Enhancement of protein kinase C-dependent O2 production in Epstein-Barr virus-transformed B lymphocytes by p120Ras-GAP antisense oligonucleotide.

The mammalian Ras GTPase-activating protein (p120Ras-GAP) interacts with activated members of the Ras superfamily of GTP-binding proteins to accelerate their deactivation by sharply increasing their rates of GTP hydrolysis. Among the Ras-family proteins interacting with p120Ras-GAP is Rap1A/Krev1, whose activity is not affected by p120Ras-GAP but which competes with Ras for p120Ras-GAP. A second protein that interacts with p120Ras-GAP is P190Rac-GAP, which activates the GTPase of guanine nucleotide-binding proteins of the Rho family (including Rac1 and Rac2). Both these p120Ras-GAP-binding proteins are of interest in connection with the regulation of the respiratory burst oxidase, Rap1A/Krev1 because it copurifies with cytochrome b558 and p190Ras-GAP because it inhibits the Rac2-dependent activation of the respiratory burst oxidase in a cell-free system. Using an 18-mer antisense oligonucleotide, we were able to decrease the expression of p120Ras-GAP in Epstein-Barr virus-transformed B lymphocytes. Under conditions where p120Ras-GAP expression was significantly depressed by antisense oligonucleotides, we observed a 40% increase in protein kinase C-dependent but not receptor-dependent O2 production. In contrast, sense and scrambled oligonucleotides had no effect on either p120Ras-GAP expression or O2 production. Our results suggest a role for p120Ras-GAP as a negative regulator in the protein kinase C-mediated activation of the respiratory burst oxidase.

The products of the mammalian N-, K-, and Ha-ras genes are 21-kDa guanine nucleotide-binding proteins that possess a low intrinsic GTPase activity (1,2). They play a significant role in both the control of normal cell growth (3) and its malignant subversion (4) and have been implicated as key elements in signal transduction in lymphoid and myeloid cell lines (5)(6)(7). The activity of Ras depends on the ratio of bound GTP to GDP, a value that is determined by certain regulatory proteins: guanine nucleotide exchange factors such as Sos (8) and Vav (9), which activate Ras by catalyzing the exchange of bound GDP for GTP, and GTPase-activating proteins such as the NF-1 gene product (10) and p120 Ras-GAP , which deactivate Ras by stimulating the conversion of bound GTP to GDP (11). p120 Ras-GAP is a monomeric 120-kDa cytosolic protein that greatly accelerates the conversion of p21 Ras ⅐GTP to the confor-mationally inactive GDP-bound form (12,13). Accordingly, p120 Ras-GAP has been proposed as a negative regulator of Ras in cellular processes. This proposal is consistent with the finding that overexpression of p120 Ras-GAP blocks oncogenic transformation (14,15) and inhibits Ras-dependent cellular signaling (16). In addition, since the site of interaction of p21 Ras ⅐GTP with p120 Ras-GAP overlaps the effector region of Ras, it has been proposed that p120 Ras-GAP may also serve as an effector protein of active Ras (17,18). p120 Ras-GAP consists of a C-terminal domain that interacts with Ras and an N-terminal portion containing two Src homology (SH2) domains and one intervening SH3 domain (reviewed in Ref. 19). The p21 Ras ⅐GTP-p120 Ras-GAP complex may interact with downstream effectors via these Src-homology domains. SH2 domains bind phosphotyrosine residues and have been implicated in the interactions of tyrosine kinases of the Src family with their effector proteins (20,21). Several members of the Src family of protein-tyrosine kinases (e.g. c-Src and Lck) have been shown to associate with and phosphorylate p120 Ras-GAP (22)(23)(24). In fibroblasts transformed by cytoplasmic and receptor-like tyrosine kinases (25), p120 Ras-GAP forms complexes with two phosphotyrosine-containing proteins, p62 and p190 Rac-GAP . The complex with p62 appears to involve the SH2 domains of p120 Ras-GAP (26,27), and the formation of the p120 Ras-GAP -p190 Rac-GAP complex is dependent on phosphorylation (28), suggesting that it too involves the p120 Ras-GAP SH2 domains. The complex itself has diminished p120 Ras-GAP activity in vitro (28). Tyrosine-phosphorylated p62 and p190 Rac-GAP have been detected in anti-p120 Ras-GAP immunoprecipitates from various activated or transformed B-and T-cell lines (30,31), and were proposed to participate in signal transduction in these cells. The p120 Ras-GAP -associated p190 Rac-GAP stimulates the intrinsic GTPase activity of Rac1 and Rac2, which are Rho family guanine nucleotide-binding proteins (32). In human phagocytes, Rac2 plays a significant role in the activation of the respiratory burst oxidase (33)(34)(35). The human oxidase consists of two membrane-bound components (gp91 phox and p22 phox ) and at least three cytosolic components (p47 phox , p67 phox , and Rac2) that after cellular activation assemble to form the active enzyme (reviewed in Ref. 36). p190 Rac-GAP has been shown to inhibit the Rac2-dependent activation of the respiratory burst oxidase in a cell-free system (37), pointing to a role for this p120 Ras-GAP -associated protein in the regulation of the oxidase. The same oxidase has also been found in normal human B-cells and EBV-transformed 1 B lymphocytes (38,39), and Rac2 plays a significant role in the activation of the lymphocyte oxidase (33). Since we were able to detect p190 Rac-GAP in p120 Ras-GAP immunoprecipitates from cultured EBV-transformed B lymphocytes and since there have been several reports of the use of antisense oligonucleotides in these cells (33,40,41), we used antisense oligonucleotides to study the effects of p120 Ras-GAP (and possibly associated proteins) on O 2 . production. In this report we describe the results of this study.
Cell Culture and Incubation with Oligonucleotides-The human EBV-transformed B lymphocyte cell line EW was a generous gift from Dr. Peter Newburger (University of Massachusetts School of Medicine). Cultures were maintained under subconfluent conditions in RPMI 1640 supplemented with 10% heat-inactivated (55°C, 30 min) fetal calf serum, 2 mM glutamine, and fungizone. Cells were kept in a humidified incubator at 37°C under 6% CO 2 . Viability of the lymphoblastoid cells was routinely Ͼ90% as determined by trypan blue exclusion. An 18-mer antisense oligodeoxynucleotide sequence complementary to a sequence spanning the published ATG initiation region of the human p120 Ras-GAP message (43) was designed with the help of the OLIGO program (45) to optimize for sites with maximum melting temperature with minimum self-complementarity and minimum dimer formation. Unmodified oligodeoxynucleotides with the antisense sequence 5Ј-GCCATCATGTT-GAAGCCG-3Ј, the corresponding sense sequence 5Ј-CGGCTTCAACAT-GATGGC-3Ј, and a scrambled sequence designed using a random number generator (5Ј-CACGTACGGCTAAGTGCT-3Ј) were synthesized in the core lab facility of The Scripps Research Institute and purified by twice precipitating them from 0.5 M NaCl with ethanol. Their purity was routinely checked by analytical reversed phase HPLC (C8 column, 0.1-50% acetonitrile in 100 mM triethylammonium acetate, pH 7.0), detecting at 260 nm. Analytical 19% urea-PAGE (44) showed that Ͼ90% of the oligonucleotides were full length.
Low passage cells 2 were washed 3 times in Opti-MEM and then incubated with 0.1 mM oligonucleotides for 24 h in serum-free Opti-MEM at a starting cell density of 4 ϫ 10 5 /ml. The cells were then washed once in HBSS and used either for immunoblotting or for the measurement of O 2 . production.
Stability of Oligonucleotides-The stability of the oligonucleotides under the chosen incubation conditions in the presence and absence of cells was evaluated by HPLC. For this purpose, 50 M oligonucleotide was incubated at 37°C in serum-free Opti-MEM medium with or without 5 ϫ 10 5 EBV-transformed lymphocytes under tissue culture conditions. At the times indicated, 0.5-ml portions of incubation mixture were withdrawn, and DNA was purified by extracting twice with phenol/chloroform. After washing twice with ether, 0.2-ml portions of the aqueous layers were analyzed by HPLC as described above. The quantity of oligonucleotide was determined from the peak height. In every case, the mobility of the extracted oligonucleotide was identical to that of the unincubated oligonucleotide.
Immunoblotting-After culture for 24 h in the presence of oligonucleotide, the cells were isolated by centrifugation (12,000 ϫ g, 30 s at room temperature) and then lysed by incubation for 15 min on ice in Triton lysis buffer (25 mM HEPES (pH 7.5) containing 150 mM NaCl, 1% Triton X-100, 10% glycerol, 2 mM EDTA, 2 mM EGTA, 10 g/ml leupeptin, 10 g/ml aprotinin, 5 g/ml pepstatin, and 1 mM phenylmethylsulfonyl fluoride). The lysates were centrifuged for 15 min at 13,000 ϫ g, and the supernatants were assayed for protein content using the micro-BCA protein assay (Pierce). p120 Ras-GAP was isolated from supernatants containing equal quantities of protein either by acetone precipitation (2 volumes of acetone (2-4 h at Ϫ20°C) followed by centrifugation) or by immunoprecipitation (monoclonal anti-p120 Ras-GAP antibody and protein G-Sepharose beads, washing the immunoprecipitates 3 times with Triton lysis buffer before release from the beads). The precipitated proteins were dissolved by boiling in Laemmli sample buffer and then separated by SDS-PAGE (7.5% gel, Laemmli buffer) and subjected to immunoblotting to detect p120 Ras-GAP (monoclonal anti-p120 Ras-GAP at 1:2000 dilution) and oxidase subunits (rabbit polyclonal anti-p47 phox at 1:5000 dilution and anti-p67 phox at 1:1000 dilution), using the oxidase subunits as standards for protein loading. The proteins were visualized with alkaline phosphatase-conjugated goat anti-mouse or anti-rabbit antibodies with nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate as phosphatase substrate. Experiments in which antisense p120 Ras-GAP levels were not diminished as compared with sense and scrambled p120 Ras-GAP levels were disregarded.
Turnover of p120 Ras-GAP in EBV-transformed B Lymphocytes-EBVtransformed B lymphocytes were washed 3 times in methionine/cysteine-free DMEM (DMEMϪ) and then incubated in the same buffer supplemented with 10 mM HEPES for 30 min at 37°C. The cells were then isolated by centrifugation, suspended in DMEMϪ plus Tran 35 Slabel (0.2 mCi/ml), and incubated for a further 60 min at 37°C. The radiolabeled cells were then washed in complete DMEM (DMEM plus cysteine/methionine) and cultured in complete DMEM containing 5% fetal calf serum. At the intervals noted, 1-ml portions of the culture were withdrawn, and the cells were isolated by centrifugation and then lysed by suspension in Triton lysis buffer. p120 Ras-GAP in the lysates was immunoprecipitated as described above. The washed immune complexes were resuspended in Laemmli sample buffer, and the proteins were separated by 7.5% SDS-PAGE. After electroblotting on nitrocellulose, the quantities of p120 Ras-GAP in the different samples were established as approximately equal by immunoblotting (see above), and the amount of 35 S in each p120 Ras-GAP band was measured with a phosphor imager.
Measurement of O 2 . Production-After incubation in the presence of the oligonucleotides, the lymphoblasts were isolated by centrifugation, washed once in HBSS, and resuspended in prewarmed (37°C) HBSS containing 1 mM CaCl 2 and 1 mM MgCl 2 . Viability of the resuspended cells as determined by trypan blue exclusion was Ͼ90%, a value not different from that obtained with untreated cells. After counting the cells and determining their viability, 2 ϫ 10 5 live cells from each incubation were transferred to 96-well microtiter luminescence plates (Labsystems Inc.) and prewarmed at 37°C. After 5 min, luminol (10 M final concentration), horseradish peroxidase (50 g/ml), and either phorbol myristate acetate (1 g/ml), Pansorbin (Calbiochem) (1 mg/ml), or ionomycin (0.1 g/ml) as indicated were added to each well, yielding final incubation mixtures of 250-l volume. Chemiluminescence from each well was then recorded at 37°C over 10 consecutive 1-min intervals, using the "repeatscan" mode of a microplate luminometer (Labsystems Luminoskan RS, Helsinki, Finland) and collecting data for 0.8 s. The results were reported as relative luminescence units.

RESULTS
Using a monoclonal antibody against human p120 Ras-GAP , we were able to immunoprecipitate p120 Ras-GAP , together with the associated tyrosine-phosphorylated proteins p62 and p190 Rac-GAP , from resting and stimulated EBV-transformed Bcells (data not shown). Blotting of cell lysates with antibodies against p120 Ras-GAP and the oxidase components p47 phox and p67 phox showed that p120 Ras-GAP could be easily detected in lysates of even small amounts of cells (Ͻ5 ϫ 10 5 ). The t1 ⁄2 for the turnover of p120 Ras-GAP was 8 -12 h, as measured in pulsechase experiments with 35 S-labeled cysteine/methionine (Fig. 1). In view of this observation and reports that oligonucleotides are readily taken up by EBV-transformed B lymphocytes (33,40,41) we decided to use antisense oligonucleotides to study the effect of p120 Ras-GAP on the respiratory burst oxidase activity in this cell type.
Subconfluent B-cell cultures were cultured without serum for 24 h in the presence of p120 Ras-GAP -specific antisense, sense, or scrambled oligonucleotides. The cells were then assayed for 1) expression of p120 Ras-GAP , and 2) O 2 . production. In preliminary experiments with oligonucleotides end-labeled 2 High passage cells were refractory to the effects of the antisense nucleotide. with 32 P, we showed that the added oligonucleotides were taken up by the cells in a time-dependent fashion, with maximum uptake at 4 h as previously reported (40) (data not shown). In addition, we found by HPLC that the oligonucleotide was not significantly degraded after culture for 24 h in medium alone and that even when cultured in the presence of the cells ϳ20% of the oligonucleotide initially added was present after 24 h (Fig. 2).
When the cells were incubated for 24 h in the presence of the antisense oligonucleotide, expression of p120 Ras-GAP was inhibited (Fig. 3). Immunoblots of cells treated with the antisense oligonucleotide showed a substantial reduction in the level of p120 Ras-GAP as compared with untreated cells or cells treated with sense or scrambled oligonucleotides. By contrast, the levels of p47 phox and p67 phox were not significantly altered by incubation with antisense oligonucleotide, as revealed by parallel immunostaining of blots with specific antibodies against these two proteins. These results indicated that the suppression of p120 Ras-GAP expression by the antisense oligonucleotide was specific and did not affect the expression of two irrelevant proteins related to the respiratory burst oxidase in this cell type. O 2 . production by EBV-transformed B-cells treated with antisense, sense, or scrambled oligonucleotides was detected by luminol-enhanced chemiluminescence of cells stimulated with phorbol, Pansorbin, or ionomycin. When antisense-treated cells were stimulated with phorbol we observed a significant enhancement of O 2 . release in comparison with cells preincubated with sense or scrambled sequences (Fig. 4, top). This enhancement always correlated with a reduced expression of p120 Ras-GAP and was not observed in cells incubated in the presence of the oligonucleotide for 3 days, conditions under which the expression of p120 Ras-GAP had returned to levels seen after incubation with sense or scrambled oligos (data not shown). At 10 min, burst enhancement averaged 147 Ϯ 10% (S.E.) in cells treated with 0.1 mM antisense oligonucleotide. When O 2 . formation was triggered by Pansorbin, however, which acts through the cell's surface immunoglobulin receptors, the burst-enhancing effect of antisense oligonucleotides was not seen (Fig. 4, middle). O 2 . production by Pansorbin-treated cells was much lower than that of phorbol-treated cells. To determine whether the difference in the effect of antisense oligonucleotides on phorbolstimulated versus Pansorbin-stimulated cells was due to the weakness of the latter stimulus, antisense experiments were carried out on cells treated with ionomycin, another weak stimulus. Experiments with the protein kinase C inhibitor GFX showed that ionomycin, like phorbol but unlike Pansorbin, activated O 2 . production by EBV-transformed B lymphocytes through a protein kinase C-dependent mechanism (Fig. 5). Although levels of O 2 . production in response to ionomycin were FIG. 1. Turnover of p120 Ras-GAP in EBV-transformed B lymphocytes. The t1 ⁄2 for p120 Ras-GAP was determined as described under "Results." Top, a representative immunoblot showing p120 Ras-GAP precipitated from aliquots of cultured cells at various times. Bottom, 35  . production in response to ionomycin was increased to 133 Ϯ 8% (S.E.) of control by the antisense oligonucleotide. p120 Ras-GAP has been connected to tyrosine kinase-mediated signaling pathways, where it is thought to function downstream of activated Src-like protein tyrosine kinases or growth factor receptor tyrosine kinases (49,56). In T-cells, p120 Ras-GAP is bound to Lck and becomes specifically phosphorylated by this Src-like tyrosine kinase (23). Since Lck has been reported to play a significant role in EBV-induced growth of transformed B lymphocytes (40), and since antisense oligonucleotides against Lck inhibited the growth of EBV-transformed B-cells, we examined the effects of the p120 Ras-GAP antisense oligonucleotide on this parameter. We found that when p120 Ras-GAP expression was significantly reduced, the cell count after 24 h was low (Fig. 6). Despite this decrease in cell proliferation, O 2 . production was increased in cells incubated with the antisense oligonucleotide.

DISCUSSION
Antisense strategy is a useful approach for elucidating the role of different proteins in signal transduction (46). Using an 18-mer antisense oligonucleotide spanning the start codon of the p120 Ras-GAP message, we examined the function of p120 Ras-GAP in the control of respiratory burst oxidase activity in EBVtransformed B lymphocytes. In agreement with Cheung and Dosch (40), we observed that these cells took up oligonucleotides in a time-dependent fashion and that the antisense oligonucleotide, but not the sense or scrambled oligonucleotide, was able to reduce the expression of p120 Ras-GAP . When p120 Ras-GAP expression was reduced, O 2 . production by phorbol-and ionomycin-stimulated lymphocytes increased, although O 2 . production by protein A (Pansorbin)-stimulated cells remained unchanged. Both phorbol-and ionomycin-stimulated O 2 . production, but not the Pansorbin-triggered burst, appear to be dependent on the activation of protein kinase C, as indicated by results obtained with GFX. The antisense results, therefore, point to a specific role for p120 Ras-GAP in the protein kinase C-dependent activation of the respiratory burst oxidase in EBV-transformed B-cells. Several laboratories have presented evidence showing a link between p120 Ras-GAP and protein kinase C-dependent signaling pathways. In T-cells, an increase in the amount of active p21 Ras that was induced by phorbol but not by receptor-dependent stimuli has been related to a protein kinase C-mediated inactivation of p120 Ras-GAP (7). In fibroblasts, overexpression of p120 Ras-GAP inhibited the activation of mitogen-activated protein kinase by phorbol but not by receptor-dependent stimuli (16), presumably by causing a selective blockade of signals from activated protein kinase C due to the inactivation of p21 Ras . Thus, for both cell types, p120 Ras-GAP has been proposed as a negative regulator in a protein kinase C-dependent signaling pathway. Our results point to a similar role for p120 Ras-GAP as a down-regulator of a protein kinase C-dependent signaling pathway leading to an activated NADPH-oxidase.  Among the factors that regulate the activity of small guanine nucleotide-binding proteins is p190 Rac-GAP , a 190-kDa protein that accelerates the deactivation of guanine nucleotide-binding proteins of the Rho class (32). p190 Rac-GAP is of interest in regard to oxidase activation because Rac2, one of its targets, is known to participate in the activation of the respiratory burst oxidase both in the cell-free system (47,48) and in whole cells (33). In accord with this idea, p190 Rac-GAP was shown to inhibit Rac2-dependent oxidase activation in the cell-free system (37,49). p190 Rac-GAP is known to associate with p120 Ras-GAP in many cell types (25,26,30,49) and using anti-phosphotyrosine antibodies, we found a phosphorylated 190-kDa protein, which could be identified as p190 Rac-GAP , in immunoprecipitates of p120 Ras-GAP from EBV-transformed B lymphocytes (data not shown). The possibility therefore exists that the deficiency of p120 Ras-GAP induced by the antisense oligonucleotide resulted in a deficiency of p190 Rac-GAP , leading to persistent oxidase activity due to a delay in the deactivation of Rac2. Unfortunately, we were unable to detect p190 Rac-GAP on Western blots of extracts from oligonucleotide-treated cells, so we have no direct evidence as to the effect of the antisense oligonucleotide on p190 Rac-GAP concentrations in the transformed lymphocytes. However, our finding that phorbol-dependent but not receptor-(i.e. Pansorbin-) dependent oxidase activation is affected by the antisense oligonucleotide is difficult to explain on the basis of an indirect effect of p120 Ras-GAP mediated through p190 Rac-GAP , because Rac2 is involved in both phorbol-and receptor-mediated activation of the respiratory burst oxidase (33) so that an effect occurring via p190 Rac-GAP should have involved both phorbol-activated and receptor-activated O 2 . production. The results therefore suggest a direct effect of p120 Ras-GAP on the oxidase activation cascade.
Rap1A is a member of the Ras superfamily of guanine nucleotide-binding proteins. Its participation in the regulation of oxidase activity is suggested by its copurification with cytochrome b 558 (50) and by the finding that overexpression of Rap1A mutants fixed in either the active or inactive conformation decrease oxidase activity in transfected B lymphocytes (51). Since p120 Ras-GAP has been shown to interact with the putative effector region of rap1A in vitro (52), p120 Ras-GAP could participate in the physiological regulation of Rap1A by competitively interfering with its deactivation by Rap1-GAP, an 85-95-kDa Rap1-specific GTPase-activating protein (53). Accelerated deactivation of Rap1A in the p120 Ras-GAP -depleted cell could therefore account for the effect of the antisense oligonucleotide on oxidase activity. Whether a mechanism based on Rap1A deactivation could account for the difference between phorbol-activated and receptor-activated cells remains to be determined.
A third alternative is presented by the interactions between the N-terminal SH2 domain of p120 Ras-GAP and various proteins containing phosphorylated tyrosines (19,25,54,55). The inhibition in the growth of EBV-transformed B-cells by the p120 Ras-GAP antisense oligonucleotide could result from interactions between p120 Ras-GAP and proteins other than p190 Rac-GAP or p62, the two proteins that are commonly found in p120 Ras-GAP immunoprecipitates. The Src-like tyrosine kinase Lck is a candidate of particular interest, since Lck forms a complex in vitro with the phosphorylated N-terminal SH2 domain of p120 Ras-GAP (23) and is thought to play a significant role in the transformation of EBV-transformed B lymphocytes (40). The inhibition of cell growth caused by p120 Ras-GAP antisense oligonucleotides might somehow be connected to the interaction between p120 Ras-GAP and Lck.