Inhibition of the c-Jun N-terminal Kinase/AP-1 and NF-κB Pathways by PICOT, a Novel Protein Kinase C-interacting Protein with a Thioredoxin Homology Domain*

Protein kinase C-θ (PKCθ) is a Ca2+-independent PKC isoform that is selectively expressed in T lymphocytes (and muscle), and is thought to play an important role in T cell receptor-induced activation. To gain a better understanding of the function and regulation of PKCθ, we have employed the yeast two-hybrid system to identify PKCθ-interacting proteins. We report the isolation and characterization of a cDNA encoding a novel 335-amino acid (37.5-kDa) PKCθ-interacting protein termed PICOT (for PKC-interactingcousin of thioredoxin). PICOT is expressed in various tissues, including in T cells, where it colocalizes with PKCθ. PICOT displays an N-terminal thioredoxin homology domain, which is required for the interaction with PKC. Comparison of the unique C-terminal region of PICOT with expressed sequence tag data bases revealed two tandem repeats of a novel domain that is highly conserved from plants to mammals. Transient overexpression of full-length PICOT (but not its N- or C-terminal fragments) in T cells inhibited the activation of c-Jun N-terminal kinase (but not extracellular signal-regulated kinase), and the transcription factors AP-1 or NF-κB. These findings suggest that PICOT and its evolutionary conserved homologues may interact with PKC-related kinases in multiple organisms and, second, that it plays a role in regulating the function of the thioredoxin system.

Members of the protein kinase C (PKC) 1 family of intracellular serine/threonine kinases play critical roles in the regulation of cellular differentiation and proliferation in many cell types and in response to diverse stimuli, including hormones, neurotransmitters, and growth factors (1)(2)(3). The PKC family consists of 11 known mammalian members that are expressed in a wide variety of tissues and cell types. Based on sequence similarities, domain structures, and cofactor requirements, PKC isoenzymes can be grouped into three subfamilies. 1) The Ca 2ϩ -dependent conventional enzymes, consisting of PKC-␣, -␤I, -␤II, and -␥, contain three conserved domains, namely the diacylglycerol/phorbol ester binding C1 domain, which contains two repeats of a cysteine-rich zinc finger, the phospholipid-and calcium-binding C2 domain, and the catalytic C3 and C4 domains.
2) The Ca 2ϩ -independent enzymes (PKC-␦, -⑀, -, -and -) are termed novel PKCs. The C2-like N-terminal domain of these enzymes can bind acidic phospholipids but not Ca 2ϩ . 3) A third PKC subfamily, termed atypical PKCs, includes PKCand -/ that possess a single cysteine-rich domain, lacking the ability to bind phospholipids or phorbol esters. PKC activity is regulated by defined cofactors that interact with specific regions of the regulatory domain as well as transphosphorylation by serine/threonine kinases and autophosphorylation. The activation is accompanied by a conformational change that releases the basic pseudosubstrate region from the catalytic cleft of the kinase domain. In addition, interaction with specific proteins, termed receptors for activated PKC, that function as selective scaffolds for activated PKCs at discrete subcellular compartments, play a role in activation of PKC (4).
In the course of isolating novel PKC isoforms that play a role in T cell antigen receptor (TCR) signaling, we isolated PKC, a Ca 2ϩ -independent PKC isoform characterized by a unique tissue distribution, i.e. in skeletal muscle, lymphoid organs, and hematopoietic cell lines, particularly in T cells (5). PKC plays a selective role in the activation of the c-Jun N-terminal kinase (JNK)/AP-1 pathway and the interleukin-2 gene in T cells (6 -8), and to colocalize with the TCR complex to the contact site between antigen-specific T cells and antigen-presenting cells (9), where it participates in the formation of a supramolecular activation cluster (10).
In order to a gain a better understanding of the function and regulation of PKC in TCR signaling, we have employed the yeast two-hybrid system to identify PKC-interacting proteins.
Here we describe the isolation of a novel PKC-interacting protein having a unique domain structure, which consists of an N-terminal thioredoxin (Trx)-homologous domain followed by two tandem repeats of a novel, evolutionary conserved protein domain. We have termed this protein PICOT (for PKC-interacting cousin of thioredoxin). Transient overexpression of PI-COT inhibited the activation of JNK and two transcription factors i.e. AP-1 and NF-B, induced by PKC or by combinations of T cell-activating stimuli, suggesting that this novel protein plays an important role in regulating T cell activation and the function of PKC.

MATERIALS AND METHODS
Antibodies and Expression Plasmids-The anti-CD3 monoclonal antibody (mAb) was affinity-purified from culture supernatants of the OKT3 hybridoma as described (11). An anti-human CD28 mAb was obtained from PharMingen. The PKC-specific mAb and rabbit polyclonal antibody were from Transduction Laboratories (Lexington, KY) and Santa Cruz Biotechnology, Inc. (Santa Cruz, CA), respectively. The anti-c-Myc mAb (9E10) was purified from culture supernatants of the corresponding hybridoma by protein A-Sepharose chromatography, and the anti-hemagglutinin (HA) mAb (12CA5) was from Roche Molecular Biochemicals. Polyclonal goat anti-JNK and rabbit anti-ERK2 antibodies were obtained from Santa Cruz Biotechnology, and phospho-c-Junor phospho-ERK2-specific antibodies were from New England Biolabs (Beverly, MA). Fluorescein isothiocyanate-coupled secondary antibodies were from Pierce. A polyclonal rabbit anti-PICOT antiserum was generated using standard protocols. In brief, a peptide comprising amino acids 90 -108 of the deduced human PICOT sequence was coupled to keyhole limpet hemocyanin and injected into two rabbits. Both antisera recognized a protein band with the predicted electrophoretic mobility of PICOT (38 kDa) in immunoblots of cell lysates. For immunohistochemistry, a portion of the antiserum was affinity-purified on a Sepharose-coupled synthetic peptide column.
The full-length PKC cDNA with a C-terminal hexahistidine tag in the pEF mammalian expression vector has been described previously (12). An HA-tagged constitutively active calcineurin mutant consisting of the catalytic subunit of calcineurin (CnA), from which the calmodulin-binding and the autoinhibitory domains were deleted (CnA⌬CaM-AI) was a generous gift from Dr. M. Karin (University of California, San Diego), and was described (8). To construct human PICOT expression vectors, the corresponding full-length cDNA or fragments encoding the N-(residues 1-146) or C-(residues 133-335) terminal regions of PICOT were subcloned into pEF, and an N-terminal hexahistidine tag followed by a subsequent HA tag was added at the 5Ј end of the cDNA. HAtagged JNK1 and c-Myc-tagged ERK2 were cloned in pSR␣ and pcDNA3, respectively. The AP-1 and NF-B luciferase reporter plasmids were obtained from M. Karin.
Yeast Two-hybrid Screen-The yeast two-hybrid system used in this study was kindly donated by E. A. Golemis (Fox Chase Cancer Center, Philadelphia, PA), and has been described previously (14). cDNAs encoding full-length PKC or fragments comprising its regulatory (amino acids 1-378) or catalytic (amino acids 379 -706) domains were subcloned into pGilda (15) to generate in-frame fusion proteins with the LexA DNA-binding domain. These baits were used to screen a Jurkat T cell cDNA library as described (15). To map the PKC-binding domain of human PICOT, the cDNAs encoding the N-or C-terminal regions of PICOT were subcloned into the activation domain fusion vector pJG4 -5 (14) Cell Culture and Transfection-Simian virus 40 large T antigen (TAg)-transfected human leukemic Jurkat-TAg cells were grown in RPMI 1640 medium (Life Technologies, Inc.) supplemented with 10 mM HEPES, pH 7.5, 10 mM minimal essential medium with non-essential amino acids, 1 mM sodium pyruvate, 10% fetal bovine serum, and antibiotics. For expression of recombinant proteins, cells were transfected for 48 h with appropriate amounts of plasmids (usually 3-20 g total) by electroporation as described (11,12). In each experiment, cells in different groups were transfected with the same total amount of plasmid DNA by supplementing expression vector DNA with the proper amounts of the corresponding empty vector. Where indicated, the cells were stimulated with anti-CD3 and/or anti-CD28 antibodies, phorbol myristate acetate (PMA; 100 ng/ml) and/or ionomycin (1 g/ml), or irradiated with UV (312 nm) for 1 min at room temperature using a transilluminator (FBTI-88; Fisher Scientific) and cultured for an additional h. Cells were lysed in 1ϫ Nonidet P-40 lysis buffer (1% Nonidet P-40, 20 mM Tris-HCl, pH 7.5, 100 mM NaCl, 5 mM NaPP i , 5 mM NaF, 5 mM Na 3 VO 4 ) supplemented with protease inhibitors (Roche Molecular Biochemicals) for 10 min on ice. The insoluble material was removed by centrifugation.
Glutathione S-Transferase (GST) Fusion Proteins and in Vitro Binding Assays-The cDNAs encoding full-length human PICOT or its Nterminal or C-terminal fragments were subcloned into the bacterial GST fusion vector pGEX-5X-1 (Amersham Pharmacia Biotech). Expression and purification of the GST fusion proteins was performed as described (16). Cell lysates (usually from 1 ϫ 10 7 cells) were incubated with 10 g of GST fusion proteins coupled to 40 l of glutathioneagarose for 2 h at 4°C. The binding mixtures were washed extensively in 1ϫ Nonidet P-40 lysis buffer and analyzed by SDS-PAGE and Western blotting.
RT-PCR and Northern Blotting-Multi Tissue cDNA Panels (CLON-TECH) were screened by RT-PCR with a pair of PICOT-specific primers according to the instructions of the manufacturer. For amplification of a 1-kilobase pair PICOT fragment, 35 cycles were used, and the glyceraldehyde-3-phosphate dehydrogenase control was amplified using 25 PCR cycles. 20 g of total RNA from Jurkat cells was prepared using standard procedures. The PICOT plasmid isolated from the yeast twohybrid screen or a full-length glyceraldehyde-3-phosphate dehydrogenase cDNA were used for generation of 32 P-labeled probes using a commercial kit (Rediprime II, Amersham Pharmacia Biotech).
Subcellular Fractionation-Jurkat cells were lysed and separated into a membrane, cytosol, and detergent-insoluble fractions as described (12).
Immunofluorescence-Transfected or nontransfected Jurkat-TAg cells were left unstimulated, or stimulated with 100 ng/ml PMA for 10 min at 37°C. Cells were then spun down, washed with cold phosphatebuffered saline (PBS), fixed with 3.7% paraformaldehyde, and permeabilized in 0.05% saponin. Transfected cells were then stained with a polyclonal rabbit anti-PICOT antibody and an anti-PKC mAb. Nontransfected cells were stained with a polyclonal anti-PKC antibody and Alexa 488-conjugated, affinity-purified anti-PICOT antibody. Samples were then incubated with fluorescein isothiocyanate-conjugated secondary antibodies (Pierce) or Alexa 594 (Molecular Probes, Eugene, OR), respectively, and were subsequently washed four times with 1% bovine serum albumin in PBS. After the final wash, the cells were mounted on glass slides using a drop of Aqua-Poly/mount (Polysciences, Inc., Warrington, PA). Samples were viewed with a Plan-Apochromat 63ϫ lens on a Nikon microscope. Images were taken using a Bio-Rad MRC 1024 laser scanning confocal imaging system.
Immunoprecipitation and Immunoblotting-Lysates (1-2 ϫ 10 7 cells) were mixed with antibodies (1-2 g) for 2 h, followed by addition of 40 l of protein A/G Plus-Sepharose beads (Santa Cruz Biotechnology) for an additional h at 4°C. Immunoprecipitates were washed twice with 1ϫ Nonidet P-40 lysis buffer and twice with PBS (pH 7.2). After boiling in 20 l 2x Laemmli sample buffer, samples were subjected to SDS-PAGE and electrotransferred to nitrocellulose membranes (Bio-Rad). Membranes were immunoblotted with the indicated primary antibodies, followed by horseradish peroxidase-conjugated secondary antibodies. Bands were visualized by chemiluminescence (Amersham Pharmacia Biotech). When necessary, membranes were stripped by incubation in 62.5 mM Tris-HCl, pH 6.7, 100 mM 2-mercaptoethanol, 2% SDS for 1 h at 65°C, washed, and then reprobed with other antibodies as indicated.
In Vitro Kinase Assays-In vitro kinase assays of immunoprecipitated JNK or ERK2 were conducted as described. Briefly, washed JNK1 or ERK2 immunoprecipitates were assayed using 2 g of GST-c-Jun fusion protein or myelin basic protein as substrates, respectively, in 20 l of JNK or ERK2 kinase buffers containing 3 Ci of [␥-32 P]ATP (30 Ci/mmol, Amersham Pharmacia Biotech). Kinase reactions were incubated for 20 min at 30°C with gentle shaking, and were stopped by addition of 20 l of 2ϫ Laemmli buffer. Proteins were resolved by SDS/13% PAGE, transferred to nitrocellulose, and subjected to autoradiography. Substrate phosphorylation was quantified by PhosphorImager (Storm 860; Molecular Dynamics, Sunnyvale, CA) analysis. The nitrocellulose membranes were routinely reprobed with anti-JNK or -ERK2 antibodies to confirm equal expression levels of the immunoprecipitated kinases.
Reporter Assays-Transfected Jurkat-TAg cells were harvested, washed twice with PBS, and lysed in 100 l of lysis buffer (100 mM KPO 4 , pH 7.8, 1 mM dithiothreitol, 0.5% Triton X-100) for 10 min at room temperature. The lysates were then centrifuged (15,000 ϫ g, 5 min at 4°C). Fifty l of the supernatant were mixed with 100 l of assay buffer (17.5 mM glycylglycine, pH 7.8, 10 mM MgCl 2 , 5 mM ATP, 0.135 mM coenzyme A, 0.235 mM luciferin), and the luciferase activity determined in a luminometer (Monolight 2010; Analytical Luminescence Laboratory, Sparks, MD). The protein content was determined using the Bio-Rad protein assay. The final results were expressed as arbitrary relative luciferase units per microgram of protein.

Isolation of a cDNA Encoding a Novel PKC-interacting Pro-
tein-To identify proteins that interact with PKC in vivo, we performed three independent yeast two-hybrid screens using the full-length PKC, its regulatory domain, or its catalytic domain as baits to screen a Jurkat T lymphoma cDNA library fused to a cDNA encoding a transcription activation domain (14). Since expression of catalytically active forms of PKC as LexA fusion proteins was toxic to yeast (data not shown), a point-mutated cDNA encoding a catalytically inactive PKC (K409R) was used for the yeast two-hybrid screening. Approximately 5 ϫ 10 7 independent clones were screened with each bait. No clones growing on selective medium were obtained with the isolated regulatory or catalytic domains of PKC, respectively (Table I). Screening with the catalytically inactive, full-length PKC bait resulted in 17 primary positive colonies, which grew on selective medium and were ␤-galactosidasepositive when tested in a filter lift assay (Table I). Thus, the observed interaction in yeast requires full-length PKC, but does not depend on its intact catalytic activity.
Ten of the positive, PKC-interacting clones were found to contain cDNAs that specifically interacted with the bait. Partial sequence analysis revealed that these clones encoded identical cDNAs. The two longest cDNAs (ϳ1,250 base pairs in length) were sequenced, and found to contain an open reading frame (ORF) encoding a putative protein of 335 amino acids (Fig. 1a). Since the putative protein product contained a Trxhomologous domain (see below), it was named PICOT (for PKCinteracting cousin of thioredoxin) (see below). Northern blot analysis using one of the yeast two-hybrid clones as a probe indicated that a hybridizing mRNA of 1.5 kilobase pairs is expressed in a preparation of total RNA from Jurkat T cells (Fig. 1b), suggesting the isolated clones contained the complete ORF. A rapid amplification of cDNA ends PCR of a human thymus cDNA library failed to reveal additional upstream sequences, thus confirming that the isolated cDNAs were full-length.
Identification of PICOT Homologous Sequences-The isolated human PICOT cDNA sequence was used to conduct a search of the GenBank data base of expressed sequence tags (ESTs) for homologous sequences. This analysis identified an identical putative human ORF (accession no. h59799) as well as homologous mouse (accession no. aa009010) and rat (accession no. aa866363) clones. The corresponding plasmids were obtained from Genome Systems or Research Genetics, respectively, and sequenced. Using the BLAST algorithm, homologous sequences were also identified in Saccharomyces cerevi-siae (Swissprot ye04 yeast), Escherichia coli (Swissprot ydhd_ecoli), Haemophilus influenzae (Swissprot ydhd_haein), Caenorhabditis elegans (GenBank g3217992), and Arabidopsis thaliana (GenBank g3335374) (see Fig. 2).
Structure and Expression of PICOT-The ORF of the PICOT cDNA encodes a putative protein with a predicted molecular mass of 37.5 kDa (Fig. 1a). The codon for the first methionine is surrounded by a consensus Kozak sequence, but is not preceded by an in-frame stop codon. To confirm the expression of the putative protein encoded by the cDNA we have isolated, rabbit antisera were generated against a synthetic peptide corresponding to amino acids 90 -108 of PICOT, a hydrophilic sequence with a high surface probability. Immunoblotting of lysates from E. coli expressing full-length PICOT as a GST fusion protein (GST-PICOT) or from untransfected Jurkat cells revealed expression of an immunoreactive ϳ74-kDa protein in E. coli, and a ϳ38-kDa protein in Jurkat cells, consistent with the predicted size of PICOT (data not shown). This antiserum was used in subsequent expression studies (see below).
To determine the cellular localization of PICOT, we fractionated Jurkat T lymphoma cells and analyzed the cellular distribution of the protein. As shown in Fig. 1c, PICOT was almost exclusively localized in the cytosol, with additional, very low TABLE I Specific interactions between PKC and PICOT in the yeast two-hybrid system EGY48 yeast cells were cotransformed with expression vectors encoding various LexA DNA-binding domain and activation domain fusion proteins. Activation of the leu2 reporter gene was monitored by growth on leucine-deficient medium, and the activity of the lacZ reporter gene was monitored using a filter assay. DNA-binding domain hybrid "bait" Activation domain hybrid "prey" expression in the membrane fraction, but no detectable expression in the detergent-insoluble cellular fraction.
The tissue distribution of PICOT was analyzed by RT-PCR using a commercial cDNA panel derived from different human tissues. The expression of PICOT mRNA was more ubiquitous than that of PKC; nevertheless, it was not expressed in all tissues. It was abundant in heart, spleen, and testis, with low but detectable expression in the other tissues tested, including in the thymus and peripheral blood leukocytes; expression of the relevant mRNA in the lung, placenta, colon, and small intestine was very low (Fig. 1d).
Sequence Analysis of PICOT-Search of the GenBank nucleotide data base revealed a 29% amino acid identity and an additional 11% similarity between the N-terminal region of PICOT (residues 12-143) and the Trx family of proteins (Fig. 2,  a and b). However, the Trx homology domain of PICOT lacks the conserved Cys-Gly-Pro-Cys motif, which is essential for catalytic activity (17), and contains instead an Ala-Pro-Gln-Cys motif. The aforementioned EST data base-derived mouse and rat homologues of PICOT encode the corresponding full-length proteins. The deduced protein sequences show Ͼ99% identity to the human sequence, with the rat protein lacking residues 90 -148 of human or mouse PICOT (data not shown).
Analysis of the protein sequence of PICOT against sequences of putative proteins derived from genome sequencing projects revealed a highly conserved sequence motif of 84 amino acids. This motif represents a novel, previously unknown domain, which is found in all organisms searched, including nematodes, yeast, bacteria, viruses, and plants (Fig. 2c). The human, mouse, and rat PICOT proteins display two tandem repeats of this domain (Fig. 2, a and c), whereas the homologous proteins present in lower organisms contain only a single repeat. Thus, PICOT shows a discrete domain structure, consisting of an N-terminal Trx-like domain followed by two novel PICOT homology domains (Fig. 2a).
Association of PKC with PICOT in Intact T Cells and in Vitro-To confirm the interaction of PICOT with PKC, and analyze its specificity at the level of PKC, we first ascertained whether these two proteins are associated in intact T cells. Jurkat-TAg cells were cotransfected with PKC plus HA epitope-tagged PICOT expression vectors. When lysates from these cells were immunoprecipitated with an anti-HA mAb, PKC was found to coimmunoprecipitate with PICOT (Fig. 3a). An unrelated control antibody did not precipitate either of these two proteins. This interaction was confirmed by precipitating T cell lysates with a GST-PICOT fusion protein in vitro. Incubation of the recombinant protein, but not the control GST protein, with lysates from Jurkat cells transfected with PKC, PKC␣, or PKC, followed by immunoblotting with antibodies against the respective PKC isoforms, revealed that PICOT bound PKC and, to a lower extent, PKC (Fig. 3b). Under the same conditions, no interaction with PKC␣ was detected, suggesting that PICOT displays some selectivity with regard to its ability to associate with PKC isoforms.
The Trx-like Domain of PICOT Is Sufficient for Interaction with PKC-In order to further define the structural basis for the interaction of PICOT with PKC, we expressed the PICOT N-terminal region corresponding to its Trx-like domain (residues 1-146) and its C-terminal fragment representing the two tandem PICOT homology domains (residues 133-335) as GST fusion proteins, and used these recombinant proteins to precipitate cell lysates from PKC-transfected Jurkat cells. As shown in Fig. 3c, the N-terminal, Trx-like domain of PICOT was sufficient for PKC binding, although it was less effective than the full-length protein. In contrast, the C-terminal region of PICOT did not display any detectable binding to PKC. In yeast, only full-length PICOT associated with PKC (Table I), suggesting that the affinity of the interaction is lower in yeast. This difference may reflect a favorable conformation of PICOT for this interaction in vitro (where it was expressed as a GST fusion protein) compared with yeast (where PICOT was expressed as a fusion protein with a transcription activation domain).
Intracellular Localization of PICOT and PKC-In order to determine the relative localization of PICOT versus PKC in situ, transfected or untransfected Jurkat cells were stained with PICOT-and PKC-specific antibodies, and analyzed by confocal microscopy. In unstimulated cells, which were transfected with both PICOT and PKC, both proteins colocalized to a distinct cytoplasmic area under the plasma membrane (Fig.  4, a and b), and their colocalization was evident when the two individual images were overlaid (Fig. 4c). PMA stimulation caused translocation of both PICOT and PKC to a more extended membrane (or submembrane) area, with colocalization of the two proteins still evident (Fig. 4, d-f).
We also analyzed the localization of the relevant endogenous protein in untransfected cells. The quality of this analysis was lower due to the lower expression of the proteins and the fact that it was necessary to use a polyclonal anti-PKC antibody, which produces a less satisfactory staining than that obtained with the monoclonal antibody used in the transfected cells. Nevertheless, endogenous PICOT and PKC also appeared to colocalize in the same area of the cell, with an overall staining pattern similar to the one observed in the transfected cells (Fig.  4, g-i).
Selective Inhibition of JNK Activation by PICOT-As PICOT was identified on the basis of its interaction with PKC, we examined the effect of transient PICOT overexpression on the activation of the stress-activated protein kinase JNK, since it represents a selective target of PKC in the TCR/CD28 signaling pathway (7,8). Jurkat-TAg cells were cotransfected with different combinations of expression plasmids encoding PI-COT, PKC and/or constitutively active calcineurin (CnA⌬CaM-AI). The latter plasmid was used since calcineurin cooperates with PKC in the activation of JNK and the interleukin-2 promoter (7,8).
As predicted, transient overexpression PKC caused activation of the cotransfected epitope-tagged JNK reporter, and this effect was somewhat augmented by CnA coexpression (Fig. 5a,  lanes 3 and 5 versus lane 1 in the two upper panels). When the cells were additionally cotransfected with a PICOT expression vector, the PKC-or PKC/CnA-induced JNK activation was significantly reduced (Fig. 5a, lanes 4 and 6). Overexpression of PICOT alone also seemed to reduce the basal JNK activity (lane 2). Densitometric analysis of the phospho-c-Jun bands revealed that PICOT reduced the PKC-or PKC/CnA-induced JNK activation by 70% and 55%, respectively. Immunoblotting of immunoprecipitated JNK with a specific antibody confirmed that all groups expressed a similar level of transfected JNK (Fig. 5a, third lane from the top). Similarly, immunoblotting of cell lysates from the same cells with antibodies specific for PKC, PICOT, or HA-CnA confirmed the expected overexpression of the corresponding proteins in the cells (Fig. 5a, three  bottom panels).
In order to examine the selectivity of this inhibitory effect of PICOT, we also assessed the effect of transient PICOT overexpression on the activation of another MAP kinase, i.e. ERK2, using similar in vitro immune complex kinase assays. As reported before (8), ERK2 can be non-selectively activated by both PKC and PKC␣ (Fig. 5b, lanes 3 and 5 versus lane 1 in the  upper panel). Coexpression of PICOT alone did not reduce ERK2 activity and, in some experiments, even enhanced it (data not shown). Similarly, coexpression of PICOT with PKC (lane 4) or PKC␣ (lane 6) did not inhibit the PKC-induced ERK2 activity. Immunoblotting with an ERK2-specific antibody confirmed the equivalent expression levels of ERK2 in most groups, with the exception of the two PKC-transfected groups, which displayed lower ERK2 expression, thereby making the PKC-induced ERK2 activation even more pronounced than the apparent level. PKC and/or PICOT were properly overexpressed in the transfected cells. Thus, the PICOT-mediated inhibition of PKC-induced MAP kinase activation is specific for JNK and, furthermore, PICOT alone can induce ERK, but not JNK, activation.
Next, we determined the effect of PICOT overexpression on JNK activation induced by several stimuli, including a physiological stimulus provided by anti-CD3/CD28 antibodies. As reported previously (8,18), stimulation of Jurkat cells with this antibody combination, a combination of PMA plus ionomycin, or by UV irradiation, all induced marked activation of the cotransfected JNK1 reporter when compared with the unstimulated cells (Fig. 5c, lane 1). Coexpression of PICOT in the same cells reduced the basal or anti-CD3/CD28-induced JNK activity by Ն80%, but had a much smaller effect on the PMA/ ionomycin-or UV-induced kinase activity (lane 2). Furthermore, the inhibitory effect of PICOT overexpression was very similar to that caused by overexpressing a dominant-negative PKC (-K/R) mutant (lane 3), supporting the notion that PI- COT exerts its inhibitory activity by interfering with the cellular function of PKC.
PICOT Inhibits the Activation of AP-1 and NF-B in T Cells-Our earlier studies demonstrated that, among several PKC isoforms tested, PKC functions as a selective AP-1 activator via a Ras-dependent pathway (6). Therefore, we assessed the effect of transient PICOT overexpression on the activation of an AP-1 reporter plasmid. As expected, PMA stimulation or transient overexpression of PKC caused a marked increase of AP-1 activity, and the constitutively active plasmid (PKC-A/E) was more active than the wild-type kinase in that regard (Fig.  6a). When the cells were cotransfected with increasing amounts of the PICOT expression plasmid, a dose-dependent inhibition of the basal or PKC-induced AP-1 activity was observed. Five g of the PICOT plasmid reduced the basal activity of AP-1 by ϳ90%, and the activities induced by wild-type or constitutively active PKC (in the absence of PMA stimulation) by ϳ95 and ϳ60%, respectively. The failure of PICOT to inhibit AP-1 activity in PMA-stimulated cells may reflect activation of AP-1 by other, endogenous PKC isoforms that are not sensitive to the inhibitory effect of PICOT, e.g. PKC␣ (see Fig. 3b).
Further analysis of the structural requirements for the inhibition of AP-1 activation by PICOT revealed that the fulllength protein was necessary. Thus, while native PICOT was capable of inhibiting AP-1 activation induced by constitutively active PKC by 60%, the isolated N-or C-terminal fragments of PICOT displayed minimal inhibitory activity (Fig. 6b).
In order to determine the effect of transient PICOT overexpression on the activity of another transcription factor that is known to be induced by stress signals, we assessed the activation of the transcription factor NF-B by the combined stimulation of anti-CD3 plus anti-CD28 antibodies. In the absence of PICOT coexpression, this combination induced a ϳ5-fold activation of NF-B. Transient expression of a lower dose of PICOT (5 g) caused a minimal reduction of activity, but at the higher dose (10 g plasmid DNA), NF-B activity was reduced by 63% (Fig. 6c). As in the case of AP-1, this inhibition required the full-length PICOT protein (data not shown). The selectivity of this effect is indicated by the fact that the PMA-induced activation of NF-B was not inhibited under the same conditions. DISCUSSION Several recent studies have relied on the yeast two-hybrid system to isolate PKC-interacting proteins, which were found to represent either substrates or regulators of distinct PKC isoforms (19 -21). Here, we have used a similar strategy to isolate proteins that interact with PKC, a Ca 2ϩ -independent PKC isoform that is expressed selectively in T cells (5) and has been implicated in T cell-specific functions (6 -10, 22). This study reports the isolation and partial characterization of a novel PKC-interacting protein termed PICOT. Our results establish the expression of the corresponding mRNA and protein in T and other cells, and the association of PICOT with PKC in intact T cells and in vitro which requires the Nterminal, Trx-homologous domain of PICOT. Since PICOT interacts with kinase-inactive PKC, and our preliminary results indicate that PICOT is not phosphorylated by PKC in vitro (data not shown), PICOT most likely does not represent a PKC substrate. Furthermore, PICOT associated in vitro not only with PKC, but also with PKC. Nevertheless, its interaction with PKC is not promiscuous since it did not associate with another PKC isoform, i.e. PKC␣. The degree of specificity of the interaction between PICOT and distinct PKC isoforms, particularly in intact cells, remains to be established.
The stoichiometry of the association between PICOT and PKC appears to be low. Although the two proteins could be coimmunoprecipitated from transfected, overexpressing cells, we could not reproducibly demonstrate coimmunoprecipitation of the relevant endogenous proteins. This may reflect the use of lysis conditions that are unfavorable for the maintenance of this association, or the requirement of other cellular factors (e.g. lipids or adaptor proteins) for optimal interaction between these two proteins. In addition, the association may take place in specific sites within the cell, and only under specific conditions. However, analysis by confocal microscopy revealed overlap, albeit incomplete, between the intracellular localization of PICOT and PKC, even in untransfected cells. Nevertheless, our data indicating that PICOT regulates cellular functions mediated by PKC or physiological stimuli in T cells suggest that the association between PICOT and PKC is physiologically relevant. Thus, the activation of two important elements in the TCR/CD28 signaling cascade leading to interleukin-2 production, i.e. JNK and AP-1, both of which are selectively activated by PKC (6 -8), was inhibited by coexpressed PICOT. This effect was selective and did not reflect a general inhibition of cellular functions, since the activation of another MAP kinase, ERK2, was not inhibited by PICOT. Since JNK positively regulates AP-1 activity by phosphorylating two regulatory serine residues in the activation domain of c-Jun (18), it is not surprising that the PKC-mediated activation of AP-1 was also inhibited by PICOT.
The JNK/AP-1 pathway is not the only target for inhibition by PICOT, since this novel protein also inhibited the anti-CD3 plus anti-CD28-induced activation of another transcription factor, i.e. NF-B. Since both of these pathways are commonly activated in response to stress signals and inflammatory stimuli (23-28), our findings suggest that PICOT may regulate stress-induced signaling pathways in other cell types and organisms. Although the N-terminal, Trx-homologous region of PICOT was sufficient to mediate PKC binding, inhibition of AP-1 or NF-B activation required the intact protein. This raises the possibility that The N-terminal domain of PICOT binds regula-tors and/or effectors, whereas the C-terminal region mediates the biological functions of this protein. At present, we do not know whether PICOT homologues expressed in lower organisms also interact with, and potentially regulate, protein kinases. Studies to examine the effects of PICOT on stress responses triggered by different stimuli are currently in progress.
Of particular interest is the novel domain that appears as two tandem repeats at the C-terminal region of PICOT. We have provisionally termed this domain PICOT homology (PIH) domain (Fig. 2c). This domain, which hitherto has not been recognized, is highly conserved in evolution from plants to mammals. However, in contrast to PICOT, only one repeat of this domain is found in lower organisms, where it constitutes most of the putative protein encoded by the corresponding EST sequences. The high degree of conservation of this domain suggests that it plays an important, yet to be identified, role in cellular functions. Studies are in progress to isolate PIH domain-interacting proteins.
Although the exact mechanism by which PICOT inhibits the activation of the JNK/AP-1 pathway or NF-B remains to be elucidated, the homology of PICOT to Trx is of particular interest. The evolutionary conserved Trx system has evolved to protect cells from damage mediated by reactive oxygen species, which are generated as part of a cellular defense mechanism against invading pathogens (17,29,30). Various cellular insults, i.e. mitogens, inflammatory stimuli, UV or ionizing radiation, ischemia, phorbol ester, and hydrogen peroxide up-regulate the expression of Trx and induce its translocation to the nucleus. Trx exerts both extracellular and intracellular functions, including its extracellular ability to protect cells from tumor necrosis factor-or Fas-mediated apoptosis (17). Trx is known to promote the DNA binding and transcriptional activ- ities of AP-1 and NF-B as well as the activity of the estrogen receptor (17,31,32). It mediates these effects by reducing cysteine residues in the p50 subunit of NF-B, the two components of AP-1, i.e. c-Jun and c-Fos, and Ref-1, an endonuclease that participates in AP-1 activation (17,31,32). These modifications are necessary for the binding of these transcription factors to their cognate DNA sequences in the promoter regions of various genes (17).
Since PICOT does not possess the essential catalytic motif of Trx (Cys-Gly-Pro-Cys) and, in fact, lacks the first of the two cysteine residues in the catalytic center, it almost certainly lacks Trx enzymatic activity. This notion is supported by the findings that mutation of either or both cysteine residues in the catalytic center of Trx abolishes its enzymatic (33) and mitogenic (34) activities, and converts the mutated proteins into competitive inhibitors of Trx reductase (34). Another type of peroxidase enzymes, the peroxiredoxins, contain a single conserved cysteine residue, but do not share homology with Trx (35). This structural feature of PICOT raises the interesting possibility that PICOT functions as an endogenous antagonist of Trx or its activating enzyme, Trx reductase, via its ability to compete for substrate binding. Since the Trx system, which is highly conserved throughout evolution, plays an important role in regulating the intracellular redox state, which is critical for both cell viability and proliferation (17,29,30), actions that are mediated in part by regulation of the transcription factors NF-B and AP-1 (30), it is possible that PICOT has a highly conserved role in regulating the Trx system. Our findings, that transient PICOT overexpression inhibits the activation of AP-1 and one of its upstream activators (JNK), as well as NF-B, are consistent with this putative function.
Finally, since the production of reactive oxygen species (27) and the concomitant induction of genetic programs that mediate defense mechanisms against pathogenic agents (36,37) are highly conserved in evolution, including in plants (38), an intriguing possibility is that PICOT and its evolutionary conserved homologues regulate innate immunity defense mechanisms. Furthermore, the conservation of Trx system (17,29,30) and the PKC superfamily (39,40) during evolution, and the findings that PKC homologues play a role in plant defense mechanisms against viral pathogens (13,41), raise the intriguing possibility that an axis consisting of PICOT/PKC/Trx homologues plays a general and well conserved important regulatory role in cellular functions. Thus, the biological significance of PICOT and its putative homologues may extend well beyond its role in T cell activation. Additional studies aimed at elucidating the physiological role and regulation of PICOT are likely to shed light on these notions.