C114 is a novel IL-11-inducible nuclear double-stranded RNA-binding protein that inhibits protein kinase R.

We have identified a cDNA (named C114) that encodes novel transcripts induced by IL-11 in mouse 3T3 L1 cells. Northern analysis of RNAs from multiple mouse tissues detects two C114 transcripts of approximately 1.0 and approximately 2.0 kb with the highest expression in liver, testis, brain, and kidney. The C114 cDNA contains an open reading frame of 187 amino acids with a predicted mass of 21 kDa. Three putative nuclear localization signals are predicted at amino acids 83-88, 126-131, and 167-178. Using green fluorescent protein (GFP)-C114 fusion plasmids, amino acids 126-131 are shown to be essential for the nuclear localization of C114. An arginine-rich region (amino acids 98-143) spanning the nuclear localization signals (amino acids 126-131) exhibits a double-stranded RNA (dsRNA) binding activity. Competition experiments with different RNA homopolymers demonstrate that C114 preferentially binds to poly(I.C). Similar to other dsRNA-binding proteins, C114 binds to the dsRNA-activated protein kinase, protein kinase R (PKR), via dsRNA-binding domains of PKR and the N-terminal region of the C114 protein. In vitro kinase assays indicate that C114 inhibits PKR activation via a dsRNA-independent mechanism. Overexpression of C114 protein inhibits the induction of eIF-2alpha phosphorylation following poly(I.C) treatment. This is the first demonstration of a novel PKR modulator induced by a gp130 superfamily cytokine that may play a role in cytokine-mediated biological functions.

IL-11 is a bone marrow fibroblast-derived cytokine with a variety of in vitro biological activities within the hematopoietic, lymphopoietic, hepatic, adipose, bone, and neuronal systems. It was originally isolated from a primate bone marrow-derived stromal cell line based on its ability to stimulate the proliferation of an IL-6-dependent mouse plasmacytoma cell line, T1165 (1). Meanwhile, Kawashima et al. (2) isolated a cDNA encoding a novel adipogenesis inhibitory factor that inhibits adipogenesis of mouse 3T3 L1 cells. The cDNA sequence of this factor is identical to IL-11 (2). Using biological assays and 125 I-IL-11 binding analysis, we previously demonstrated the existence of functional IL-11 receptors on 3T3 L1 mouse preadipocytes (3). Despite the participation of IL-11 in diverse physiological processes, its precise role in cell growth and differentiation remains unclear.
Protein kinase R (PKR) 1 is a 68-kDa serine/threonine kinase that manifests two distinct kinase activities, one for its own phosphorylation and the other for the phosphorylation of other substrates. Once activated, PKR becomes autophosphorylated and acquires the capacity to phosphorylate its well studied substrate, eIF-2␣. Phosphorylation of eIF-2␣ inhibits protein synthesis by impairing the activity of guanine nucleotide exchange factor eIF-2B (4). As a serine/threonine kinase found in a latent state in most cells, it plays an important role in cellular antiviral defense. In addition, PKR has been implicated in other cellular processes including apoptosis (5), cellular transformation (6), differentiation (7,8), splicing (9), and transcription (10). PKR also functions as a signal transducer in signaling pathways activated by different stimuli, including dsRNA, IL-3, interferon-␣, interferon-␥, IL-1, tumor necrosis factor, and platelet-derived growth factor, through a variety of effectors including NF-B, IRF-1, ATF-2, STAT3, and eIF-2␣ (4,11). Although PKR has been implicated in different cytokine actions, its involvement in IL-11 signaling has not been reported. Meanwhile, the physiological events and signaling pathways that trigger PKR activation by cellular RNA or proteins are not well established (4). Several PKR modulators have been identified. PKR inhibitors include cellular proteins such as P58, HIV-1 transactivating response (TAR) RNA-binding protein (TRBP), ribosomal protein L18, and NF90 and its homologues (12)(13)(14). On the other hand, other cellular proteins such as PKR activating protein (PACT) and PKR associate protein X (RAX) function as PKR activators (15). Many of these proteins also contain well defined dsRNA-binding domains (DRBDs).
In this study, we identified a novel IL-11-inducible transcript, C114, in 3T3 L1 cells. C114 is a nuclear protein exhibiting higher expression in the nucleolus. C114 interacts with dsRNA and also binds and inhibits the activity of PKR. We propose that this novel cellular protein acts as a negative regulator of PKR.

MATERIALS AND METHODS
Antibodies and Reagents-3T3 L1, NIH-3T3, and HEK 293 cells were cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and penicillin/streptomycin. Anti-FLAG monoclonal M2 antibody was from Sigma, and anti-V5 monoclonal antibody was from Invitrogen. The anti-eIF-2␣ and phosphospecific anti-eIF-2␣ antibodies were from BIOSOURCE International (Camarillo, CA). Recombinant human IL-11 was kindly provided by Genetics Institute (Cambridge, MA). Various polynucleotide resins were from Amersham Biosciences.
PCR-based cDNA Subtraction-Total cellular RNA was isolated from starved or IL-11-stimulated (500 ng/ml) mouse 3T3 L1 cells using TRIzol reagent (Invitrogen), and poly(A) RNA was purified using oligo(dT) beads (Invitrogen). cDNA synthesis and subtraction were done using PCR-Select cDNA subtraction kit (Clontech) following manufacturer's instructions. mRNA of IL-11-treated 3T3 L1 cells was used as a tester, and mRNA of untreated cells was used as a driver. Subtracted PCR products were cloned into pGEM-T Easy vector (Promega) to obtain a subtraction library. DNA fragments were prepared from individual cDNA clones by PCR and used as [␣-32 P]dCTP-labeled hybridization probes for Northern blot analysis. Total RNA extracted from IL-11-treated or untreated 3T3 L1 cells was prepared as mentioned above. RNA was electrophoresed in a 1.0% agarose gel containing formaldehyde and transferred onto a nylon membrane (Hybond Nϩ, Amersham Biosciences). After a cDNA fragment was confirmed by Northern blot to show a differential expression pattern, the full-length cDNA was cloned from a mouse liver cDNA library (Stratagene) by rapid amplification of cDNA ends. Poly(A)ϩ RNA blot of mouse tissues (Mouse MTN blot, Clontech) was also used for Northern analysis.
Cellular Localization of C114 Protein-The entire coding region and various mutant forms of C114 cDNA were amplified by PCR using primers containing a EcoRI or BamHI site and cloned into pEGFP (Clontech) so that the coding sequence continued in-frame with that of the enhanced green fluorescent protein (EGFP). These pC114-EGFP hybrid plasmids were transfected into NIH 3T3 cells by LipofectAMINE transfection reagent (Invitrogen). Cells were grown at 37°C in Dulbecco's modified Eagle's medium plus 10% fetal bovine serum for 48 h and examined by fluorescence microscopy to detect the distribution of C114-EGFP proteins. Mock pEGFP vector and mitochondrial protein-fused EGFP constructs were also transfected into NIH 3T3 cells.
Polynucleotide Binding Studies-In vitro transcription/translation reaction of C114 was carried out using the TNT-coupled reticulocyte lysate system (Promega) in the presence of [ 35 S]methionine (16). Radiolabeled C114 protein was incubated with various resins (Amersham Biosciences), which were first washed and resuspended in wash buffer 1 (20 mM Tris, pH 7.5, 5 mM ␤-mercaptoethanol, 4 mM Mg(C 2 H 3 O 2 ) 2 ⅐4H 2 O, 200 mM KCl, 0.5% Nonidet P-40). Equal aliquots of 35 S-labeled C114 protein were added to 40 l of a 1:1 volume of each resin and allowed to incubate at room temperature for 30 min. The resins were spun down, washed three times with 500 l of wash buffer 1, and washed twice with 500 l of wash buffer 2 (20 mM Tris, pH 7.5, 5 mM ␤-mercaptoethanol, 4 mM Mg(C 2 H 3 O 2 ) 2 ⅐4H 2 O, 50 mM KCl). The proteins bound to beads after the wash were analyzed by SDS-PAGE followed by autoradiography. For quantitation, bound proteins were eluted by the addition of an equal volume of 2ϫ Laemmli loading buffer, and equal aliquots were subjected to scintillation counting to determine the amount of proteins bound to each resin. For analysis, the percentage of proteins bound to each resin was calculated, taking into consideration the specific activity of each deletion construct, and comparison was made relative to the levels of the wild-type protein bound to each resin. For competitive binding studies, analyses were performed as above in the presence of competitor polynucleotides. Various concentrations of competitors (either poly(I⅐C) or poly(dI⅐dC)) ranging from 0.05 to 100 g were added to the in vitro translated 35 S-labeled C114 protein followed by the addition of resins. Analysis of bound proteins was carried out as above.
Construction of Eukaryotic Expression Vector Encoding C114 with a FLAG Epitope-To generate a FLAG epitope-tagged version of C114 for expression in mammalian cells, the coding region of C114 was amplified by PCR from C114/pGEM-T using 5Ј-AGGAATTCTATGGCGAGTCCT-GCTGCC-3Ј and 5Ј-ACGGATCCTCGCCCCATGATGACGAC-3Ј as the forward and reverse primer with a BamHI and EcoRI restriction site, respectively. The 600-bp PCR product was ligated in-frame with a FLAG tag into the pCMV-2 vector.
Proteins were transferred to polyvinylidene difluoride membranes and probed with monoclonal anti-V5 or anti-FLAG (1:5,000) followed by appropriate horseradish peroxidase-conjugated secondary antibody and ECL detection (Amersham Biosciences).
Pull-down Assays-35 S-labeled wild-type or mutated forms of PKR were in vitro translated using TNT-coupled reticulocyte system from Promega (18). FLAG epitope-tagged C114 protein was immunoprecipitated from transfected HEK 293 cell lysates with the M2 anti-FLAG (Sigma) and Protein A-agarose (Roche Molecular Biochemicals) as described above. Twenty l of the in vitro translated 35 S-labeled proteins was incubated with FLAG-tagged C114 protein in immunocomplexes containing Protein A-agarose and anti-FLAG antibody in 500 l of pull-down buffer (20 mM HEPES-KOH, pH 7.5, 100 mM KCl, 1 mM dithiothreitol, 0.5 mM EDTA, 10% glycerol, and 0.5% Nonidet P-40) at 4°C for 2 h. The resin was washed four times with 500 l of pull-down buffer. Proteins were eluted with 40 l of Laemmli sample buffer, boiled for 5 min, and resolved by SDS-PAGE. The proteins were transblotted to a polyvinylidene difluoride membrane followed by autoradiography and processed for Western blotting with M2 anti-FLAG antibody. Five l of in vitro translated 35 S-labeled proteins was also resolved by SDS-PAGE to visualize the input proteins.
Kinase Assays-HEK 293 cells were transfected with V5-tagged PKR constructs. In some samples, PKR-expressing plasmid was co-transfected with various forms of FLAG-tagged C114 plasmids (see Fig. 2B). Twenty-four hours after transfection, cells were washed with phosphate-buffered saline, pelleted, and lysed in an equal volume of lysis buffer. After centrifugation at 14,000 rpm for 10 min, the supernatants were collected. PKR was immunoprecipitated from aliquots containing 100 g of total protein using an anti-V5 monoclonal antibody in radioimmune precipitation buffer (50 mM Tris, pH 7.4, 50 mM KCl, 400 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, 100 units/ml aprotinin, 0.2 mM phenylmethylsulfonyl fluoride, 20% glycerol, 1% Triton X-100) at 4°C overnight. Protein A-agarose was added for an additional hour followed by washing four times with 500 l of TNEN buffer (20 mM Tris, 100 mM NaCl, 1 mM EDTA, 0.5% Nonidet P-40, 0.1 M phenylmethylsulfonyl fluoride, 100 M vanadate, 1 mM NaF, pH 8.0) and twice with activity buffer (20 mM Tris, pH 7.5, 50 mM KCl, 2 mM MgCl 2 , 2 mM MnCl 2 , 200 units/ml aprotinin, 0.1 mM phenylmethylsulfonyl fluoride, 5% glycerol). The immune complex containing PKR was incubated with activity buffer containing 1 Ci of [␥-32 P]ATP at room temperature for 30 min. One g/ml poly(I⅐C) or 1 unit/ml heparin was added as a PKR activator. In some samples, kinase activities were assayed in the presence of the affinity-purified wild-type or mutant form IV of C114 protein. Labeled proteins were analyzed by SDS-PAGE and visualized by autoradiography and Western blotting.
Partial Purification of FLAG-C114 Proteins-HEK 293 cells were transfected with plasmids expressing FLAG-tagged wild-type or the mutant form IV of C114 protein (see Fig. 2B). Cells expressing FLAGtagged fusion proteins were harvested and lysed as described above. The clear lysate was then applied to the anti-FLAG M2 (Sigma) and Protein A-agarose for overnight incubation at 4°C. The agarose beads were then washed three times with Tris-buffered saline buffer (50 mM Tris, pH 7.4, 0.15 mM NaCl). Bound FLAG-tagged fusion protein was eluted with six 1-ml aliquots of FLAG peptide (Sigma) at 0.1 mg/ml in Tris-buffered saline, washed, and concentrated by centrifugation in Centricon 10 (Amicon), and purified FLAG-tagged fusion proteins were rinsed and dissolved in kinase activity buffer.

RESULTS
Cloning of Mouse C114 -C114 was cloned using PCR-based cDNA subtraction strategy from 3T3 L1 cells treated with human IL-11 (500 ng/ml) for 30 min. Northern blot analysis confirmed induction of C114 expression in IL-11-treated 3T3 L1 cells as early as 30 min and persisting for 24 h (Fig. 1A). Sequencing of C114 revealed that it is identical to several expressed sequence tags deposited in GenBank TM . To obtain the full-length cDNA sequence for C114, PCR-based rapid amplification of cDNA ends was performed using a mouse liver cDNA library based on the overlapping sequences of several mouse expressed sequence tags. The full-length cDNA contains an open reading frame of 561 bp, which encodes a 187-amino acid polypeptide with a predicted molecular mass of 21 kDa. The deduced protein sequence contains three putative nuclear localization signals, a basic amino acid-rich region, as well as a highly acidic amino acid tail at its C terminus (Fig. 1B). An identical full-length sequence has since been submitted by others (GenBank TM accession number AK010961). C114 also shares a very high sequence homology with a recently submitted human hypothetical protein FLJ13902 (GenBank TM accession number AK023964). The striking similarity (91% identity at the amino acid level) suggests that FLJ13902 is likely a human counterpart of mouse C114. Sequence alignment with the human genome data base revealed that the human FLJ13902 gene consists of four exons and is localized at human chromosome 7q22.1. C114 protein also has significant homology with two other proteins with unknown functions from Caenorhabditis elegans and yeast (GenBank TM accession numbers AE003808 and P41880).
Northern blot analysis of mRNA prepared from mouse tissues showed that C114 mRNA is widely expressed in mouse tissues with higher abundance in liver, kidney, brain, and heart and low expression in lung and skeletal muscle (Fig. 1C). Two major C114 transcripts of 1.0 and 2.0 kb were detected in most tissues. The length of the cloned cDNA corresponds to the size of the shorter transcript.
Cellular Localization of the C114 Protein-To elucidate pos-sible functions of C114, we examined the cellular localization of the C114 protein. The C114 cDNA was fused in-frame into the N or C terminus of EGFP in the pEGFP vector, which expresses a green fluorescent protein to facilitate protein localization in living cells by confocal microscopy. Addition of C114 to pEGFP resulted in a nuclear localization of the C114 fusion protein with enrichment in the nucleolus ( Fig. 2A, panel c). In contrast, the green fluorescent signal distributed throughout the cytoplasm and nucleus of cells transfected with pEGFP plasmid (Fig. 2A, panel a), whereas the signal was detected in mitochondria when transfected with pEGFP-mitochondrial protein plasmid ( Fig. 2A, panel b).
There are three clusters of Arg/Lys-rich sequences present in FIG. 1. Molecular cloning of C114. A, time course of C114 mRNA expression in 3T3 L1 cells after IL-11 treatment. Total RNAs were prepared from 3T3 L1 cells at various times after treatment with IL-11, and 10 g of total RNA was loaded. The same filter was stripped and then hybridized with ␤-actin as a loading control. B, nucleotide and deduced amino acid sequences of C114 cDNA. The nucleotide sequence is presented in 5Ј to 3Ј orientation, and the amino acid sequence is displayed using the single-letter codes. Putative nuclear localization signals are bold, and the C-terminal acidic region is underlined. The sequence has been deposited to GenBank TM by the RIKEN Genome Exploration Group (accession number AK010359). C, Northern blot analysis of C114 mRNA expression in mouse tissues. A mouse multiple tissue poly(A) ϩ RNA blot was probed with 32 P-labeled random primed C114 cDNA. Arrows indicate the two molecular weight markers on the blot. the basic region of C114 ( 83 RHLRRR 88 , 126 KRRKKR 131 , and 133 KLKEKK 138 ). To determine the sequences responsible for nuclear localization of C114, wild-type C114 and a series of deletion mutants were expressed as GFP fusion proteins, and confocal fluorescence microscopy was performed 24 h after transfection of NIH 3T3 cells. The pattern of various EGFP-⌬C114 proteins indicated that the last two clusters of the Arg/Lys-rich sequences ( 126 KRRKKRQKLKEKK 138 ) are the major nuclear localization signals for the C114 protein (Fig. 2B).
Nucleotide Binding Analysis of the C114 Protein-Since C114 protein accumulates in nucleolus, and many nucleolar proteins are dsRNA-binding proteins (12,13), 35 S-labeled in vitro translated C114 was tested for its ability to bind various polynucleotide resins. As shown in Fig. 3, A and B, unlike in vitro translated luciferase protein, wild-type C114 bound poly(I⅐C) but not poly(A) or poly(U) RNA. Competitive binding studies were next performed to compare the binding of wildtype C114 to dsRNA versus dsDNA. For these experiments, in vitro translated C114 was incubated with poly(I⅐C)-agarose in the presence of either free poly(I⅐C) or poly(dI⅐dC) competitors and assayed for the ability to bind to the poly(I⅐C) resin (Fig.  3C). In the presence of increasing amounts of poly(I⅐C) RNA, the binding of wild-type C114 to poly(I⅐C)-agarose decreased 79.5% in the presence of the highest concentration of competitors. In contrast, poly(dI⅐dC) failed to compete for wild-type C114 binding to the poly(I⅐C) resin with the highest concentration of competitors. These results suggest that wild-type C114 preferentially binds dsRNA.
Although C114 clearly binds dsRNA, no known dsRNA-binding motifs were identified in the highly basic amino acid-rich region of C114 by domain search analysis. To determine the regions of C114 that are important for dsRNA-binding, a series of deletion constructs of C114 were assayed for their ability to bind dsRNA (Fig. 3D). The N-terminal 56 amino acids (amino acids 1-56 , mutant fragment III) and the C-terminal peptide (amino acids 143-187, mutant C114 fragment VI) showed only 15 and 25% poly(I⅐C) binding capacity as compared with wildtype C114 protein, respectively. On the other hand, the middle region (amino acids 51-143) containing highly basic amino acid residues contains 88% poly(I⅐C) binding capacity (Fig. 3D). These results suggest that the binding of C114 to dsRNA is mediated by its middle region.
C114 Interacts with the dsRNA-dependent Protein Kinase PKR-Different dsRNA-binding proteins, including the viral proteins HIV-1 transactivator of transcription (Tat), TRBP, and vaccinia virus E3 protein, and cellular proteins such as PACT and NF90, interact with PKR (13,18). We therefore tested whether C114 can interact with PKR. Co-immunoprecipitations were performed with overexpressed FLAG-tagged C114 and V5-tagged PKR constructs in HEK 293 cells. When FLAG-tagged C114 was immunoprecipitated with an antibody against FLAG, V5-tagged PKR could be detected when the immunoprecipitates were immumoblotted with anti-V5 antibody (Fig. 4A). In addition, a comparable amount of FLAG-C114 (14%) was found to associate with PKR as compared with FLAG-PACT (17%), a well characterized PKR-interacting activator (Fig. 4B). The co-immunoprecipitation results also demonstrate that the N-terminal 50 amino acids of C114 are responsible for its association with PKR in cells (Fig. 4A, lane 4,  and B, lane 6).
We also mapped the regions of PKR that interact with FLAG-tagged C114 protein. Wild-type PKR or its deletion mutants were individually expressed as [ 35 S]methionine-labeled proteins and incubated with purified FLAG-tagged C114 protein. As shown in Fig. 4C, panel b (lanes 1-3), wild-type PKR and two deletion mutants containing two DRBDs can interact with the C114 protein. Purified FLAG-C114 associated with Protein A-agarose did not pull down the PKR deletion mutant containing only the kinase domain or in vitro translated luciferase (lanes 4 and 5). These results suggest that C114 can interact with DRBDs in the N terminus of PKR.
C114 Inhibits PKR Kinase Activity-Since several RNAbinding proteins including cellular proteins such as PACT and binding to the poly(I⅐C) resin. In vitro transcribed/translated 35 S-labeled luciferase (Luc) or C114 protein (5 l of each) and proteins bound to polynucleotide resins after extensive washing (45 l of each) were analyzed by SDS-PAGE and autoradiography as indicated. B, polynucleotide binding analysis. C114 was in vitro translated in rabbit reticulocyte lysates, and the 35 S-labeled translation products were incubated with excess amounts of various polynucleotide resins. Bound proteins were eluted, and the amount of bound proteins was determined relative to the luciferase binding. C, competitive binding studies. In vitro translated wild-type C114 was incubated with various concentrations of competitor nucleic acid (either poly(I⅐C) or poly(dI⅐dC), as indicated) followed by incubation with dsRNA-agarose. Bound proteins were eluted, and the relative amount analyzed with wild-type C114 in the absence of any competitors was determined. D, regions of C114 for its poly(I⅐C) binding. In vitro translated products of various deletion mutants (as shown in Fig. 2B) were incubated with poly(I⅐C) resin. Bound proteins were eluted, and the amount of bound proteins was determined relative to the wild-type binding.
NF90 (14 -16) and viral proteins such as TRBP and E3L (19,20) are known to affect PKR activity, we tested whether C114 can modulate the activity of PKR. Expression constructs of FLAG-tagged C114 or its mutant form IV and V5-tagged PKR were co-transfected into HEK 293 cells. The V5-PKR protein was immunoprecipitated with monoclonal antibody against V5 and co-immunoprecipitated with transfected C114 proteins. When kinase assays were performed, PKR immunoprecipitated FIG. 4. C114 interacts with PKR. A, PKR co-immunoprecipitation assays with deletion mutants of C114 (as described in the legend for Fig. 2B). WT, wild type; IP, immunoprecipitation; WB, Western blot. B, relative amount of FLAG-C114 and FLAG-PACT associated with PKR. HEK 293 cells were transfected in 10-cm culture dishes with 10 g of FLAG-C114, FLAG-PACT, and/or V5-PKR construct DNAs. At 48 h after transfection, cells were harvested, and cell extracts were prepared. Total cell extracts were used to immunoprecipitate with anti-FLAG monoclonal antibody and analyzed by Western blot analysis with the anti-V5 and anti-FLAG antibodies. Ab, antibody. C, panel a, structure of wild-type and deletion mutants of PKR (18). Luc, luciferase; DRBD, dsRNA-binding domain region mutant; KD, kinase domain region mutant; 7M, link region (between DRBD and KD) deletion mutant. Panel b, C114 pull-down assays with deletion mutants of PKR. The immunoprecipitated FLAGtagged C114 protein was incubated with untagged PKR or its mutant protein synthesized by in vitro translation using T7 TNT system. Con, control.
C114, a Novel IL-11-inducible Nuclear dsRNA-binding Protein 22842 from C114-expressing cells was less activated in the presence of poly(I⅐C) or heparin (Fig. 5A, lanes 3 and 6). However, similar levels of PKR autophosphorylation were observed with immunoprecipitated V5-PKR from V5-PKR expression alone or the mutant form IV of C114 co-expressing cells (Fig. 5A, lanes 2, 4,  5, and 7). Comparable results were observed when kinase assays were performed using V5-PKR mixed with partially im-munoaffinity-purified FLAG-C114 or FLAG-tagged mutant form IV of C114 protein in vitro (as described under "Materials and Methods," data not shown). Although the IV mutant form of C114 retained 98% of the poly(I⅐C) binding ability as compared with wild-type C114 (Fig. 3D), it lost its binding ability to PKR (Fig. 4A, lane 4) and the effect on PKR activation. When PKR mixed with the IV deletion mutant of C114, similar levels  Fig. 2B) from transfected HEK 293 cells were performed in the presence or absence of poly(I⅐C) (1 g/ml) or heparin (1 unit/ml). PKR autophosphorylation was detected after proteins were separated by SDS-PAGE. Equal loading of the input proteins was determined by Western blot (WB) analysis with anti-V5. FLAG-C114 and the FLAG-tagged mutant form IV of C114 protein were detected by Western blot with anti-FLAG monoclonal antibody. B, inhibition of eIF-2␣ phosphorylation by C114. HEK 293 and HT1080 cells were transfected with plasmid expressing FLAG-tagged C114 protein 46 h after transfection, and cells were transfected with poly(I⅐C) (0.6 g/ml) for different times. C, inhibition of eIF-2␣ phosphorylation by C114 through its PKRinteracting region. HEK 293 cells were transfected with plasmids expressing FLAG-tagged C114 or the mutant IV protein 46 h after transfection, and cells were transfected with poly(I⅐C) (0.6 g/ml) for 90 min. Cell extracts were prepared at the time points indicated and subjected to Western blot analysis with monoclonal antibodies to native and phosphorylated forms of eIF-2␣. Bar charts are relative activity measurements indicating a relative quantification for PKR autophosphorylation (A) or eIF-2␣ phosphorylation (B and C).
of PKR phosphorylation could be achieved by the PKR activators in the same assay (Fig. 5A, lanes 4 and 7). These results suggest that the inhibition of PKR activity is mediated through direct interaction of C114 and PKR and is independent of dsRNA binding. This scenario is supported by the fact that wild-type C114 can also inhibit PKR activation in the presence of heparin (Fig. 5A, lane 6). To further test the negative regulatory role of C114 on PKR function in cells, we examined the levels of phosphorylation of endogenous eIF-2␣ in cells following poly(I⅐C) treatment of cells for 2 h. Overexpression of C114 in HEK 293 or HT 1080 cells resulted in lower levels of poly(I⅐C)-stimulated eIF-2␣ phosphorylation (Fig. 5B). In contrast, overexpressed C114 IV mutant did not complex with PKR (Fig. 4, A and B) and had no effect on eIF-2␣ phosphorylation (Fig. 5C, lane 6). Taken together, these results suggest that C114 inhibits cellular PKR function through the interaction between C114 and PKR protein. DISCUSSION In this study, we have identified C114 as a novel 21-kDa dsRNA-binding protein that interacts with PKR and inhibits PKR activity by a dsRNA-independent mechanism. C114 was isolated through a differential screening for IL-11-inducible genes in a mouse preadipocyte cell line, 3T3 L1. IL-11 is a multifunctional cytokine in the bone marrow microenvironment (1,21). It shares a common signal transducer, gp130, with other IL-6-type cytokines, such as IL-6, leukemia inhibitory factor, oncostatin M, ciliary neurotrophic factor, and cardiotrophin-1 (CT-1). These cytokines play important roles in the regulation of complex cellular processes such as gene activation, proliferation, and differentiation (22). The main components of their signaling pathways are Janus kinases, Jak1, Jak2, and Tyk2, and the signal transducers and activators of transcription STAT1 and STAT3 (23). However, mechanisms underlying IL-11 in growth and differentiation signaling have not been elucidated. Accumulated evidence suggest the requirement of PKR in response to a variety of biological stimuli, including cytokines (4). Recently, PKR has been shown to interact with STAT3 activated by platelet-derived growth factor to induce c-fos expression (11). However, the direct signaling link between IL-6 family cytokines and PKR function has never been established.
PKR is critical in interferon-mediated antiviral defense, apoptosis, signal transduction, and cell growth but is also implicated in the onset of differentiation (4,(7)(8). Analysis of phosphoproteins has demonstrated increased phosphorylation of PKR and eIF-2␣ during 3T3 F442A adipogenesis. Consistently, inhibition of adipocyte differentiation by cat serum correlated with reduced PKR activation, diminished eIF-2␣ phosphorylation, and elevated growth rate (7,8), suggesting a role of PKR in adipogenesis. Furthermore, introduction of a kinaseinactive dominant negative PKR mutant into 3T3 L1 prevented proliferation arrest and interfered with adipogenic differentiation (23). In addition to its mitogenic activity on a plasmacytoma cell line, IL-11 was cloned as an adipogenesis inhibitory factor (2). Our observation that C114 is an IL-11-inducible PKR modulator suggests its potential role in adipogenesis. Interestingly, besides IL-11, expression of C114 in 3T3 L1 can also be induced by insulin (data not shown). Although insulin induces a decrease in eIF-2␣ phosphorylation in chondrocytes (24), IL-11 significantly reduces LPS-induced binding activity of NF-B (25), a process known to be modulated by PKR (4,5). Further studies of PKR and C114 interaction may clarify the regulation of PKR in adipocyte differentiation and potential cross-talk between IL-11, insulin signaling pathways, and PKR functions.
PKR has a bipartite structure consisting of N-terminal DRBDs and a C-terminal catalytic domain. Its DRBDs are known to interact with its catalytic domain and act as autoregulators for PKR activation (26). PACT binds to DRBDs of PKR, which leads to a conformational change facilitating PKR activation (14). C114, like NF90 and many other cellular PKR antagonists, also binds DRBDs of PKR (13). C114 protein inhibits PKR autophosphorylation in a dsRNA-independent manner, suggesting inhibitory effects mediated by its binding to DRBDs of PKR. This interaction may result in a conformational change of PKR, leading to its inability to be activated by polyanionic agents. PKR is mainly localized in the cytoplasm but is also expressed in the nucleolus (27,28). Under normal condition in Daudi cells, the cytoplasmic to nuclear ratio is ϳ5:1. Following interferon treatment of Daudi cells for 16 h, this ratio could increase to around 17:1 (27). DRBD1 domain in the N terminus of PKR plays an important role for PKR nucleolar localization (27). Moreover, previous findings also demonstrated that cytoplasmic PKR is predominantly in a dimeric form, whereas in the nucleus it is a basic (i.e. unphosphorylated) protein. PKR appears to be actively transported to the nucleus by unknown mechanisms and translocated to the nucleolus by binding to structural RNAs, possibly rRNA (27). In contrast, C114 is predominantly nuclear but has been found in the cytoplasm by fractionation (data not shown). It is highly enriched in nucleolus, determined by its second and third Arg/Lys-rich clusters. Our data also demonstrate that C114 can interact with the DRBDs of PKR protein. Transient transfection experiments showed an increased amount of PKR in the nuclear fraction when coexpressed with C114 in the HEK 293 cells (data not shown), suggesting that C114 might play a role in nuclear import and export of PKR in cells. In many of the experimental systems, what regulates PKR in the cells has remained an enigma, especially in the cases where synthetic or viral dsRNA is not involved. As a novel cellular inhibitory factor of PKR, expression of C114 in 3T3 L1 cells can be induced by IL-11 and insulin as well as poly(I⅐C) treatment but not by interferon-␥, transforming growth factor-␤1, IL-4, and IL-9 (data not shown). C114 may modulate intracellular PKR activity under a variety of growth and differentiation conditions and in response to various extracellular stimuli, within and outside the context of virus infection. Further functional studies of this novel PKR modulator should provide important insights into the role of PKR in cell growth and differentiation.