-1 capsid protein probably generates some repulsive forces among proteinprotein interactions that participate in viral core destabilization and viral particle uncoating.
During the past few years, we reported the presence of at least two cellular host cell serine-threonine kinases incorporated within HIV-1 particles. One of these proteins was identified as an active form of the extracellular signal-regulated kinase 2 (ERK-2)/mitogen-activated protein kinase (20). The functional integrity of virus-associated ERK-2 was found to be crucial for virus infectivity. As a confirmation, ERK-2 packaged within HIV-1 particles was found to be responsible for the phosphorylation of serine residues within MAp17 gag protein, indicating that this virus-associated kinase regulates viral infectivity by promoting membrane translocation of the reverse transcriptase complex in the host cell and thus participating in viral uncoating (19).
Despite evidence obtained for the incorporation of other protein kinases within HIV particles, including a 53-kDa serinethreonine kinase that remains unidentified so far (20), the precise nature of proteins responsible for CAp24 gag proteins phosphorylation remains to be defined. The present work was then designed to identify serine-threonine kinases incorporated within HIV particles. We demonstrate here by using subtilisin-digested viruses that the catalytic subunit of the cellular protein kinase A (C-PKA) is incorporated within HIV-1 virions. The virus-associated C-PKA was found to be enzymatically active. The implication of incorporated C-PKA in virus infectivity was analyzed. Virions produced from cells devoid of C-PKA kinase activity displayed reduced infectivity in single infection assays and in infection of lymphoblastoid T cell lines, despite normal assembly and fully mature morphology of virions. Analysis of viral protein phosphorylation revealed that virus-associated C-PKA interacts with and phosphorylates the CAp24 gag capsid protein. Altogether, our data support the hypothesis that the active catalytic subunit of PKA associated to HIV-1 particles regulates viral infectivity. A possible function could rely on its ability to interact with and to phosphorylate the CAp24 gag capsid protein.
Virus Purification-Supernatants from H9 cells and H9 cells infected with HIV-1 NL4.3 strain were cleared from cellular debris by low speed centrifugation, filtered on a 0.45-m-pore-size membrane (Millipore) and concentrated by centrifugation at 4,000 rpm for 40 min on Centricon Plus-80, Biomax-PB membrane with a cut-off of 100 kDa (Millipore). Concentrated supernatants were then digested in a solution of 1 mg/ml subtilisin (Sigma) in 10 mM Tris, pH 8.0, and 1 mM EDTA for 16 h at 37°C. Digestion was stopped with 5 g/ml phenylmethylsulfonyl fluoride (final concentration) for 30 min at room temperature. The digested supernatants were then centrifuged through 20% sucrose cushion at 25,000 rpm for 2.5 h at 4°C in a 28.38 rotor (Kontron Instruments). The pellets were solubilized in kinase assay lysis buffer (50 mM HEPES, pH 7.5, 0.5% Nonidet P-40, 1 mM phenylmethylsulfonyl fluoride, 100 g/ml leupeptin, 100 g/ml aprotinin, 200 M sodium orthovanadate, 20 mM ␤-glycerophosphate, and 50 mM sodium fluoride).
The concentration of proteins in viral lysates was normalized by quantification of CAp24 gag antigen concentration assessed with anti-CAp24 gag enzyme-linked immunosorbent assay (Beckman Coulter).
PKA in Vitro Kinase Assays-PKA activity in viral and cellular lysates was assayed using a PKA assay kit (Upstate Biotechnology, Inc.). Briefly, the lysates were incubated for 30 min at 30°C with 100 M Kemptide, a PKA-specific substrate, and 10 Ci of [␥-32 P]ATP in kinase buffer provided by the kit manufacturer, supplemented or not with PKA inhibitor peptide. The phosphorylated substrate is separated from the residual [␥-32 P]ATP using P81 phosphocellulose paper and quantitated by using a liquid scintillation counter.
Reverse Transcriptase Assay and CAp24 gag Antigen Enzyme-linked Immunosorbent Assay-Supernatants of transfected or infected cells were harvested and cleared of cells and cellular debris by centrifugation. For each sample, 1 ml of culture supernatant was tested for RT activity using a synthetic template primer that permitted the RT to neosynthesize radioactive DNA, as previously described (23). The CAp24 gag antigen concentration was determined by enzyme-linked immunosorbent assay.
Electron Microscopy and Immunoelectron Microscopy-Thin layer electron microscopy was processed as follows. The cells were washed once with medium and fixed with 0.05% glutaraldehyde, 4% paraformaldehyde in 0.1 M Sörensen buffer, pH 7.4 (24), for 1 h at 4°C. After washing, the cells were included in Lowicryl K4M at Ϫ20%, sectioned, and processed for conventional electron microscopy or immunoelectron microscopy. Mouse monoclonal anti-C-PKA antibody (Santa Cruz Biotechnology) and M35/2F8 rat anti-p6 gag mAb (obtained from M. G. Sarngadharan) (25) were used for immunoelectron microscopy, with the corresponding colloidal gold-labeled complementary antibodies, 10-nm gold-tagged anti-mouse Ig antibody, and 5-nm gold-tagged anti-rat, respectively.
For conventional electron microscopy, the cells were fixed with 2.5% glutaraldehyde in 0.1 M phosphate buffer, pH 7.5, postfixed with osmium tetroxide (2% in H 2 O), and treated with tannic acid (0.5% in H 2 O). After dehydration, the specimen were embedded in Epon (Epok-812, Fullam), and sections were stained with 2.6% alkaline lead citrate and 0.5% uranyl acetate in 50% ethanol. The specimens were observed under an Hitachi HU7100 electron microscope.
Viral Infectivity Assays-MAGI Cells (21) that stably express the ␤-galactosidase reporter gene cloned downstream of the HIV-1 long terminal repeat promoter were platted at 8 ϫ 10 4 cells/ml in 24-well plates. The cells were exposed to HIV stock solutions normalized according to RT activity. 48 h post-infection, virus infectivity was monitored by measurement of ␤-galactosidase activity from the cell lysates as previously described (26). Briefly, 200 l of total cellular extracts were incubated for 1 h at 37°C in 1.5 ml of buffer containing 80 mM Na 2 HPO 4 , 10 mM MgCl 2 , 1 mM 2-mercaptoethanol, and 6 mM o-nitrophenyl ␤-D-galactopyranoside. ␤-Galactosidase activity was evaluated by measuring absorbance at 410 nm and was normalized according to total protein content in the cell lysate. Viability of cells exposed to viruses produced in the presence of H89 was assessed as previously described (27). Briefly, the cells were incubated for 3 h at 37°C in the presence of 0.5 mg/ml (final concentration) MTT. The cells were then lysed by addition of 100 l of 0.1 N HCl in isopropanol, and absorbance was read at 570 nm with the solubilization buffer serving as a blank.
In Vitro Phosphorylation Assays and Immunoprecipitation of CAp24 gag -CAp24 gag protein was immunoprecipitated by incubating cell lysates or viral lysates in the presence of 2 g of anti-CAp24 gag mAbs (ICN) and 50 l of protein G-conjugated magnetic beads (Miltenyi Biotech). After 2 h at 4°C, immune complexes were separated on magnetic columns. For in vitro kinase experiments, immunoprecipitates were incubated in the presence of 2.5 g of recombinant C-PKA, 5 Ci of [␥-32 P]ATP in the appropriate buffer containing 13.5 mM Mg 2ϩ , 90 M ATP, and 2 M cAMP for 30 min at room temperature. After several washes, phosphorylated products were eluted, loaded onto 15% SDS-PAGE, transferred onto PVDF membrane, and revealed by autoradiography. Phosphorylated CAp24 gag was identified by incubating the membrane with anti-CAp24 gag goat polyclonal serum (Biogenesis).

cAMP-dependent Protein Kinase Catalytic Subunit Is Specifically Incorporated within Highly Purified HIV-1 Viral
Particles-We and others have previously reported that several serine-threonine and tyrosine kinases are packaged within HIV-1 particles. In an attempt to identify virus-associated protein kinase(s) incorporated within HIV-1 particles, lysates of highly purified viruses were probed with antibodies specific for various cellular protein kinases in Western blotting experiments. HIV-1 virions prepared from the supernatant of H9 cells infected by NL4.3 virus strain were digested with subtilisin protease and isolated by sucrose density ultracentrifugation as previously described (28). Viral lysates were then probed with a number of antibodies directed to cellular kinases. Antibodies specific for PKR, ERK-1, GSK3, PI3K, CK2, CSK, and c-Src protein kinases failed to detect the corresponding kinase in preparation of purified viruses (data not shown). In contrast, we found that p56 lck tyrosine kinase is specifically associated with subtilisin-digested HIV-1 particles (data not shown) as previously observed by high pressure liquid chromatography and Western blotting analysis (29). In addition, the presence of a protein with a molecular weight of 40,000 was revealed by the mean of a serum specific for the catalytic subunit of cAMP-dependent protein kinase (C-PKA) (Fig. 1A, lane 4), suggesting that C-PKA is incorporated within purified HIV-1 particles. As previously reported, the main difficulties for characterization of host cell proteins associated to viral particles lies in the co-sedimentation of cellular microvesicles with viral preparations. For this particular reason, the specificity of our observation was ascertained by analyzing the presence of C-PKA subunit in mock virus preparations (culture supernatants from H9 uninfected cells) that were prepared in the same manner as virion lysates (Fig. 1A, lane 3) without any detectable staining.
Our results indicate that the presence of contaminant C-PKA proteins evidenced from undigested preparations of mock virus supernatants (visualized from Fig. 1A, lane 1), because of the co-sedimentation of microvesicles with HIV-1 viral particles, is completely removed by digestion of purified virions with protease. To ascertain the relevance of our results, the immunoblot was then reprobed with anti-ERK-2 antibody (Fig. 1B). As previously reported, the presence of the mitogen-activated protein kinase ERK-2 was evidenced from purified virions (20) but not from subtilisin-digested preparations of H9 supernatants.
The efficiency of subtilisin treatment was also ascertained by analyzing gp41 env envelope glycoprotein pattern in subtilisindigested viral preparations. As shown from Fig. 1C, a reduction in the molecular masses of the proteins detected by anti-gp41 env serum was observed after subtilisin treatment, indicating that proteins outside the viral envelope have been efficiently digested. It is worthwhile to notice that the size of the C-PKA is not altered after digestion of viral preparations with subtilisin, indicating that in contrast to gp41 env , the size of which is lowered from 41 to 22 kDa after treatment, C-PKA is protected from enzymatic digestion. This result suggests that virus-associated C-PKA is located inside the virus particle. Finally, the amount of protein loaded in each lane was controlled by probing the membrane with anti-CAp24 gag mAbs (Fig. 1D). Similar experiments performed from supernatants of cells chronically infected with HIV-2 ROD or SIV strains did not allow the detection of the PKA catalytic subunit in the HIV-2 and SIV particles (data not shown). These results indicate that the catalytic subunit of PKA is selectively incorporated within HIV-1 particles.
The association of C-PKA to HIV-1 viral particles was then controlled by immunoelectron microscopy after double labeling of both chronically HIV-1-infected cells and free subtilisindigested viral particles. Chronically HIV-1-infected 8E5 cells, that derive from a CEM parental cell line or CEM-uninfected cells used as a control were embedded in Lowicryl, and the sections were double labeled with anti-p6 gag mAbs allowing identification of viral structures and anti-C-PKA serum. Primary antibodies were revealed with anti-rat Ig or anti-rabbit Ig secondary reagents coupled with 5-or 10-nm gold particles, respectively. As shown in Fig. 2A, the sections obtained from HIV-1-infected cells with anti-C-PKA mAbs revealed specific structures labeled by anti-p6 gag mAbs, and low background was detected in the section of uninfected CEM cells (Fig. 2B). These data confirmed the presence of PKA catalytic subunit within HIV-1 virions. Similar observations were performed from inclusions of cell-free virions previously digested with subtilisin (data not shown). It is important to notice that labeling of the sections with anti-p6 gag and anti-C-PKA allowed us to monitor the efficiency of subtilisin digestion by comparing the presence of microvesicles in untreated and subtilisintreated viral preparations.
cAMP-dependent Protein Kinase Incorporated within HIV-1 Viral Particles Is Catalytically Active-In the cell context, the PKA activity is regulated by the association of C-PKA catalytic subunits with two regulatory domains termed R-PKA that inhibit the catalytic properties of C-PKA. To further investigate whether the C-PKA incorporated within highly purified HIV-1 particles is enzymatically active, the presence of R-PKA regulatory subunit was investigated by Western blotting experiments performed from lysate of subtilisin-digested HIV-1 particles. In our experimental conditions, the presence of R-PKA subunits could not be proved in preparations of viruses (data not shown). In the absence of detection of virus-associated regulatory R-PKA subunits, we next investigated for the presence of virus-associated PKA enzymatic activity. Virus particles or subtilisin-treated virions were lysed and incubated in appropriate buffer in the presence of Kemptide, a PKA synthetic substrate. PKA kinase activity was measured by counting the incorporation of [␥-32 P]ATP in Kemptide. As shown in Fig. 3A, phosphorylation of Kemptide was detected from lysate of concentrated HIV-1 NL4-3 particles and was maintained in subtilisin-digested viruses. Basal phosphorylation level measured in the absence of Kemptide is indicated for each sample. Specificity of PKA activity was next assessed by analyzing phosphorylation of Kemptide substrate in the presence of highly specific inhibitors of either PKA activity or protein kinase C/calmodulin kinase activity as a control. Incorporation of [␥-32 P]ATP within the substrate was abolished when PKA inhibitor peptide was added to the samples but not when protein kinase C/calmodulin kinase inhibitor mixture was added to the reaction mixture (Fig. 3B). Altogether, these data indicate that C-PKA kinase activity is associated with highly purified HIV-1 virions.
C-PKA Activity Affects HIV-1 Particles Production but Is Not Required for Viral Maturation or Assembly-We next studied whether virion-associated C-PKA activity affects viral infectivity. Virus particles were produced from cells cultured in the presence of H89, a synthetic inhibitor specific for PKA. 40 h after transfection of Cos-7 cells with the pNL4-3 plasmid, culture medium was replaced for a 4-h time period, either by medium alone or medium containing 100 M of H89, a concentration of inhibitor that does not affect cell viability as determined by MTT testing (data not shown). H89 was then removed, and the cells were maintained in culture for an additional 4 h to allow the release of viral particles before supernatants were collected. First, we analyzed the ability of H89 concentrations added to the culture medium to efficiently inhibit cellular PKA activity. Lysates of H89-treated cells prepared at various times following the removal of the inhibitor of the culture medium were assayed for PKA activity by in vitro phosphorylation assays of Kemptide. As shown in Fig. 4A, addition to the culture medium of 100 M H89 for a 4-h period was sufficient to inhibit C-PKA cells for at least a 4-h period after the inhibitor was removed from the culture medium.
Then virus production from transfected cells cultured in the presence or absence of H89 was assessed by measuring RT activity in culture supernatants. In these experimental conditions, we found that the addition of H89 to the cells significantly lowers RT activity measured in cell culture supernatants (Fig. 4B). These results indicate that host cell PKA activity might regulate HIV-1 particle production.
The consequences of H89 treatment on viral product maturation were next investigated by Western blot experiments. Cell-associated viral proteins or protein content of samples prepared from supernatants of mock transfected cells or Cos-7 cells transfected with pNL4-3 and cultured in medium alone or medium supplemented with H89 were subjected to 10% SDS-PAGE and immunoblot analysis. The amounts of proteins loaded were normalized according to total protein content for cell associated proteins and according to RT activities for cell free viral products. The blots were sequentially probed with rabbit anti-RT and anti-gp41 env sera and anti-CAp24 gag mAbs (Fig. 5). According to densitometry scanning, no significant difference was observed for amounts of RT, Gag precursor (p55 gag ), and gp41 env products detected from lysates of H89treated or untreated cells. In addition, analysis of cell free viral products allowed us to detect similar amounts of RT heterodimer (p51 pol and p66 pol ), gp41 env , intermediate processing product (p41 gag ), and cleaved protein (CAp24 gag ) in supernatants collected from untreated cells or cells maintained in the presence of H89, indicating that inhibition of C-PKA activity does not affect the maturation of HIV-1 viral particles. The consequences of H89 treatment on HIV-1 particles assembly were also evaluated by electron microscopy. Transfected Cos-7 cells cultured in the presence of H89 were embedded in Epon, and the morphology of viruses budding at the cell surface was analyzed. No modification of virus morphology was detected from viruses produced from H89-treated cells as compared with those produced from untreated cells (data not shown). Altogether these data indicate that inhibition of host cell PKA enzymatic activity, despite lowering RT titers in culture supernatants, does not generate any detectable defect in viral proteins maturation nor in virus morphology as tested by electronic microscopy experiments.

FIG. 5. Inhibition of host cell C-PKA activity do not interfere with expression and maturation of viral proteins.
Cell-associated viral proteins or protein content of viral particles produced from Cos-7 cells transfected with pNL4-3 and maintained or not in the presence of H89 inhibitor were subjected to 10% SDS-PAGE and immunoblot analysis. Amounts of viral protein lysates were normalized for RT activity content. The blots were sequentially probed with rabbit anti-RT or anti-gp41 env sera and anti-CAp24 gag mAbs, and the protein bands were visualized using chemiluminescent detection. The positions of migration of RT heterodimer (p51 pol and p66 pol ), Gag precursor (p55 gag ), intermediate processing product (p41 gag ), cleaved capsid protein (CAp24 gag ), and gp41 env are indicated on the right.

FIG. 4. Inhibition of C-PKA activity in host cell lowers HIV particle production.
A, analysis of PKA activity in Cos-7 cells treated or not with a PKA specific inhibitor (H89). The cells were treated for 4 h with 100 M of H89, inhibitor was removed, and cellular PKA activity was measured as described under "Materials and Methods" from the cell lysates at different times after treatment removal (0, 2, and 4 h). The values correspond to one representative experiment of three. B, release of HIV-1 particles was quantified by measuring RT activity in culture supernatants from Cos-7 cells transfected with pNL4-3 and cultured in medium alone or in medium supplemented with 100 M H89. H89 inhibitor was added to the cells for a 4-h time period 40 h post-transfection. The inhibitor was then removed, and the cells were maintained in culture for an additional 4 h. The supernatants were then assayed for RT activity (each value corresponds to the mean of six separate experiments).
virus-associated protein kinase has been previously shown to modulate virus infectivity (19). Implication of virion-associated C-PKA in HIV-1 infectivity was then assessed in a single cycle transactivation assay. Viruses collected from transfected Cos-7 cells maintained in the presence or in absence of H89 inhibitor were normalized for RT activity, and various virus concentrations were used to infect the MAGI indicator cell line. These cells expressing both HIV-1 receptor CD4 and co-receptors (CXCR4 and CCR5) and a ␤-galactosidase reporter gene driven by an HIV-1 viral long terminal repeat allowed us to quantify the efficiency of virus infection. After 2 days in culture, ␤-galactosidase activity was determined from the cell lysates. As shown in Fig. 6A, infection was barely detected from MAGI Cells exposed to viruses produced from H89-treated cells, whereas high X-galactosidase hydrolysis levels were observed from cells exposed to comparable levels of viruses produced in absence of PKA inhibitor. Differences in infectivity were observed at any given viral input tested (5000, 10,000, or 20,000 cpm of RT), with a marked difference when high infectious doses were used. To ascertain that differences in infectivity observed were not related to a cytotoxic effect of residual H89 present in the viral innoculum, cell viability was determined by MTT testing 48 h after exposure of the cells to the virus. As shown from Fig. 6, no significant cytotoxic effect was noted at any viral input used. This observation suggests that HIV-1 virions produced from cells with impaired C-PKA activity display a lower infectivity as compared with wild type viruses.
Infectivity of PKA-deficient virions was next analyzed by infection of H9 CD4 ϩ T lymphoblastoid cells. The cells were exposed to various amounts of virions produced by transfection of Cos-7 maintained in the presence or absence of H89 inhibitor and normalized according to RT activity. After 1 h at 37°C the cells were washed three times, resuspended in RPMI medium, and maintained in culture. The amount of virus produced by the cells was determined by measuring RT levels every 3 days from the cell supernatants. We observed that virus replication was significantly delayed when H9 cells were exposed to viruses produced from H89-treated cells, as compared with when these cells were exposed to wild type virus. Indeed, a delay of at least 3 days was observed for replication of viruses produced from C-PKA-deficient cells at any concentration tested (Fig.  6B, upper panel). As a representative example, values measured at day 10 after virus exposure for various concentrations of wild type and C-PKA-deficient viruses are shown in Fig. 6B  (lower panel). The absence of cytotoxic effect of residual H89 contained in the viral innoculum was controlled at different times after virus exposure. As mentioned in Fig. 6B (insert), the differences observed between replication of wild type and C-PKA-deficient virions in H9 cells at day 10 after infection are not related to toxicity of the viral input as cell viability is unaffected. Despite this, after 14 days in culture, significant RT activities were detected from H9 cells exposed to H89-treated viruses. This result may reflect either the appearance or the propagation of replication efficient viruses. Such a phenomenon might be the consequence of an unstable inhibitory effect of H89 viral particles maintained in the culture medium. Consequently, the discrepancy observed between wild type and H89treated viruses during follow-up of replication in H9 cells was significantly reduced at day 14 after viral exposure. Altogether our data indicate that viruses produced from cells expressing kinase-dead C-PKA, despite normal maturation and morphology, display reduced infectivity for CD4 ϩ cells as compared  H89). The lower panel represents the RT activities measured at day 10 after infection from the supernatant of H9 cells exposed to various concentrations of wild type or H89-treated viruses. Cell viability determined at day 10 post viral exposure by MTT cytotoxicity assay is shown as a control (inset). Each value is the mean of duplicate experiments. with viruses produced from cells expressing fully active C-PKA.
The CAp24 gag Protein Is a Substrate of C-PKA-Having demonstrated that active C-PKA associated with purified HIV-1 particles regulates viral infectivity, we attempted to explain its functional role. We have previously shown that several structural proteins of HIV-1 are phosphorylated by virus-associated kinases. Among these, the CAp24 gag protein was found to be phosphorylated by a host cell serine-threonine kinase incorporated within viral particles (14). The contribution of ERK-2 virus-associated kinase has been previously questioned, and its contribution in CAp24 gag phosphorylation was rejected (14). We thus investigated here the involvement of virus-associated C-PKA in viral protein phosphorylation with a particular focus to define the putative contribution of this kinase in CAp24 gag phosphorylation. In a first assay, we determined the capacity of CAp24 gag protein to interact in vivo with C-PKA. To this end, CAp24 gag protein was immunoprecipitated from extracts of H9 cells expressing the pNL4-3 plasmid or from lysates of viruses produced from these cells. Proteins co-precipitated with the HIV-1 capsid protein were then revealed with anti-C-PKA serum. In these experimental conditions C-PKA subunit was found to interact with CAp24 gag protein both in extracts prepared from infected cells and in virus lysates, because these proteins were found to co-precipitate after incubation of the membrane with anti-C-PKA serum (Fig. 7A).
The consequences of such interaction on CAp24 gag phosphorylation was then investigated. As a first attempt to analyze phosphorylation of CAp24 gag by C-PKA, recombinant CAp24 gag protein was incubated with active recombinant C-PKA and [␥-32 P]ATP in an in vitro kinase assay. As shown in Fig. 7B, C-PKA was found to phosphorylate recombinant CAp24 gag pro-tein in vitro. The specificity of the reaction was assessed by the addition of C-PKA inhibitor to the reaction mixture that was found to abolish CAp24 gag phosphorylation. Finally, the physiological relevance of this observation was estimated in an in vitro kinase experiment performed in the presence of nonrecombinant CAp24 gag protein. CAp24 gag was immunoprecipitated from H9 cells infected with the HIV-1 NL4-3 strain and was used in an in vitro kinase experiment performed in the presence of recombinant C-PKA. We found that mature CAp24 gag protein is phosphorylated in vitro by recombinant C-PKA (Fig.  7C). Altogether, these data support the hypothesis that CAp24 gag is likely to be a substrate for C-PKA in vivo.

DISCUSSION
In the present study we demonstrated the selective incorporation of active catalytic C-PKA subunit within HIV-1 virions. Experimental approaches used to this end were based on protease digestion and sucrose density separation of virions. This method is efficient for removal of 95% of the microvesicleassociated proteins from virion preparations as assessed by inclusion of viral preparations and electron microscopy of thin layer sections. This strategy was previously used to report association to the inner face of HIV-1 virions of several host cell proteins including cytoskeletal components (28), glutaredoxin (30), and several actin-binding proteins (EF1-␣ elongation factor, Pin1 prolyl isomerase, Nm23-H1, Lck, and HS-1 protein kinases) (29). Our present result corroborates previous observations indicating that protein kinase activity is associated with HIV-1 particles as for a number of enveloped DNA-and RNA-containing animal viruses. The presence of two cellular serine-threonine protein kinases incorporated in HIV-1 parti-  .). The reaction products were separated on 12% SDS-PAGE and transferred onto PVDF membrane, and phosphorylation of CAp24 gag protein was revealed by autoradiography ( 32 P). The membrane was successively probed with anti-CAp24 gag mAbs and an anti-C-PKA serum to ascertain the protein content in each lane. The positions of phosphorylated CAp24 gag (P-CAp24 gag ), CAp24 gag , and C-PKA are indicated by arrows. C, CAp24 gag protein was immunoprecipitated from lysates of HIV-1-infected H9 cells by using an anti-CAp24 gag or an irrelevant mAbs (IR.). Immunoprecipitates were subjected to in vitro phosphorylation assay in the presence of recombinant C-PKA, separated on 15% SDS-PAGE, transferred onto PVDF membrane, and revealed by autoradiography ( 32 P). Phosphorylated products corresponding to the CAp24 gag protein (P-CAp24 gag ) were identified by probing the membrane with anti-CAp24 gag mAbs and are indicated by arrows. cles has previously been reported, and virus-associated ERK-2 was identified as being one of them (20). We describe here that C-PKA is an additional serine-threonine kinase associated to HIV-1 particles.
A very large spectrum of host cell components have been reported to be incorporated within HIV particles. For most of them, questions have been unanswered about their precise contribution to the viral life cycle, although they are known for their functional role in cells. Host cell proteins embedded in the viral envelope were mostly proposed to enhance virus affinity for its target cell. Functions for cytosolic virus-associated compounds are less clearly defined, and their packaging within viral particles helps to create an appropriate environment for propagating the virus into the target cell. The association with HIV-1 particles of C-PKA devoid of regulatory subunit and associated to enzymatic activity suggests a functional contribution for this protein during the retroviral life cycle. Indeed, virus-associated ERK-2 enzymatic activity was previously reported to play a predominant role in the establishment of a functional reverse transcription complex by phosphorylating MAp17 gag protein, promoting dissociation of the reverse transcription complex and influencing its cellular localization in infected cells (19). We found here that impairment of host cell C-PKA activity by a synthetic inhibitor at the time of virus release results in the production of virions with reduced infectivity as assessed in single round infection assays performed in MAGI cells and in in vitro infection of the CD4 ϩ lymphoblastoid H9 cell line. We cannot rule out the possibility that H89 inhibitor affects a late stage of virus replicative cycle, but the absence of modification of viral precursor maturation or virus assembly as shown by conventional biochemical and electron microscopy methods suggests that defects in viral infectivity might rely on the inhibition of virus-associated C-PKA activity. Virus-associated C-PKA might thus contribute to the regulation of HIV-1 infectivity.
It is also important to notice that C-PKA activity in the host cell was found to influence viral production. Indeed the addition of H89 inhibitor to cells expressing the pNL4-3 plasmid was found to lower RT activities measured from the cell supernatants, suggesting a possible inhibitory effect on viral particle release. The role of cellular PKA is complex (for review see Ref. 31), and such alteration of viral production may be explained at different levels. First, PKA activity is required for regulation of host cell DNA-binding proteins involved in HIV transcription, including cAMP-responsive element-binding protein and NF-B factors, which are affected by cAMP-dependent phosphorylation. Second, PKA activity was also found to participate in chromatin remodeling through histone phosphorylation. Therefore, PKA activity might interfere with both integration and transcription of HIV genome. Third, PKA regulates exocytosis, endocytosis, and transcytosis by acting on transport events from the endoplasmic reticulum to the plasma membrane across the Golgi stacks (32).
Because inhibition of PKA by the addition of H89 inhibitor to the cells was known to alter vesicle-mediated transport along the exocytic route, the experimental procedures used in the present study and aimed at producing virus having incorporated an inactive PKA were designed to limit the consequences of PKA inhibition on envelope glycoprotein release and incorporation into virions. To this end, H89 treatment of the cells was followed by the removal of H89 and a 4-h culture time period before recovery of the virus. Western blot analysis of cell-associated viral proteins and virions protein content from cells exposed or not to H89 demonstrated that similar levels of gp41 env transmembrane proteins were incorporated in wild type and H89-treated virions, and viruses produced from cells cultured in the presence of efficient concentrations of PKA inhibitors were found to display fully mature protein content and morphologic characteristics similar to wild type virions. As a consequence we propose that the reduction of viral particle release observed from H89-treated cells is rather due to a direct consequence of inhibiting either transcription or integration steps of HIV-1.
Characterization of viral protein(s) targeted by virus-associated C-PKA will help to elucidate its precise contribution in HIV life cycle. C-PKA recognizes a consensus sequence (RX(S/ T)X) that can be found in a number of HIV proteins including MAp17 gag , CAp24 gag , gp41 env , Rev, Vpr, and Vif. In the present paper we report that the CAp24 gag protein is phosphorylated in vitro by recombinant C-PKA, and the physiological relevance of this event was confirmed by co-immunoprecipitation of CAp24 gag and C-PKA in experiments performed from both cell extracts and virus lysates. The putative role of C-PKA in CAp24 gag phosphorylation is in agreement with our previous observations indicating that viral capsid protein is phosphorylated by virus-associated serine-threonine kinase distinct of ERK-2 (14).
Capsid phosphorylation is a common feature reported for a number of viruses and retroviruses. The herpes simplex virus type 1 tegument protein VP22 is phosphorylated on serine residues by both virion-associated and cellular kinases. Such phosphorylations were proposed as a regulatory mechanism in the dissociation of structural components of this virus (33,34). Similar observations have been performed on the destabilizing effect that phosphorylations of the viral capsid have in uncoating of poliovirus (35). For HIV-1, CAp24 gag phosphorylations were found to participate in the early events of HIV replication, mainly in the late retrotranscription steps (14). Although the consequences of PKA-dependent phosphorylation of CAp24 gag remains to be defined, a possible implication of capsid protein phosphorylation in uncoating events can be proposed. Indeed, although evidence has been obtained from previous and present studies that CAp24 gag may be phosphorylated by a virusassociated kinase in in vitro phosphorylation assays and that C-PKA may participate in such phosphorylation events, it remained difficult to prove the presence of phosphorylated forms of capsid protein in cell-free assembled HIV-1 mature particles either from one-or two-dimensional gel analysis (data not shown). This observation suggests that CAp24 gag phosphorylation by C-PKA may be a transient event, occurring at a precise stage of HIV life cycle. If CAp24 gag phosphorylation appeared to participate in the HIV-1 uncoating process, then the addition of negatively charged phosphate groups to capsid monomers could generate some repulsion force and destabilize the viral core. The contribution of PKA in such events is currently under investigation in our group.
In conclusion, host cell protein kinases associated to HIV-1 particles, including C-PKA, despite minor structural components of the viral architecture, might play an important role in the virus life cycle by modifying virion structural proteins, interfering with viral assembly, or uncoating and regulating the release of nucleic acid into the host cell. It is also important to consider that the presence of virus-associated C-PKA might be of special interest in deregulating the host cell activation level by triggering abnormal signaling. Interestingly, HIV infection was shown to up-regulate PKA activity and several cAMP-inducing mediators. In T cells from untreated HIV-infected patients, elevated levels of cAMP have been detected, and PKA subunits are constitutively activated (36). Identifying the precise relationship between the virus-associated protein kinases and the viral life cycle will reveal novel targets for the development of specific and new antiviral agents.