was obtained from Sigma. Unless indicated otherwise,2 U of. PC-PLC per ml was used to stimulate cell cultures. TNF and. PMA were used at concentrations of ...
Vol. 67, No. 11
JOURNAL OF VIROLOGY, Nov. 1993, p. 6596-6604 0022-538X/93/1 16596-09$02.00/0 Copyright © 1993, American Society for Microbiology
Phosphatidylcholine Hydrolysis Activates NF-KB and Increases Human Immunodeficiency Virus Replication in Human Monocytes and T Lymphocytes FERNANDO ARENZANA-SEISDEDOS,I* BELEN FERNANDEZ,' ISABEL DOMINGUEZ,2 JEAN MARC JACQUE,' DOMINIQUE THOMAS,' MARIA TERESA DIAZ-MECO,2 JORGE MOSCAT,2 AND JEAN LOUIS VIRELIZIER' Unite dImmunologie Virale, Institut Pasteur, 28 rue du Dr. Roux, 75724 Paris Cedex 15, France,' and Centro de Biologia Molecular, Consejo Superior de Investigaciones Cientificas-UAM, Canto Blanco, 24049 Madrid, Spain2 Received 11 May 1993/Accepted 3 August 1993
We have tested whether breakdown of phosphatidylcholine (PC) initiated by exogenous addition of a PC-specific phospholipase C (PC-PLC) from Bacillus cereus or by endogenous overexpression of PC-PLC induces functional activation of NF-KB and increases human immunodeficiency virus (HIV) enhancer activity. PC-PLC-activated hydrolysis of PC was found to induce bona fide p5O/p65 NF-KB binding activity in three different cell lines of human or murine origin. No significant changes in the turnover of other cellular phospholipids were detected in PC-PLC-treated cells. Induction of NF-KB by PC-PLC did not depend on de novo synthesis of proteins or autocrine secretion of either tumor necrosis factor or interleukin 1. In human monocytic and lymphoblastoid T-cell lines, induction of NF-KB by PC-PLC resulted in clear induction of luciferase expression vectors placed under the control of synthetic KB enhancers or wild type, but not KB-mutated, HIV long terminal repeat constructs. HIV replication was increased by PC-PLC in chronically infected monocytes and T lymphocytes. NF-KB activation promoted by addition of exogenous PC-PLC correlated with an intense production of diacylglycerol. However, addition of a phosphatidylinositol-specific PLC from B. cereus also induced diacylglycerol but did not activate KB enhancer-directed vectors. PC-PLCinduced NF-KB activation could not be blocked by a specific inhibitor of phorbol ester-inducible protein kinases C. These results indicate that a cellular transduction pathway, dependent on specific PC breakdown, is functional in T lymphocytes and monocytes and may be used by various transmembrane receptors to activate HIV transcription through NF-KB-dependent induction of the HIV enhancer. as the inducer, since some PKC isoforms, such as the calciumdependent ax, 3, and y PKCs, are intracellular targets for phorbol esters. Furthermore, direct demonstration that purified PKC indeed phosphorylates IKB has been obtained with cell lysates (13). The intracellular messenger which links membrane signals to PKC is diacylglycerol (DAG). An important but transient source of DAG is the hydrolysis of phosphatidylinositol (PI) which takes place following activation signals through specific membrane receptors in many cell types, especially T lymphocytes (5). However, hydrolysis of phosphatidylcholine (PC) is now largely accepted to account for the production of the sustained amounts of DAG detected in activated cells, in contrast to DAG derived from PI hydrolysis (11, 23). The latter declines within minutes after the onset of cell activation. The sustained production of DAG generated after specific breakdown of PC is associated with the induction of DNA synthesis in fibroblasts (21). Moreover, induction of the interleukin 2Rcx (IL-2RoL) gene and DNA synthesis in human T lymphocytes require the prolonged presence of high levels of DAG (4), suggesting that a sustained production of this messenger is necessary to maintain transcription of inducible genes involved in cell activation and proliferation. Stimulation of T lymphocytes either with phorbol esters or through the CD3-T-cell receptor activates NF-KB and HIV LTR transcription and also induces breakdown of PC, with generation of DAG independently of PI hydrolysis (34). Furthermore, activation of p21 ras, which plays a critical role in cell activation and induces PC breakdown and DAG formation without significant modification of PI metabolism (20, 28)
Initiation of human immunodeficiency virus (HIV) genome transcription, after integration of the HIV provirus into host cell DNA, appears to be determined by transcription factors interacting with the long terminal repeat (LTR) (14). Transcriptional activation of the LTR depends largely on a major enhancer made of two directly repeated sequences able to respond to the transcription factor NF-KB (25, 33). For this event to occur and HIV transcription to be initiated, NF-KB, usually associating p5O (18) and p65 (30) subunits, must be activated and translocated into the nucleus from the cytoplasm, where it is normally retained by interaction with inhibitory proteins named IKB (3, 22). The biochemical events leading to this critical dissociation of NF-KB heterodimers from IKB have not been elucidated. Induction of transcriptionally active NF-KB complexes occurs after cellular activation either following physiological signalling engaging the CD3-T-cell receptor complex in T lymphocytes, in particular antigen recognition (15, 16) and anti-CD3 antibodies (35), or following stimulation with artificial inducers such as phorbol esters in both T lymphocytes and monocyte/macrophages, the two main targets of HIV infection (37). Inhibition of type C protein kinases (PKC) in T lymphocytes blocks induction of the HIV enhancer by either CD3 antibodies or phorbol esters (35). Thus, it is assumed that in all these cases, NF-KB induction depends on the activation of PKC. This is obviously the case when phorbol esters are used
*
Corresponding author. 6596
VOL. 67, 1993
PHOSPHATIDYLCHOLINE HYDROLYSIS INDUCES HIV REPLICATION
transactivates the HIV enhancer in human cells (la). Finally, recent data from our laboratory now directly document in Xenopus laevis oocytes a cascade of transduction events involving ras-induced hydrolysis of PC and DAG production mediated by a PC-specific phospholipase C (PC-PLC) (10). This was shown to lead to both nuclear translocation of NF-KB and enhanced transcription of KB-dependent vectors through activation of the zeta isoform of PKC (PKC ~), a calciumindependent and non-phorbol 12-myristate 13-acetate (PMA)inducible isotype of PKC (26). Altogether, the above-mentioned data indicate that, in human monocytes and T lymphocytes, the two types of target cells of HIV infection, phosphodiesterase-mediated hydrolysis of PC has the potential to be a major activation pathway leading to induction of nuclear NF-KB-binding activity and transactivation of the HIV enhancer. Using cell systems relevant to HIV pathogenesis, including an IL-2-dependent T-cell clone, we demonstrate that specific hydrolysis of PC, but not PI, induces nuclear translocation of bona fide NF-KB and enhances HIV enhancer activity independently of classical, phorbol ester-inducible PKC. Furthermore, we show that increased HIV enhancer activity correlates with augmented HIV replication in chronically infected cells treated with PC-PLC. MATERIALS AND METHODS Cells. J. Jhan is a human lymphoblastoid cell line derived from Jurkat. U937 is a human monocytic cell line. Ul, derived from U937, and ACH2, derived from the lymphoblastoid T-cell line CEM, are chronically HIV-infected cell lines. NIH 3T3 cells are murine fibroblasts. The SPBenv cell line was obtained by infection of SPB21 cells, an IL-2-dependent CD4+ T-cell clone (15), with a recombinant HIV virus carrying an amphotropic envelope. The recombinant HIV virus was obtained by cotransfecting in COS7 cells a noninfectious HIV provirus, a derivative of HIV R7/neo (12) (a gift from M. Feinberg) carrying a 1,300-bp deletion from nucleotide 5930 to nucleotide 7210 in the envelope-coding region, and a simian virus 40 promoter-directed expression vector (a gift from M. Alizon) of the amphotropic 4070A Moloney leukemia virus envelope (7). Selection of infected cells was obtained by using a neomycin selectable marker located at the 3' end (nef region) of the HIV genome. Reagents. PC-PLC from Bacillus cereus was either obtained and purified essentially as previously described (21) or purchased from Calbiochem. PI-specific phospholipase C (PlPLC) from B. cereus was purchased from Boehringer Mannheim. Recombinant human tumor necrosis factor (TNF) was a gift from W. Fiers. The bisindolylmaleimide GF 109203X (GF) (36) is an inhibitor of PKC ot, 3, and y isoforms and was a kind gift from J. Kirilovsky (Glaxo, Les Ulis, France). PMA was obtained from Sigma. Unless indicated otherwise, 2 U of PC-PLC per ml was used to stimulate cell cultures. TNF and PMA were used at concentrations of 340 IU/ml and 20 ng/ml, respectively. Plasmids. Vector HIV-LTR-Luc (2) consisted of the XhoI (-640)-HindIII (+86) fragment containing the U3 and R regions of the HIV type 1 strain LAI cloned in a luciferase expression vector (pC-Luc) (15). HIV-LTRAKB-Luc (2) was obtained by insertion into the pC-Luc plasmid of the KB consensus motif-deleted XhoI-HindIII fragment of the LAI LTR. The kb-CONA-Luc vector derives from a CONA-CAT vector described by Kimura et al. (19). It carries a luciferase gene (pCLuc) placed under the control of three synthetic copies of the KB consensus of the immunoglobulin K-chain
6597
promoter cloned in the BamHI site located upstream of the conalbumin (CONA) transcription start site. The CONA-LUC vector is identical to plasmid kb-CONA-Luc but does not contain the KB sequences. The PC-PLC expression vector, pCMV-PLC, was constructed from the B. cereus PC-PLC gene. It carries the 21-amino-acid signal peptide of the Escherichia coli outer membrane protein OmpA. This construct, including 19 bp upstream of the start codon, was then inserted into the mammalian expression vector pRcCMV (pCMV), purchased from Invitrogene. Analysis of products of phospholipid metabolism. For determination of the levels of different phospholipids, cells were labelled with 10 ,uCi of [methyl-'4C]choline for PC and sphingomyelin and with 10 ,uCi of [2-'4C]ethan-1-olamine for phosphatidylethanol. Identical amounts of L-[U-'4C]serine and [myo-2-3H]inositol were used for labelling of phosphatidylserine or polyphosphoinositides, respectively. After induction with agonists, cell cultures were treated with methanol and cell extracts were fractionated into chloroform and aqueous phases. Organic phases were processed essentially as described elsewhere (21). For determination of DAG, we used a DAG detection kit (Amersham) according to the instructions of the manufacturer. Results are expressed in terms of the 32p label incorporated into phosphatidic acid, which results from the in vitro action of a DAG kinase on the pool of cellular DAG. Thus, the amounts of radiolabelled phosphatidic acid can be directly correlated to the levels of DAG produced by cells. EMSA. Crude nuclear cell extracts were prepared as described elsewhere (17). For electrophoretic mobility shift assay (EMSA), 3 ,ug of protein extract was incubated with [a-32P] dCTP-labelled double-stranded oligodeoxynucleotides corresponding either to the KB motif found in the promoter of the immunoglobulin K-chain gene or to the tandem of KB sequences of the HIV enhancer. The binding reaction was analyzed by electrophoresis in nondenaturing 5% polyacrylamide gels. In DNA-binding competition assays, unlabelled oligonucleotides were used in a 40-fold molar excess with respect to the corresponding radiolabelled probes. When indicated, specific polyclonal rabbit antibodies (a gift from R. T. Hay) generated against recombinant p50 or p65) or a C-terminal peptide of c-Rel (a gift from I. Verma) were incubated with the nuclear extracts for 10 min before addition of probes. Similarly, 20 ng of recombinant IKB/MAD-3 (a gift from R. T. Hay) was used to block the binding of NF-KB to the radiolabelled oligonucleotides. Transfection assays. Plasmids were transfected in U937 or J. Jhan cells by the DEAE-dextran method. Briefly, 5 x 106 cells were incubated with 5 ,g of DNA in a buffer containing 500 ,ug of DEAE-dextran per ml for 20 min at room temperature. Subconfluent NIH 3T3 cells were transfected by the calcium phosphate precipitation technique for 4 h. After transfection, cells were washed and allowed to remain in culture for 48 h in the corresponding culture medium enriched either with 5% fetal calf serum in the case of U937 and J. Jhan cells or with 0.2% fetal calf serum for NIH 3T3 fibroblasts. Transfected cells were induced with the different stimuli for the last 6 h of the culture period. For detection of luciferase activity, total cell extracts were prepared in a lysis cell buffer, and samples were analyzed, as previously described (15), in a luminometer (Berthold, Gosheim, Germany). Data are expressed in terms of relative luciferase activity units, calculated as (light emission from experimental sample - light emission of lysis buffer alone)/micrograms of cellular protein in the sample. Detection of p24 HIV antigen. Ul or ACH2 cells were seeded at 5 x 105/ml and stimulated for 24 h. At this time,
6598
ARENZANA-SEISDEDOS ET AL.
-
_
+ *
-
+
+_ +_ +
I';
VP _C __
+
WM MM WTWWT
.
COMPET.
-
_ C
_ ^ .,
~ ~M J.JIIAN
13937
INDUC. CHX
J. VIROL.
9.
-.N
-+ ++-
+-+-+ M WTWT
FIG. 1. PC-PLC from B. cereus induces nuclear KB-binding activity on human monocytes (U937) and T lymphocytes (J. Jhan) independently of de novo synthesis of proteins. Cells were preincubated for 45 min with 50 p.g of cycloheximide (CHX) per ml, and then they were stimulated for 90 additional min with PC-PLC or TNF (J. Jhan). A double-stranded oligonucleotide (5'-ACACAGGGAC'TTCCGCTGGGGACTTTC CAGGGA-3') containing the tandem of KB wild-type (WT) sequences of the HIV enhancer was used as a radiolabelled probe. For DNA-binding competition (COMPET.), unlabelled oligonucleotides (WT or the mutated [M] counterpart [5'-ACACACTCACTT[CCGCTGCTCACTCC AGGGA-3']) were used in a 40-fold molar excess. *, a sample containing radiolabelled probe incubated in the absence of nuclear extracts.
culture supernatants were harvested, and p24 viral core antigen was quantitated by enzyme-linked immunosorbent assay (DuPont/NEN). SPBenv-cells were starved of IL-2 for 18 h before induction, and viral antigen was evaluated in the supernatants of 24-h-stimulated cells. RESULTS PC-PLC induces bona fide nuclear KB-binding activity in human monocytes, T lymphocytes, and murine fibroblasts. Analysis by EMSA of crude nuclear extracts from monocytes, T lymphocytes, or murine fibroblasts incubated with PC-PLC from B. cereus revealed the presence of specific KB-binding activity in the nuclei of treated cells (Fig. 1 and 2). Specific binding of NF-KB proteins from either U937 or J. Jhan cells to the radiolabelled probes (HIV enhancer or K-chain immunoglobulin KB motifs) was abrogated by addition of a 40-fold molar excess of the corresponding unlabelled, wild-type oligonucleotide but not its mutated counterpart (Fig. 1). In some experiments (Fig. 1, J. Jhan cells, and Fig. 2, NIH 3T3 cells), nuclear extracts from TNF-treated cells were introduced as a source of control bona fide NF-KB. Results displayed in Fig. 1 show that inhibition of protein synthesis did not block the induction of NF-KB complexes by PC-PLC. On the contrary, at least in T lymphocytes, stimulation of cycloheximide-treated cells with either PC-PLC or TNF resulted in increased induction of NF-KB (Fig. 1, J. Jhan cells). In order to characterize the components of the KB-retarded complexes induced by PC-PLC, we preincubated nuclear extracts from U937 cells with either recombinant IKB/MAD3 or rabbit specific antisera directed against p50, p65, or c-Rel. Either anti-pSO or anti-p65 antibodies detect single bands in Western immunoblot analysis of cytoplasmic extracts from U937 cells, corresponding in apparent molecular mass to proteins of 50 and 65 kDa, respectively (1). The presence of
the p50 subunit in the complexes induced by PC-PLC was demonstrated (Fig. 2, U937 cells), as an anti-pSO antibody, but not a preimmune rabbit serum, greatly impaired binding of NF-KB proteins and induced the up-shifting of the KB-binding activity. It should be noted that addition of p50 antiserum equally blocked the formation of both the more slowly and the faster-migrating KB-binding complexes (Fig. 2, U937 cells). In contrast, Fig. 2 shows that incubation of PC-PLC-induced U937 nuclear extracts with anti-p65 antiserum, but not with an anti-c-Rel antiserum, dramatically retarded the migration of the more slowly migrating complex but did not affect the mobility of the faster-migrating complex. Together, these data demonstrate that two KB-binding activities mainly composed of homodimeric p50 (lower arrowhead) or p5O/p65 (upper arrowhead) complexes were induced in U937 cells by PC-PLC. This conclusion is reinforced by the ability of recombinant IKB to prevent binding of heterodimeric pSO/p65 complexes in the three cell lines (Fig. 2) without affecting the formation of DNA-binding complexes formed by p50 homodimers (lower arrowhead) detected in both U937 and NIH 3T3 cells (Fig. 2). PC hydrolysis induces transcriptional activation of the HIV LTR and HIV replication in chronically infected cell lines. To demonstrate that nuclear NF-KB complexes induced by PCPLC in U937 or J. Jhan cells were functionally active, we analyzed the transcription activity of transfected luciferase expression vectors placed under the control of the HIV LTR. After 42 h of transfection, cells were stimulated for 6 additional h with PC-PLC, and then luciferase activity was evaluated in whole-cell lysates. TNF was used as a positive control of HIV LTR transactivation. Results of transient transfection depicted in Fig. 3 show a clear induction of HIV LTR activity and demonstrate that the enhancer sequences mediated the transactivation effect promoted by PC-PLC inasmuch as deletion of the tandem of KB motifs abolished the LTR responses
VC)L. 67, 19'93
6599
PHOSPHATIDYLCHOLINE HYDROLYSIS INDUCES HIV REPLICATION NIH3T3
U937
_s
INDUC.
ANTIB.
- -
J,JHAN
x,
K#-q
INDUC.
,
,,
__SeCOMPEl..WTWTWT ++
PKC-INH..++
+
1kB
__w 6' .
j
FIG. 2. EMSA characterization of proteins integrated in the KB-binding complexes induced by exogenous PC-PLC and independence of their induction from PKC ot, and y. U937 and J. Jhan cells were stimulated (INDUC.) for I h with 2 U of PC-PLC per ml. One unit of PC-PLC per ml was used to induce NIH 3T3 cells. Nuclear extracts from U937 cells were preincubated for 10 min with specific anti-p50, anti-p65, or anti-c-Rel rabbit antibodies (ANTIB.) before addition of the radiolabelled probes. A rabbit preimmune (pre.) serum was used as a control. U937 cells were preincubated for 45 min with 2.5 ,uM PKC inhibitor GF before stimulation with PC-PLC. To prevent the formation of NF-KB DNA, 20 ng of recombinant IKB/MAD3 (IKB) was added to the nuclear extracts from U937, J. Jhan, or NIH 3T3 cells. For both U937 and J. Jhan nuclear extracts, the KB wild-type sequence of the HIV enhancer was used as a probe. For NIH 3T3 extracts, a synthetic oligonucleotide with the KB motif from the immunoglobulin K-chain promoter was used as a radiolabelled probe. Specific inhibition (COMPET.) of DNA binding by unlabelled oligonucleotides was performed (NIH 3T3 and J. Jhan cells) by adding a 40-fold molar excess of the corresponding wild-type (WT) oligonucleotide. The positions of heterodimeric p5O/p65 complexes and homodimeric p50 complexes are indicated (upper and lower arrowheads, respectively).
to either PC-PLC or TNF. Maximal induction was usually obtained at PC-PLC concentrations of 1 to 2 U/ml, although concentrations as low as 0.1 U/ml induced detectable transactivation (data not shown). Similar results were obtained with NIH 3T3 fibroblasts by using the KB-CONA-Luc vector. It should be noted that the HIV LTR transactivation induced by PC-PLC correlated with its PC-hydrolyzing activity and DAG production (Table 1 and Fig. 4a). In order to obtain further evidence for the specificity of this effect, we compared the effects of PI-PLC and PC-PLC, both obtained from B. cereus, on U937 cells transfected with the KB-CONA-Luc vector. Results shown in Fig. 4b indicate that PC-PLC, but not PI-PLC, was able to increase the activity of the KB-directed vector. The inability of PI-PLC to enhance the activity of KB enhancers contrasted with its capacity to induce a clear increase of DAG in the same cells (Fig. 4a). The above results, obtained with exogenously added PCPLC, were confirmed in experiments whose results are shown in Fig. 5 by using an expression vector for the PC-PLC from B. cereus. Transient expression of this enzyme in NIH 3T3 cells resulted in a clear and enhancer-dependent transactivation of the KB-CONA-Luc vector in transient cotransfection assays. Interestingly, in these experiments the level of transactivation of KB-CONA-Luc obtained by cotransfection of the PC-PLC expression vector was not further modified by exogenously added PC-PLC (data not shown). To determine whether HIV enhancer transactivation induced by PC-PLC resulted in augmented HIV genome transcription and virus production, we induced hydrolysis of PC in three different cell lines (Ul, ACH2, and SPBenv-) harboring integrated and replicative HIV provirus. Results shown in Fig. 6 demonstrate that PC-PLC treatment of Ul, a monocytic cell line, or ACH2, a lymphoblastoid cell line, consistently led to enhanced levels of p24 Gag antigen in cell supernatants, which correlated well with the levels of HIV LTR transactivation promoted by PC-PLC in control monocytic or lymphocytic cells (Fig. 3). It should be noted that the enhanced production
*
13.7X
300-
HIV-LTR-Luc
HIV-LTRAkB-Luc 200-
100_
1X
~~~~~0.8X
11
I..
m
ix J.Jhan
0)
3.5X
40
30' ,
20
3.3X -
0 DE
HIV-LTR-Luc
HIV-AkB-Luc
1
1IO U)i
0
1X
1.2X
l.lX
o
0
H
kB-CONA-Luc
D
CONA-Luc
ix
Inducers
none
PC-PLC
TNF
none
ix 0!5~
PC-PLC
ix
TNF
FIG. 3. Addition of exogenous PC-PLC induces transcription activity of transfected HIV LTR and KB enhancer-directed luciferase expression vectors in U937, J. Jhan, and NIH 3T3 cells. Fold amplification is indicated at the top of each bar. The data are representative of at least three independent experiments for each cell line. RLU, relative luciferase activity units.
6600
J. VIROL.
ARENZANA-SEISDEDOS ET AL.
TABLE 1. Phospholipid turnover in cells treated with PC-PLC Cells
Phospholipid level' (% of control) PE PS SM
PC-PLC PC
U937
+
J. Jhan
+
100 74 ± 3
100 94 ± 3
100 98 ± 2
100 100 ± 3
PIPs
100 100 ± 4
6x104
( >1000)
X 5x104 -j Lu
0
0
N 4x1
4 -
a
100 75 ± 2
100 98 ± 2
100 100 ± 3
100 100 ± 5
100 100 ± 4
"Cells were labelled with different precursors as described in Materials and Methods. Cells were treated with 1 U of PC-PLC per ml for 45 min. Control values (in disintegrations per minute per well) for U937 cells were as follows: PC 510,000; phosphatidylethanolamine (PE); 415,000; phosphatidylserine (PS), 76,000; sphingomyelin (SM), 34,000; polyphosphoinositides (PIPs), 130,000. Control values for J. Jhan cells were as follows: PC, 400,000; PE, 840,000; PS, 63,000; SM, 23,000; PIPs, 112,000.
4C2
3x10 4
(625)
F I. X,) CL o 0.
4
2x10
INC 0
(78)
(66)
O of extracellular p24 antigen in PC-PLC-treated cells correlated with an increased detection of intracellular p24 (data not shown). This rules out the possibility that the effect of PC-PLC in HIV-infected cells was due to an enhanced shedding of virus particles and not to increased expression of the viral genome. In an attempt to analyze the role of PC hydrolysis in the reactivation of HIV replication in normal, nontransformed human T lymphocytes, we used an IL-2-dependent T-cell line harboring an integrated HIV provirus replicating at very low levels. Stimulation of these cells with PC-PLC or with PMA for 24 h in the absence of IL-2, significantly enhanced production of p24 antigen (Fig. 6). Addition of exogenous PC-PLC specifically hydrolyzes PC but not other, cellular phospholipids. In a previous report (21), we documented that addition of purified PC-PLC from B. cereus to NIH 3T3 cells activates breakdown of PC in the absence of significant changes in the turnover of other phospholipids. Results given in Table 1 show that a similar phenomenon occurred in J. Jhan and U937 cells treated for 30 min with 1 U of PC-PLC per ml. Indeed, under these experimental conditions a clear effect on PC metabolism was observed without modifications of the sphingomyelin or PI turnover. Very modest modifications of the hydrolysis of phosphatidylserine or phosphatidylethanol were detected. Concomitant with the hydrolysis of PC, a dramatic and prompt increase in DAG levels was observed in U937 cells treated with PC-PLC
(Fig. 4a).
Classical, PMA-inducible PKC a, 1, and y are not involved in the transduction of PC-PLC-induced activation of NF-KB and HIV genome transcription. To investigate whether NF-KB nuclear translocation induced by PC-PLC requires activation of PMA-inducible PKC isoforms, we stimulated U937 cells in the presence of GF, a potent and highly specific inhibitor of the cx, 1, and -y PKC isoforms. Figure 2 shows that 2.5 ,iM GF did not impair induction of nuclear NF-KB complexes by PC-PLC in U937 cells. In agreement with the lack of inhibitory effect of GF on the nuclear translocation of NF-KB, transactivation of the HIV LTR and induction of HIV replication were unaffected by this compound (Fig. 7). The failure of GF to block activation by PC-PLC was not due to lack of effect of this inhibitor on its targets, since GF abolished induction of both LTR transactivation and viral replication induced by PMA
(Fig. 7). DISCUSSION
Activation of the PC turnover by exogenous PC-PLC was shown to induce a clear expression of nuclear KB-binding
Inducers
none
Incubation Time
none
60'
30' 15.7X
100 -i
5.5X
0 ._
as
10
a.._
0L
3 0
co 01
A
,1. Inducers
none
TNF
PC-PLC
PI-PLC
FIG. 4. Production of DAG by PC- or PI-PLC-treated U937 cells and comparative analyses of the abilities of the two phosphodiesterases to induce the transactivation of a KB-directed vector. (a) U937 cells were induced with 1 U of either PC-PLC or PI-PLC per ml for 30 or 60 min. The production of DAG in picomoles is indicated at the top of each bar. (b) After 42 h of transfection of U937 cells with the KB-CONA-Luc expression vector, cells were stimulated for 6 additional h with PC-PLC or PI-PLC at a concentration of 1 U/ml. RLU, relative luciferase activity units.
activity in human monocytes and T lymphocytes and also murine NIH 3T3 fibroblasts. These results are in agreement with a recent report (32) suggesting that TNF-induced PCPLC activity activates nuclear NF-KB binding activity and our previous observation that expression of activated p2lras, an inducer of PC breakdown (20, 28), activates the HIV enhancer in human cells (la). The present results were obtained with two different, highly purified preparations of PC-PLC, one commercially available and the other prepared by us. Furthermore, expression of a transfected B. cereus PC-PLC gene, the same which was used to prepare purified proteins, was shown to induce transcriptional effects on KB-dependent genes similar to that observed in cells treated with exogenous PC-PLC proteins. It should be stressed, however, that nuclear translocation of NF-KB does not always result in functional transactivation of the HIV enhancer. We have indeed shown a dissociation between these two events in human T-cell clones (15). We here
VOL. 67, 1993
PHOSPHATIDYLCHOLINE HYDROLYSIS INDUCES HIV REPLICATION
300 ,
CONA-Luc
10
6601
5 -
Ul 42.2X
kB-CONA-Luc
-J
10
4 -
-i
1.0
10
200
-
3 -
ix
._
40
10
(0
13X
CU 8> (U
0 ._
0 L-
100
-
6.6X 0
Co-transfected plasmid
pCMV
pCMV-PLC
FIG. 5. Endogenous PC-PLC expressed from a cytomegalovirus immediate-early gene promoter-directed vector induces transactivation of a KB enhancer. One microgram of the luciferase reporter plasmid (KB-CONA-Luc or CONA-Luc) and 2.5 ,ug of either the vector carrying the PC-PLC gene (pCMV-PLC) or a control without the inserted gene (pCMV) were cotransfected in NIH 3T3 cells. For detection of luciferase activity, cell lysates were analyzed 48 h after transfection. Fold amplification is indicated at the top of each bar. The results are representative of three independent cotransfection experiments. RLU, relative luciferase activity units.
show that PC breakdown induced by exogenous PC-PLC results in enhanced transcription of luciferase vectors under the control of either the whole (U3 plus R) HIV LTR or synthetic oligonucleotides with KB-responsive motifs cloned upstream from the weak CONA promoter. The induction of HIV LTR activity by PC-PLC correlated with detection of bona fide NF-KB complexes in EMSA using the HIV enhancer as a probe. Moreover, deletion of the two NF-KB-responsive motifs in the HIV LTR abolished responsiveness to exogenous PC-PLC. To investigate this phenomenon under conditions as relevant as possible to the pathogenesis of AIDS, we generated an appropriate cell system consisting of a CD4-positive, IL-2dependent human T-cell clone stably transfected with an infectious HIV-1 provirus clone. The constitutive transcriptional activity of this integrated provirus was shown to be upregulated by addition in the culture medium of PC-PLC. Thus, PC-PLC-inducible transduction pathways can activate HIV transcription in normal, nontransformed T cells under conditions likely to mimic HIV reactivation from quiescence in circulating T cells. PC-PLC clearly induced functional nuclear bona fide NF-KB complexes in monocytes, T lymphocytes, and fibroblasts. Concomitantly with the induction of p5O/p65 heterodimers, a significant increase of homodimeric p50 complexes was also detected in U937 and NIH 3T3 treated cells (Fig. 2). Indeed, EMSA of the NF-KB complexes detected in the nuclei of
none
PMA
PC-PLC
FIG. 6. Exogenous PC-PLC induces HIV replication in chronically infected monocytes and T lymphocytes. SPBenv- cells were seeded at 2 x 105/ml for 24 h after being cultured for 18 h in an IL-2-starved culture medium. To evaluate the production of p24, culture supernatants were harvested after 24 h of induction. At least two independent experiments were performed with each cell line. Fold amplification of p24 production with respect to uninduced cultures is indicated at the top of each bar.
PC-PLC-treated U937 cells indicated that they were made mainly of p5O/p65 heterodimers and p50 homodimers, as their binding to KB oligonucleotides was upshifted by either antibodies to p50 or those to p65 but not by an anti-c-Rel antibody. In contrast to anti-p50 antibodies, which promote a clear inhibition and mobility retardation of both homodimeric and heterodimeric complexes in U937 nuclear extracts (Fig. 2), addition of anti-p65 exclusively retarded the migration of heterodimeric complexes, leaving unchanged the binding of p50 homodimers (faster-migrating complexes). Furthermore, our conclusion that the more slowly migrating complexes induced by PC-PLC are composed mainly of p5O/p65 heterodimers is reinforced by the ability of recombinant IKB/ MAD3 to prevent their binding to the radiolabelled probes. The specificity of the inhibitory effect of IKB is underlined by the fact that its addition to nuclear extracts did not substantially modify the binding of p50 homodimers (faster-migrating complexes in Fig. 2, U937 and NIH 3T3 cells).
J. VIROL.
ARENZANA-SEISDEDOS ET AL.
6602 400 :
U937
350
cc
C
300
0
CL cm
250
>,
200
o
150
a) C
100
a) 0
50,
II
01.n5_
1
Z-.
E CD
0..i
103
0. IV
1o2 f Inducers
none
ca/iy PKC inhibitor
__
TNF
PMA _
PC-PLC -
none
TNF
PMA
+
+
PC-PLC +
FIG. 7. Inducibility of the HIV LTR and HIV replication by PC-PLC are not mediated by classical, PMA-inducible PKC ax, or y. U937 cells were transfected for 42 h with the HIV-LTR-Luc vector and then were preincubated for 45 min with 2.5 puM PKC inhibitor GF and subsequently stimulated for 6 additional h with PC-PLC, TNF, or PMA. Ul cells were induced for 24 h with the same inducers. GF was added 45 min before induction. Production of p24 was evaluated in culture supernatants harvested 24 h after induction. RLU, relative luciferase activity units.
Addition of exogenous PC-PLC from the same source as that used in this study was previously shown to promote DNA synthesis and cell proliferation in fibroblasts (21). We now show that exogenous PC-PLC enhances the activity of KBdependent luciferase expression vectors and induces the replication of integrated HIV provirus. To rule out that PC-PLC would act through transcription of cellular genes, in particular that of cellular factors known to induce NF-KB such as TNF and IL-1, we induced cells in the presence of cycloheximide. Our results indicate that blockade of de novo protein synthesis does not prevent activation of NF-KB complexes in PC-PLCtreated U937 cells. These results, together with the absence of either TNF or IL-1 activity in the supernatants of PC-PLCtreated cells (data not shown), rule out the possibility that autocrine secretion of cytokines plays a role in the induction of NF-KB complexes induced by PC-PLC. PC-derived DAG is known to be a pivotal, long-lasting second messenger in cell activation and mitogenesis (11, 23). Prolonged treatment of U937 cells with PC-PLC resulted in high-level production of DAG, still observed 1 h after induction. This is not surprising, since we have previously reported that 24 h after incubation of fibroblasts with PI-PLC or PC-PLC, DAG production remains elevated (9). We compared the efficiencies with which PC-PLC and PI-PLC induce NF-KB activation and DAG production. Both enzymes in-
duced DAG, but only PC-PLC activated NF-KB in the U937 monocytic cell line. DAG production induced by PC-PLC was usually higher than that induced by PI-PLC. This is unlikely to account for the difference in activities, since we observed similar differences in other cell systems (8), despite induction of functional effects of PI-PLC such as p80 phosphorylation or PKC and epidermal growth factor receptor downregulation (9). We have demonstrated previously (21) and confirm here that PC-PLC induces hydrolysis of PC but very few, if any, modifications of other cellular phospholipids. We also confirmed that exogenous PI-PLC induces hydrolysis of PI (data not shown). If DAG production indeed has a role in the transduction pathways elicited by exogenous PC-PLC and leading to NF-KB activation, two alternative explanations are possible. One is that molecular forms of DAG induced by the two enzymes are functionally different. The other is that PC-PLC, but not PI-PLC, induces in addition to DAG other, yet uncharacterized signals that function independently or in association with DAG to ultimately activate NF-KB. It was recently reported that addition of synthetic forms of DAG induces sphingomyelinase activity, which in turn would be responsible for NF-KB activation (32). It should be noted, however, that we could not detect sphingomyelin breakdown after PC-PLC treatment of the cell lines used here. We previously demonstrated that incubation of fibroblasts with PC-PLC induces PKC translocation (8, 9). NF-KB activation induced by PC-PLC could not, in the present study, be blocked by GF, a highly specific inhibitor of PMA-inducible PKC isoforms, including a, I, and -y PKCs. This is in agreement with previous work showing that PMA-induced PKC downregulation does not block the mitogenic effect of exogenous PC-PLC on fibroblasts (21). These results, together with the lack of effect of PI-PLC, a known inducer of PMAinducible PKCs (8), suggest that the ox, P, and y isoforms of PKC do not mediate the NF-KB-activating effects of PC-PLC. Although the cascade of transduction events induced by PC-PLC and resulting in NF-KB activation is not yet characterized for mammalian cells, an exciting possibility is that a new isoform of PKC, (, has a pivotal role in this cascade, integrating and transducing the signals induced by PC-PLC. Indeed, recent data from our laboratory (10), obtained with specific inhibitory peptides (pseudosubstrates), show that C, but not a-, PKC mediates the functional activation of NF-KB induced by microinjection of either p2lras or PC-PLC proteins into Xenopus oocytes. Recent results (9a) in our laboratory are compatible with this hypothesis, since cells expressing constitutively high levels of 4 PKC show permanent NF-KB nuclear binding activity. An original cellular transduction pathway generated by cell membrane signalling and involving p2lras activation, PC breakdown, and C PKC induction may thus be envisaged. This pathway might be used by all membrane signals which can induce PC breakdown. This is the case with CD3 stimulation (34), TNF (31), IL-1 (29), and even PMA stimulation, which, through cross-talk, can induce PC hydrolysis (6, 23). In contrast, sphingomyelinase activation and ceramide production cannot account for NF-KB activation by all these inducers, since stimulation of the EL4 T-cell line through CD3 or by PMA, which induces intense NF-KBdependent HIV LTR transactivation (12a), does not elicit a ceramide response (24). Thus, PC breakdown triggered on the occasion of T-cell and macrophage activation by physiological inducers may have a major role in the induction of HIV replication. In monocytes, HIV replication itself induces, in an ot-, 3-, and y-PKC-independent manner sustained NF-KB activation responsible for self-perpetuating HIV enhancer activity (2, 27). The characterization of a transduction pathway
VOL. 67, ]1993
PHOSPHATIDYLCHOLINE HYDROLYSIS INDUCES HIV REPLICATION
controlling HIV genome transcription, induced by physiological cell signalling and possibly by HIV replication itself, offers potential cellular targets for antiviral intervention.
15.
ACKNOWLEDGMENTS We are indebted to S. Michelson and A. Garcia for helpful discussions and critical reading of the manuscript. This work was supported in part by funds of Agence Nationale pour la Recherche sur le SIDA (France) and by grants SAL-90-0070 and PTR 91-0022 from Comision Interministerial de Ciencia y Tecnologia and PB90.0074 from Direccion General de Investigacion Cientifica y Tecnologica (Spain). B.F. is supported by a fellowship from the European Communities. I.D. and M.T.D-M. are fellows of Gobierno Vasco and Ministerio de Educacion (Spain), respectively. REFERENCES 1. Arenzana-Seisdedos, F. Unpublished data. la.Arenzana-Seisdedos, F., N. Israel, F. Bachelerie, U. Hazan, J. Alcami, F. Dautry, and J. L. Virelizier. 1989. c-Ha-ras transfection induces human immunodeficiency virus (HIV) transcription through the HIV-enhancer in human fibroblasts and astrocytes. Oncogene 2:1359-1362. 2. Bachelerie, F., J. Alcami, F. Arenzana-Seisdedos, and J. L. Virelizier. 1991. HIV enhancer activity perpetuated by NF-KB induction on infection of monocytes. Nature (London) 350:709-712. 3. Baeuerle, P., and D. Baltimore. 1988. IKB: a specific inhibitor of the NF-KB transcription factor. Science 242:540-546. 4. Berry, N., K. Ase, A. Kishimoto, and Y. Nishizuka. 1990. Activation of resting human T cells requires prolonged stimulation of protein kinase C. Proc. Natl. Acad. Sci. USA 87:2294-2298. 5. Berry, N., and Y. Nishizuka. 1990. Protein kinase C and T cell activation. Eur. J. Biochem. 189:205-214. 6. Besterman, J. M., V. Duronio, and P. Cuatrecasas. 1986. Rapid formation of diacylglycerol from phosphatidylcholine: a pathway for generation of a second messenger. Proc. Natl. Acad. Sci. USA 83:6785-6789. 7. Chattopadhyay, S. K., A. I. Oliff, D. L. Linemeyer, M. R. Lander, and D. R. Lowry. 1981. Genomes of murine leukemia viruses isolated from wild mice. J. Virol. 39:777-791. 8. Diaz-Laviada, I., P. Larrodera, M. T. Diaz-Meco, M. E. Cornet, P. E. Guddal, T. Johanssen, and J. Moscat. 1990. Evidence for a role of phosphatidylcholine-hydrolysing phospholipase C in the regulation of protein kinase C by ras and src oncogenes. EMBO J. 9:3907-3912. 9. Diaz-Laviada, I., P. Larrodera, J. L. Nieto, M. E. Cornet, M. T. Diaz-Meco, M. J. Sanchez, P. H. Guddal, T. Johanssen, A. Haro, and J. Moscat. 1991. Mechanism of inhibition of adenylate cyclase by phospholipase C-catalyzed hydrolysis of phosphatidylcholine. J. Biol. Chem. 266:1170-1176. 9a.Diaz-Meco, M. T., E. Berra, M. M. Municio, L. Sanz, J. Lozano, I. Dominguez, V. Diaz-Golpe, M. T. Lain de Lera, J. Alcami, C. V. Paya, F. Arenzana-Seisdedos, J. L. Virelizier, and J. Moscat. 1993. A dominant negative protein kinase C 4 subspecies blocks NF-KB activation. Mol. Cell. Biol. 13:4770-4775. 10. Dominguez, I., L. Sanz, F. Arenzana-Seisdedos, M. T. Diaz-Meco, J. L. Virelizier, and J. Moscat. 1993. Inhibition of protein kinase C 4 subspecies blocks the activation of an NF-KB-like activity in Xenopus laevis oocytes. Mol. Cell. Biol. 13:1290-1295. 11. Exton, J. H. 1990. Signalling through phosphatidylcholine breakdown. J. Biol. Chem. 265:1-4. 12. Feinberg, M. B., D. Baltimore, and A. D. Frankel. 1991. The role of tat in the human immunodeficiency virus life cycle indicates a primary effect on transcriptional elongation. Proc. Natl. Acad. Sci. USA 88:4045-4049. 12a.Fernandez, B. Unpublished data. 13. Ghosh, S., and D. Baltimore. 1990. Activation in vitro of NFkappaB by phosphorylation of its inhibitor IkappaB. Nature (London) 334:678-682. 14. Haseltine, W. 1988. Regulation of replication of HIV-1, p. 139158. In B. R. Franza, Jr., B. R. Cullen, and F. Wong-Staal (ed.),
16.
17.
18.
19.
20. 21.
22.
23. 24.
25. 26.
27.
28.
29.
30.
31.
6603
Control of human retrovirus gene expression. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. Hazan, U., D. Thomas, J. Alcami, F. Bachelerie, N. Israel, H. Yssel, J. L. Virelizier, and F. Arenzana-Seisdedos. 1990. Stimulation of a human T cell clone with anti-CD3 or tumor necrosis factor induces NF-KB translocation but not human immunodeficiency virus 1 enhancer-dependent transcription. Proc. Natl. Acad. Sci. USA 87:7861-7865. Horvat, R. T., and C. Wood. 1989. HIV promoter activity in primary antigen specific human T lymphocytes. J. Immunol. 132:2745-2751. Israe, N., U. Hazan, J. Alcami, A. Munier, F. Arenzana-Seisdedos, F. Bachelerie, A. Israel, and J. L. Virelizier. 1987. Tumor necrosis factor stimulates transcription of HIV-1 in human T lymphocytes, independently and synergistically with mitogens. J. Immunol. 143:3956-3960. Kieran, M., V. Blank, F. Logeat, J. Vandekerekhove, F. Lottspeich, 0. Le Bail, M. B. Urban, P. Kourilsky, P. A. Baeuerle, and A. Israel. 1990. The DNA binding subunit of NF-KB is identical to factor KBF1 and homologous to the rel oncogene product. Cell 62:1007-1018. Kimura, A., A. Israel, 0. Le Bail, and P. Kourilsky. 1986. Detailed analysis of the mouse H-2KB promoter: enhancer-like sequences and their role in the regulation of class gene expression. Cell 44:261-272. Lacal, J. C., J. Moscat, and S. A. Aaronson. 1987. Novel source of 1,2-diacylglycerol elevated in cells transformed by Ha-ras-oncogene. Nature (London) 330:269-272. Larrodera, P., M. E. Cornet, M. T. Diaz-Meco, M. Lopez-Barahona, I. Diaz-Laviada, P. E. Guddal, T. Johansen, and J. Moscat. 1990. Phospholipase C-mediated hydrolysis of phosphatidylcholine is an important step in PDGF-stimulated DNA synthesis. Cell 61:1113-1120. Liou, H. C., G. P. Nolan, S. Ghosh, T. Fujita, and D. Baltimore. 1992. The NF-KB p50 precursor, p105, contains an internal IKB-like inhibitor that preferentially inhibits p50. EMBO J. 11: 3003-3009. Liscovitch, M. 1992. Crosstalk among multiple signal-activated phospholipases. Trends Biochem. Sci. 17:393-399. Mathias, S., A. Younes, C. C. Kan, I. Orlow, C. Joseph, and R. N. Kolesnick. 1993. Activation of the sphingomyelin signaling pathway in intact EL4 cells and in a cell-free system by IL-lb. Science 259:519-522. Nabel, G., and D. Baltimore. 1987. An inducible transcription factor activates expression of human immunodeficiency virus in T cells. Nature (London) 326:711-713. Ono, J., T. Fuji, K. Ogita, K. Kikkawa, K. Igarashi, and Y. Nishizuka. 1989. Protein kinase C C subspecies from rat brain: its structure, expression and properties. Proc. Nati. Acad. Sci. USA 86:3099-3103. Paya, C. V., R. M. Ten, C. Bessia, J. Alcami, R. T. Hay, and J. L. Virelizier. 1992. NF-KB-dependent induction of the NF-KB p50 subunit underlies self-perpetuation of human immunodeficiency virus transcription in monocyte cells. Proc. Natl. Acad. Sci. USA 89:7826-7830. Price, B. D., J. D. Morris, C. J. Marshall, and A. Hall. 1989. Stimulation of phosphatidylcholine hydrolysis, diacylglycerol release and arachidonic acid production by oncogenic ras is a consequence of protein kinase C activation. J. Biol. Chem. 264: 16638-16643. Rossof, P. M., N. Savage, and C. A. Dinarello. 1988. Interleukin 1 stimulates diacylglycerol production in T lymphocytes by a novel mechanism. Cell 54:73-81. Ruben, S., P. J. Dillon, R. Schreck, T. Henkel, C. H. Chen, M. Maher, P. A. Baeuerle, and C. Rosen. 1991. Isolation of a rel-related human cDNA that potentially encodes the 65-Kd subunit of NF-KB. Science 251:1490-1493. Schutze, S., D. Berkovic, 0. Tomsing, C. Unger, and M. Kronke. 1992. Tumor necrosis factor induces rapid production of 1'2'
diacylglycerol by a phosphatidylcholine-specific phospholipase C. J. Exp. Med. 174:975-988. 32. Schutze, S., K. Posthoff, T. Machleidt, D. Berkovic, K. Wiegmann, and M. Kronke. 1992. TNF activates NF-KB by phosphatidylcho-
6604
ARENZANA-SEISDEDOS ET AL.
line phospholipase C-induced "acidic" sphingomyeline. Cell 71: 765-766. 33. Siekevitz, M., S. F. Josephs, M. Dukovich, N. Pieffer, F. WongStaal, and W. C. Greene. 1987. Activation of the HIV-LTR by T cell mitogens and the trans-activator protein of HTLV-1. Science 238:1575-1578. 34. Stewart, S. J., G. Z. Cunningham, J. A. Strupp, F. S. House, L. L. Kelley, G. S. Henderson, J. H. Exton, and S. B. Bocckino. 1991. Activation of phospholipase D: a signaling system set in motion by perturbation of the lymphocyte antigen receptor/CD3 complex. Cell. Regul. 2:841-850.
J. VIROL.
35. Tong-Starksen, S., P. A. Luciw, and B. M. Peterlin. 1989. Human immunodeficiency virus long terminal repeat responds to T cell activation signals. Proc. Natl. Acad. Sci. USA 87:78617865. 36. Toullec, D., H. Pianetti, H. Coste, P. Bellevergue, T. Grand-Perret, M. Akajane, V. Baudet, P. Boisson, E. Boursier, F. Loriolle, L. Duhamel, D. Charron, and J. Kirilowsky. 1991. The bisindolylmaleimide GF109203x is a potent and selective inhibitor of protein kinase C. J. Biol. Chem. 266:15771-15781. 37. Virelizier, J. L. 1990. Cellular activation and human immunodeficiency virus infection. Curr. Opin. Immunol. 2:409-413.
