ras Protein Activity Is Essential for T-cell Antigen Receptor Signal Transduction*

In a Jurkat cell model of T-cell activation an inter- leukin-2 promoter/reporter gene construct was activated by antigen receptor agonism in combination with the lymphokine interleukin- 1. Antigen receptor sig- nals could be mimicked by suboptimal activation of protein kinase C (PKC) with phorbol esters in combi- nation with calcium mobilization by an ionophore. In cotransfection experiments, oncogenic ras obviated the need for PKC stimulation but did not replace either the calcium signal or interleukin-1. Activated ras expres- sion also replaced the requirement for PKC stimulation in activation of the T-cell transcription factor NF-AT. A dominant inhibitory ras mutant specifically blocked antigen receptor agonism, indicating that ras activity is required for antigen receptor signaling. In addition, an inhibitor of PKC blocked both activated ras and phorbol ester stimulation, suggesting a role for ras upstream of PKC.

A crucial event in some subsets of T-cells is the induction of interleukin-2 (IL-2) and IL-2 receptor (IL-2R) expression in the target cells. Expression of IL-2 and IL-2R results in autocrine growth stimulation and is coincident with commitment to differentiation. The signals required for induction of 1L-2 expression reflect the signals required for T-cell activation (for review, see . IL-2 expression is strictly dependent on PKC activity and calcium mobilization (Macchia et al., 1990;Baldari et al., 1991). In addition, the macrophage-derived lymphokine IL-1 (for review, see Dinarello, * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
IL-2 expression is controlled primarily at the level of transcription. In the absence of stimulation, the promoter is virtually inactive (reviewed in Ullman et al. (1990)). Of the several transcription factors involved in IL-2 gene expression, AP-1, NFKB, and NF-AT are known to respond to T-cellactivating stimuli. NFKB and AP-1 are stimulated by PKC activity and respond to IL-1 but do not respond to calcium and function in the absence of extracellular calcium (Emmel et al., 1989;Muegge et al., 1989;Serfling et al., 1989;Espel et al., 1990;Baldari et al., 1992). NF-AT, however, requires both PKC activity and calcium mobilization for activation, which may explain the absolute dependence of IL-2 promoter activity on extracellular calcium. The NF-AT transcription factor functions as a heterodimer. In the human lymphoma cell line Jurkat, PKC activation results in induction of expression of a nuclear component of NF-AT. Increases in intracellular free calcium result in translocation of a second, constitutively expressed component from the cytoplasm to the nucleus where an active complex is formed (Flanagan et al., 1991).
Recent evidence has suggested a role for the ras family of small GTP-binding proteins in transduction of TCR signals. First, Downward et al. (1990) have shown that TCR triggering or PKC activation results in an increase in the GTP-bound active form of ras proteins in Jurkat cells and in peripheral blood lymphocytes. Second, Baldari et al. (1992) have shown that a constitutively active form of Ha-ras expressed in the murine thymoma line, EL4, can replace the requirement for PKC in activation in IL-2 promoter induction. Here we report that expression of constitutively active Ha-ras protein can replace PKC in the activation of the IL-2 promoter and of a multimer of NF-AT in the human lymphoma cell line Jurkat. We show that TCR induction of NF-AT is inhibited by expression of the dominant negative mutant ras protein N17 (Feig and Cooper, 1988) indicating that ras proteins are essential for TCR signal transduction. Interestingly, direct stimulation of PKC by phorbol ester was considerably less sensitive to N17 than TCR activation, suggesting that ras proteins may function both upstream and downstream of PKC. In support of this an inhibitor of PKC activity, but not an inhibitor of cyclic nucleotide-dependent kinases, blocked oncogenic ras activation of NF-AT.

MATERIALS AND METHODS
Reagents-Recombinant human IL-10 from Escherichia coli (Casagli et al., 1989) was used at 1 ng/ml. PMA (Sigma) and ionophore A23187 (Boehringer Mannheim) were dissolved in dimethyl sulfoxide at 100 pg/ml and 10 mg/ml, respectively. PHA (Wellcome Diagnostics, Dartford, United Kingdom) was dissolved in phosphate-buffered saline at 1 mg/ml. For protein determination the kit BCA from Pierce (Rockford, IL) was used. Acetyl coenzyme A (Boehringer Mannheim) ras Mediates TCR Signal Transduction and ['4C]chloramphenicol (Amersham International, United Kingdom) were used for chloramphenicol acetyltransferase (CAT) assays as described by Gorman et al. (1982). Restriction and modification enzymes (Boehringer Mannheim and Promega, Madison, WI) were used according to the manufacturers' instructions. Nucleotide sequence determinations were performed using the Sequenase kit (U. S. Biochemical Co.). Polymerase chain reaction was carried out using a Gene Amp kit (Perkin-Elmer Cetus Instruments). Oligonucleotides were synthesized on an Applied Biosystems 391 DNA synthesizer using cyanoethyl phosphoramidate chemistry. Plasmids-IL-2/CAT is a derivative of the Bluescript SK plasmid (Stratagene, San Diego) containing a 2,000-base pair fragment of the human IL-2 promoter upstream of the CAT gene (Macchia et al., 1990). NF-AT/CAT contains a trimer of the NF-AT binding site of the IL-2 promoter upstream of the CAT gene (Emmel et al., 1989). The pDOL-expression vector has been described elsewhere (Korman et al., 1987). pSVT7/hIL-lR has been described elsewhere (Heguy et al., 1991).
T24-ras containing the murine leukemia virus long terminal repeat cloned from pDOL-has been described in Baldari et al. (1992). The N17 inhibitory mutant (Feig and Cooper, 1988) was generated by subcloning the XbaI fragment of T24-ras containing the first exon into the XbaI cloning site of pEMBL18 (Dente et al., 1985) and first removing the activating mutation by site-directed mutagenesis of G to T at codon 12 (valine to glycine) on the single-stranded plasmid as described (Kunkel et al., 1987). A clone containing the wild-type sequence identified by DNA sequencing was then used for a second round of site-directed mutagenesis of G to A at position 17 (serine to asparagine). The XbaI fragment encoding glycine 12 and asparagine 17 was finally cloned into T24-ras to replace the original homologous fragment.
Cell Culture, Transfections, and CAT Assays-The human lymphoma line Jurkat was maintained in RPMI supplemented with 2 mM L-glutamine, 20 mM HEPES (pH 7.9), and 10% heat-inactivated (56 "C for 90 min) fetal calf serum (Boehringer Mannheim). The medium used for transfections included 200 units/ml penicillin (Farmitalia, Italy). Transfections were carried out using a modification of the DEAE-dextran procedure as described (Banerji et al., 1983) using 1 X 10' cells and 10-15 pg/sample for IL-Z/CAT transfections or 1 X IO6 cells and 1-2 pg of DNA/sample for NF-AT/CAT transfections.
T o avoid error caused by variation in transfection efficiency, comparisons of stimuli or inhibitors were carried out on identical aliquots of single pools of transfected cells. When the effect of T24 or N17 was tested, a single transfection mix was prepared containing all components except the plasmid to be tested. This mixture was then aliquoted, and the test plasmid or an appropriate quantity of control plasmid was added. Cells were allowed to recover for 24 h hefore activation. The PKC inhibitor H7 or HA1004 (Hidaka et al., 1987) was added when required 30 min before stimulation. After incubation for 8-10 h cells were collected by centrifugation, washed in Trisbuffered saline, resuspended in 0.25 M Tris hydrochloride (pH 7.5), and extracted by freeze-thawing. Equal amounts of proteins, determined according to a modification of the method described by Lowry (1951), were used for CAT assays. CAT enzyme activity was assayed using ['4C]chloramphenicol according to Gorman et al. (1982). Autoradiographs of the thin layer chromatograms were scanned using an Ultrascan XL enhanced lasrr densitometer.

RESULTS
The human lymphoma cell line Jurkat has been widely used to study IL-2 promoter activation by mitogenic stimuli (Fujita et al., 1986;Siebenlist et al., 1986;Durand et al., 1988). A reporter construct containing the bacterial gene for CAT under the control of the IL-2 promoter (IL-2/CAT) can be activated in Jurkat cells by the TCR agonist PHA in combination with high concentrations of PMA (>lo ng/ml) ( Fig.  1A and Baldari et al. (1991)). Singly these stimuli have no effect. In addition the IL-2 promoter can be activated by a combination of PMA (10 ng/ml) and the calcium ionophore A23187. Treatment with high concentrations of PMA (10 ng/ ml) results in translocation from the cytoplasm to the plasma membrane and consequent activation of more than 80% of PKC' (Nishizuka, 1984;. This massive PKC * C. T. Baldari  activation overcomes the requirement for accessory signals but does not reflect the physiological role of signals such as IL-1, which has been shown to be independent of PKC (Abraham et al., 1987;Macchia et al., 1990). Jurkat cells lack IL-1 receptors; however, they can be converted to IL-1 responsiveness by cotransfection with a construct capable of expressing T-cell type IL-1 receptors (Baldari et al., 1991;Heguy et al., 1991). In this case, the IL-2/CAT construct responds to a combination of TCR stimulation by PHA and IL-1R stimulation by IL-1 (Fig. 1A).
PHA activation of the TCR results in a low level activation of PKC and mobilization of calcium (for review, see  and Gardner (1989)). The PHA signal can therefore best be mimicked with suboptimal concentrations of PMA (<1 ng/ml), which result in approximately 15% of PKC associated with the membrane (data not shown), plus the calcium ionophore A23187. This combination, like PHA alone, had very little effect on IL-2/CAT activity; however it synergized effectively with IL-1 (Fig. lA, right panel). Thus in this model of T-cell activation the three signals, suboptimal PKC activation, calcium mobilization, and an IL-1-induced signal, are required for IL-2 promoter activation.
Activated ras Protein Replaces PKC Activation"T24 is an oncogenic form of human Ha-ras which is activated by a mutation of a glycine to valine at position 12 (Santos et al., 1982). This mutation results in reduced GTPase activity a n d accumulation of GTP-bound active ras protein. Activation of the IL-2/CAT construct in Jurkat cells cotransfected with the IL-1R construct, and a construct capable of expressing T24ras (Baldari et al., 1992) no longer required PMA treatment.
In the presence of T24, treatment with IL-1 plus A23187, which in control cells had no effect, resulted in significant activation of IL-2/CAT ( Fig. 1R and Table I).
No CAT activity was detected after treatment with IL-1 alone, A23187 alone, or PMA plus IL-1 in T24-cotransfected cells. The T24ras therefore replaced the PMA stimulus but the not the   -, below detection.
ND, not determined.
-calcium-mediated signal. We conclude that constitutively active ras can, a t least in part, replace the PKC-activating component of TCR signaling. ray Activation of IL-2 Expression Is Mediated by NF-AT-TCR activation of the IL-2 enhancer is mediated by two short nucleotide sequences known as antigen receptor response elements (Durand et al., 1988). One of these sequences binds the activated T-cell-specific transcription factor, NF-AT (Shaw et al., 1988). As was reported previously (Emmel et al., 1989), a construct containing the CAT gene under the control of a synthetic promoter containing several copies of the binding site for NF-AT was activated by PHA alone or a combination of PMA and A23187 ( Fig. 2A and Table 11). No accessory signals were required, and PMA alone or A23187 alone had no effect. In cells cotransfected with the IL-1R construct, IL-1 treatment either alone or in combination with PMA or A23187 had no effect on NF-AT/CAT activity; however, PHA stimulation was somewhat enhanced ( Fig. 2A and Table 11). The NF-AT factor therefore responds to TCR triggering in the absence of accessory signals. Presumably, in the context of the IL-2 enhancer, NF-AT alone is not sufficient to mediate activation of the IL-2 promoter by TCR signals.
Cotransfection of NF-AT/CAT with the T24 construct in part obviated the need for PMA treatment. In the presence of T24-ras, A23187 alone induced significant CAT activity, whereas PMA alone had no effect ( Fig. 2B and Table I). We conclude that T24-ras activation of the IL-2 promoter is mediated, at least in part, by NF-AT.
ras Activity Is Necessary for TCR Signaling-The data presented so far show that activated ras can replace PKC activation of the IL-2 promoter and NF-AT. To address the question of whether ras proteins play an essential role in TCR signal transduction we have tested the effect of inhibition of endogenous ras activity on PHA induction of NF-AT/CAT. A single amino acid substitution from serine to asparagine at position 17 in the ras protein not only inactivates the protein but also results in a molecule capable of inhibiting wild type ras function (Feig and Cooper, 1988). The mutation interferes with GTP-GDP exchange, and thus the mutated protein does not inhibit oncogenic ras activity (Medema et al., 1991;Stacey et al., 1991). Using site-directed mutagenesis we have modified the T24 construct to remove the Val"-activating mutation and to introduce the Asn" mutation. This construct (N17) was cotransfected with the NF-AT construct into Jurkat cells which were subsequently treated with either PHA or PMA plus A23187.
The results of an experiment in which aliquots of a single pool of competent cells were transfected with a fixed amount of reporter plasmid and varying amounts of N17 DNA are shown in Fig. 3A. At ratios of 0.5 pg or 1 pg of Nl7 DNA to ras Mediates TCR Signal Transduction lo6 cells, CAT activity obtained on activation with PHA was significantly reduced, while PMA/A23187-induced CAT activity was not affected. 1.5 pg of N17 DNA resulted in reduction of both the PHA-and the PMA/A23187-induced CAT activity. Table 111 shows the results of laser densitometry of thin layer chromatograms from several experiments. In all experiments, PHA induction was considerably more sensitive to N17 than was PMA/A23187 induction. When different experiments were compared, transfection efficiency, as measured by CAT activity in a control sample, had a greater influence on N17 efficacy than the amount of DNA used. Thus when transfection was more efficient, a greater inhibition of both PHA and PMA/A23187 stimulation was observed (see Table 111). The level of N17 protein expressed and consequently its ability to inhibit endogenous ras proteins presumably depend on the amount of N17 DNA taken up by the cells.
Coexpression of T24-ras with N17 overcame the reduction of the PHA signal (Fig. 3B), confirming that the N17 effect is caused by inhibition of ras proteins and not by any nonspecific effects of the N17 DNA.
As can be seen by comparison of Fig. 2A with Fig. 3A, the level of CAT activity induced by PHA compared with PMA/ A23187 was significantly less when lower amounts (0.3 pg/ lo6 cells) of the NF-AT/CAT plasmid were used in the transfections, suggesting that at the higher concentration (0.7 pg/ lofi cells) CAT activity was reaching saturation in this system (see also Table 11). Thus at lower doses of reporter gene PHA delivers a weaker signal than PMA/A23187, which may explain the difference in sensitivity to N17. Fig. 4 shows the effect of N17 transfection on activation of NF-AT/CAT by suboptimal PMA (1 ng/ml) in combination with A23187 compared with its effect on PHA or PMA (10 ng/ml) and A23187. No significant difference was observed in the sensitivity of PHA or suboptimal PMA/A23187 to N17, whereas optimal PMA/A23187 was less sensitive. The results of three independent experiments are shown in Table IV.
The preceding experiments indicate that ras activity is essential for TCR signaling. Maximal stimulation of PKC with PMA was less sensitive to inhibition by N17 than PHA stimulation. This may simply reflect the difference in strength of the stimulations; however, it may indicate a role for ras proteins upstream of PKC, inhibition of which could be overcome by sufficient direct stimulation of PKC by PMA. In support of this, H7, an isoquinoline sulfonamide which blocks PKC and cyclic nucleotide-dependent protein kinases (Hidaka et al., 1984), completely blocked activation of NF- ' Percent of chloramphenicol conversion in the control sample.
*Experiment shown in Fig. 3.   AT/CAT by T24/A23187, whereas HA1004, a similar compound which blocks preferentially cyclic nucleotide-dependent kinases, had no effect (Fig. 5). Ha-ras may therefore function upstream of PKC.

DISCUSSION
A role for ras proteins in the transduction of extracellular signals has long been suggested by their structural similarity to the trimeric G-nucleotide-binding proteins, their location on the inner flap of the plasma membrane, and their role in oncogenesis (for review, see Barbacid, 1987). This view has been considerably strengthened by reports of ras involvement in nerve growth factor-induced phosphorylation of MAP kinases (Wood et al., 1992;Thomas et al., 1992) and insulininduced gene expression (Medema et al., 1991). A role for ras in TCR signal transduction was suggested by reports that TCR engagement in Jurkat cells and peripheral blood leukocytes resulted in an increase in the active GTP-bound form of the protein (Downward et al., 1990) and that a constitutively active ras mutant in EL4 cells could contribute part of the signals necessary for IL-2 gene activation (Baldari et al., 1992). We have used Jurkat cells as a model system to analyze further the role of ras in mediating the TCR-derived signals and show a direct involvement of ras proteins in transduction of part of the complex TCR signals which lead to activation of IL-2 gene expression. TCR engagement results in phosphatidylinositol metabolism to produce inositol triphosphate and diacylglycerol. Inositol triphosphate causes a transient increase in intracellular free calcium, and diacylglycerol activates PKC. A further sustained increase in calcium from extracellular stores is mediated by an as yet poorly understood mechanism (reviewed in ), Finkel et al. (1991), Truneh et al. (1985). These two signals can be mimicked by directly activating PKC with phorbol esters and by increasing intracellular free calcium by means of calcium ionophores. In this work we have used a combination of suboptimal concentrations of PMA and a calcium ionophore which, like TCR engagement, have little effect on IL-2 expression but which synergize effectively with the macrophage-derived lymphocyte-activating factor IL-1.
Our results show that constitutively active ras oncoprotein can substitute for suboptimal PKC activation but not for IL-1 or calcium-mediated signals in activation of the IL-2 promoter. In addition we have shown that the NF-AT transcription factor mediates, at least in part, this effect. A dominant negative ras mutant reduced NF-AT activation, indicating that ras proteins are essential for TCR signaling. N17 inhibited TCR signaling under conditions which had little effect on direct activation of PKC by saturating concentrations of phorbol esters. This and the fact that an inhibitor of PKC blocked NF-AT activation by constitutively active Ha-ras indicate that ras protein functions upstream of PKC. This interpretation is in agreement with other reports that oncogenic ras mutants result in an increase in diacylglycerol, a physiological activator of PKC (Wolfman and Macara, 1987;Price et al., 1989) and that down-regulation of PKC abrogates oncogenic ras-induced mitogenesis of Swiss 3T3 fibroblasts (Lacal et al., 1987).
In contrast, Downward et al. (1990) have shown that both TCR engagement and phorbol esters cause an increase in GTP-bound ras protein in Jurkat cells. In addition, phorbol ester stimulation of tyrosine phosphorylation of MAP kinases is inhibited by N17 ras (Thomas et al., 1992;Wood et al., 1992). On the basis of these data the authors suggest that PKC activates ras. It should be noted that others (Leevers and Marshall, 1992) have reported that oncogenic ras activation of MAP kinases is abrogated by down-regulation of PKC.
We also observed inhibition of optimal PMA stimulation in cells transfected with high concentrations of N17. In addition, suboptimal concentrations of PMA were as sensitive to N17 as PHA stimulation. One possible explanation of these apparently contrasting results is that ras proteins play a role both upstream and downstream of PKC. Perhaps significant is the observation by Downward et al. (1990) that phorbol ester activates predominantly N-ras and Ki-ras in lymphocytes. It cannot be excluded that PKC activates ras in a feedback loop capable of amplifying the intracellular signal.
Both active ras and PKC have been shown to activate AP-1 (Boyle et al., 1991;Bin6truy et al., 1991)) NFKB (Ghosh and Baltimore, 1990;Baldari et al., 1992) and NF-AT (Flanagan et al., 1991;this work). The recent demonstration that the PKC-inducible subunit of NF-AT is in fact AP-1 (Jain et al., 1992) may explain the ras activation of NF-AT. These three transcription factors are known to be involved in induction of IL-2 expression (reviewed in Ullman et al. (1990)). We propose that TCR activation of ras results in activation of PKC, which in turn results in suboptimal stimulation of each of these factors. This stimulation is insufficient to activate IL-2 expression fully. Activation of the second NF-AT subunit by calcium (Flanagan et al., 1991) and additional activation of NFKB by an as yet unknown IL-1-induced signal (Espel et al., 1990;Baldari et al., 1992) would then result in the formation of fully competent transcriptional initiation complexes.