Role of Tyrosine Phosphorylation in the Regulation of the Interaction of Heterogenous Nuclear Ribonucleoprotein K Protein with Its Protein and RNA Partners*

The heterogeneous nuclear ribonucleoprotein K protein recruits a diversity of molecular partners and may act as a docking platform involved in such processes as transcription, RNA processing, and translation. We show that K protein is tyrosine-phosphorylated in vitro by Src and Lck. Treatment with H 2 O 2 /Na 3 VO 4 , which induces oxidative stress, stimulated tyrosine phosphorylation of K protein in cultured cells and in intact livers. Tyrosine phosphorylation increased binding of Lck and the proto-oncoprotein Vav to K protein in vitro . Oxidative stress increased the association of K protein with Lck and Vav, suggesting that tyrosine phosphorylation regulates the ability of K protein to recruit these effectors in vivo . Translation-based assay showed that K protein is constitutively bound to many mRNAs in vivo. Native immunoprecipitated K protein-mRNA complexes were disrupted by tyrosine phosphorylation, suggesting that the in vivo binding of K protein to mRNA may be responsive to the extracellular signals that activate tyrosine kinases. This study shows that tyrosine phosphorylation TBST incubated TBST (25 following concen-trations of secondary used: alkaline phosphatase-con- jugated anti-rabbit Ig (Promega) 1:3000 anti-mouse Ig 1:3000 (Promega); horseradish peroxidase conjugated anti-mouse Ig Biotech) 1:2000 anti-rabbit 1:5000 (Amer-sham Biotech). For colorimetric detection using alkaline phosphatase, the membranes were developed using the phosphatase phosphate/nitro blue tetrazolium substrate solution For chemiluminescence, using horseradish peroxidase, membranes were developed using the ECL kit as per the manufacturer’s protocol (Am-ersham Pharmacia Biotech). K Protein Plasmid Constructs— The point mutations of Ser 302 to Ala or Glu and of Tyr 236 , Tyr 230 , and Tyr 234 to Phe were generated using the QuickChange Site-Directed Mutagenesis Kit (Stratagene). The triple point mutant, where Tyr 236 , Tyr 230 , and Tyr 234 were all mutated to Phe, generated two-stage polymerase chain reaction protocol

with 20% nonfat dry milk in TBST buffer for 30 min. Membranes that were used in anti-phosphotyrosine blots were blocked in TBST containing 2% bovine serum albumin. After blocking, membranes were washed three times with TBST for 5 min and were then incubated for 1 h with TBST containing the primary antibody (25°C). Rabbit anti-K protein (antibody 54) was used at 1:10,000 dilution, mouse anti-phosphotyrosine (PY99, Santa Cruz) at 1:1000 dilution, and mouse monoclonal anti-FLAG (M3, Sigma) at 1:200 dilution. Membranes were then washed three times with 10 ml of TBST and then incubated with secondary antibody in TBST for 30 min (25°C). The following concentrations of secondary antibodies were used: alkaline phosphatase-conjugated anti-rabbit Ig (Promega) at 1:3000 and anti-mouse Ig at 1:3000 (Promega); horseradish peroxidase conjugated anti-mouse Ig (Amersham Pharmacia Biotech) at 1:2000 and anti-rabbit at 1:5000 (Amersham Pharmacia Biotech). For colorimetric detection using alkaline phosphatase, the membranes were developed using the phosphatase 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium substrate solution (Kirkegaard & Perry Laboratories, Gaithersburg, MD). For chemiluminescence, using horseradish peroxidase, membranes were developed using the ECL kit as per the manufacturer's protocol (Amersham Pharmacia Biotech).
K Protein Plasmid Constructs-The point mutations of Ser 302 to Ala or Glu and of Tyr 236 , Tyr 230 , and Tyr 234 to Phe were generated using the QuickChange Site-Directed Mutagenesis Kit (Stratagene). The triple point mutant, where Tyr 236 , Tyr 230 , and Tyr 234 were all mutated to Phe, was generated by a two-stage polymerase chain reaction protocol that allows the simultaneous introduction of multiple mutations (29). All plasmids were purified by CsCl gradient before use in transient transfections, and the mutations were confirmed by automated sequencing using the DyeDeoxy Terminator Cycle Sequencing Kit (Applied Biosystems, Foster City, CA).
In Vitro Transcription and Translation-In vitro transcription and translation was performed in the presence of [ 35 S]methionine using the TNT T7 Quick Coupled Transcription/Translation System as per the manufacturer's protocol (Promega, Madison, WI).
Translation Assay to Define K Protein-mRNA Interactions-RNA bound to the beads bearing K protein was eluted with 100 l of RNA elution buffer. RNA was phenol-purified and ethanol-precipitated. RNA pellets were dissolved in 10 l of water and were translated in vitro using Flexi Rabbit Reticulocyte Lysate System as per the manufacturer's protocol (Promega) in the presence of [ 35 S]methionine. 35 S-Labeled translational products were resolved by SDS-PAGE and were analyzed by autoradiography. In K protein-mRNA interaction assay extractions, binding and phosphorylation reactions were carried out in the presence of 1 unit/l RNase inhibitor.

RESULTS
K Protein Is Tyrosine-phosphorylated in Vitro-Analysis of K protein amino acid sequence reveals that Tyr 230 , Tyr 234 , Tyr 236 , and Tyr 380 meet the criteria (CANSITE) for phosphorylation by the Src family of tyrosine kinases. To test if K protein is tyrosine-phosphorylated, increasing amounts of purified GST-K fusion protein was incubated with the tyrosine kinase, Lck (30), in Lck phosphorylation buffer containing [␥-32 P]ATP (20 min, 30°C). Phosphorylation of increasing amounts of enolase, an established tyrosine kinase substrate (31), was done in parallel. Phosphorylation of GST protein was also done as a control. After phosphorylation, proteins were resolved by SDS-PAGE, gels were stained (Fig. 1C) and autoradiographed (Fig.  1B), and radioactivity of the relevant bands was measured in a scintillation counter (Fig. 1A). These results show that fulllength GST-K protein was as good a substrate for phosphorylation by Lck as was enolase (Fig. 1A, compare enolase, bars 1-5 with GST-K, bars 6 -10). These results also show that Lck was autophosphorylated ( Fig. 1, B, lane 11, and A, bar 11) to a level lower than either enolase or GST-K phosphorylation. Because GST was not phosphorylated by Lck (Fig. 1, B and C, lane 11,  (54) and GST-K fusion protein were phosphorylated by 1 l of baculovirusexpressed Lck in Lck phosphorylation buffer containing [␥ -32 P]ATP for 20 min at 30°C. Reaction of GST with Lck was done as a control. Reactions were stopped by boiling the samples in SDS-loading buffer. Samples were resolved by SDS-PAGE, and gels were stained with fast stain (C) and autoradiographed (B). Bands corresponding to enolase (lanes 1-5), full-length GST-K protein (lanes 6 -10), GST (lane 11) and Lck (lane 11) were cut out from the gel (C), and radioactivity was measured. Incorporation of phosphate (pmol/min) is shown in A as follows. Bars 1-5, enolase; bars 6 -10, GST-K; bar 11, Lck; bar 12, GST. D, glutathione beads (40 l) bearing either wild-type K protein (GST-K) or one of the K protein point mutants (GST-KY236F, GST-KY234F, and GST-KY230F) were suspended in 50 l of reaction buffer containing either 1 l of baculovirus-expressed Src (upper panel) or Lck (lower panel) kinase. The phosphorylation reaction was carried out for 20 min (30°C), beads were washed twice with binding buffer, and proteins were eluted by boiling in loading buffer. Eluted proteins were resolved by SDS-PAGE, gels were Coomassie-stained, and 32 P-labeled GST-K protein was visualized by autoradiography. and A, bar 12), Lck-mediated phosphorylation of GST-K fusion protein reflects phosphorylation of K protein itself. GST-K was also effectively phosphorylated by the tyrosine kinase Src (32) (Fig. 1D).
To begin to identify sites that are phosphorylated by the tyrosine kinase, beads bearing either wild-type or mutated GST-K fusion protein were phosphorylated (20 min, 30°C) in vitro by either Src or Lck. Mutation of either Tyr 230 or Tyr 236 to Phe (GST-KY230F and GST-KY236F) greatly reduced the level of GST-K protein phosphorylation by Src. Mutation of Tyr 234 (GST-KY234F) also reduced the level of phosphorylation, but not as much as the mutations of the other two sites (Fig. 1D, upper panel, compare lanes 2-4 with lane 1). These results suggest that all three sites are phosphorylated by this enzyme in vitro. Compared with the wild-type GST-K protein (Fig. 1D, lower panel, lane 1), the level of Lck-mediated phosphorylation of GST-KY236F and GST-KY234F (Fig. 1D, lower panel, lanes 2 and 3) was greatly reduced, but not the level of phosphorylation of GST-KY230F (Fig. 1D, lower panel, lane 4). These experiments suggest that Tyr 236 and Tyr 234 are the preferable sites for phosphorylation by Lck in vitro. These results show that although both Src and Lck effectively phosphorylate K protein in vitro, their phosphorylation site preferences overlap but are not identical.
K Protein Is Tyrosine-phosphorylated in Vivo-Treatment of cells with H 2 O 2 induces oxidative stress and stimulates tyrosine phosphorylation of many intracellular proteins (33,34). This effect is thought to result from both protein tyrosine kinase activation and the inhibition of protein-tyrosine phos-phatases. We used this agent, in combination with the Na 3 VO 4 phosphatase inhibitor, to test if K protein is inducibly tyrosinephosphorylated in vivo. Human Jurkat thymoma cells (24) were treated with H 2 O 2 /Na 3 VO 4 , and cells were harvested after different time intervals. Extracts were prepared, and K protein was isolated by immunoprecipitation using a highly specific anti-K protein antibody (number 54) (22). Immunoprecipitated proteins were analyzed by Western blotting using monoclonal anti-phosphotyrosine ( Fig. 2A, lanes 1-4) and polyclonal anti-K protein ( Fig. 2A, lanes 5-8) antibodies. The antiphosphotyrosine blot revealed a major 65-kDa band whose intensity was higher in extracts from cells treated with H 2 O 2 / Na 3 VO 4 ; the peak effect occurred at 30 min and returned to base line after 60 min of treatment. This inducible 65-kDa band ( Fig. 2A, lanes 1-4) overlaps with the K protein band seen on the anti-K protein blot ( Fig. 2A, lanes 5-8), indicating that the transiently tyrosine-phosphorylated 65-kDa band is probably K protein. The anti-phosphotyrosine blot also shows that there are several other well defined and inducibly tyrosine-phosphorylated proteins that co-immunoprecipitate with K protein.
These bands may represent proteins that inducibly bind to K protein.
To ensure that the 65-kDa inducibly tyrosine-phosphorylated band is K protein, we transfected HeLa cells with either wild-type FLAG tag K fusion protein, FLAG-K, or a mutant, FLAG-KY3F, where the predominant three sites, Tyr 230 , Tyr 234 , and Tyr 236 , of in vitro phosphorylation by Src/Lck (Fig.  1), were all mutated to Phe. Twenty-four hours after transfection, cells were treated for 5 min with either media or H 2 O 2 / After incubation, 20 l of protein A/G-agarose beads were added to each sample, and the suspension was mixed for 2 h (4°C). After centrifugation, beads were washed four times with 1 ml of immunoprecipitation buffer, and proteins were eluted by boiling in SDS-loading buffer. Immunoprecipitated (IP) proteins were resolved on SDS-PAGE; and after electrotransfer and Western blotting, colorimetric immunostaining (IS) was carried out with either an anti-phosphotyrosine (lanes 1-4, ␣P-Tyr) or anti-K protein (lanes 5-8, ␣K) antibody. B, HeLa cells were transiently transfected (2.5 g of each plasmid/60-mm plate) with a FLAG tag mammalian expression plasmid containing wild-type K protein (lanes 1 and 2 and lanes 5 and 6), FLAG-K, or K protein where Tyr 230 , Tyr 234 , and Tyr 236 residues were all mutated to Phe (lanes 3 and 4 and lanes 7 and 8), FLAG-KY3F. After an overnight transfection, cells were treated with or without 3.5 mM H 2 O 2 , 0.1 mM Na 3 VO 4 for 5 min, and cell extracts were prepared and proteins were immunoprecipitated with 1 l of anti-FLAG monoclonal antibody (1 h) and protein A/G beads (30 min) as in A. Immunoprecipitated proteins were analyzed with chemiluminescence Western blotting using either anti-phosphotyrosine (lanes 1-4, ␣P-Tyr) or anti-K protein (lanes 5-8, ␣K) antibody. C, portal veins of a group of three ether-anesthetized mice (20 g each) were injected with 500 ml of either saline (Ϫ) (lanes 1-3 and 7-9) or 2 mM H 2 O 2 0, 0.5 mM Na 3 VO 4 (lanes 4 -6 and 10 -12) (ϩ). 30 min following the injection, livers were procured and frozen at Ϫ70°C. 200 -300 mg of frozen tissues were used to prepare extracts as described previously (36). K protein was immunoprecipitated with anti-K protein antibody and protein A/G beads. K protein immunoprecipitates were analyzed by SDS-PAGE, Western blotting, and chemiluminescence immunostaining using either anti-phosphotyrosine (␣P-Tyr) (lanes 1-6) or anti-K protein (lanes 7-12) antibody. Na 3 VO 4 . After treatment, cells were harvested, cell extracts were prepared, and FLAG-K protein was immunoprecipitated with monoclonal anti-FLAG antibody. The same samples of precipitated proteins were analyzed by Western blotting using both anti-phosphotyrosine (Fig. 2B, lanes 1-4) and anti-K protein (Fig. 2B, lanes 5-8) antibody. The anti-K protein blot revealed one 65-70-kDa band. The size of this protein is identical to a strongly inducible band seen in the anti-phosphotyrosine blot from cells transfected with the wild-type FLAG-K. Compared with the wild-type K protein, a much weaker level of tyrosine phosphorylation induction of the FLAG-KY3F mutant was seen (compare lanes 1 and 2 with lanes 3 and 4).
These results confirm that K protein is tyrosine-phosphorylated in vivo in response to H 2 O 2 /Na 3 VO 4 treatment. Mutation of three Tyr sites, Tyr 230 , Tyr 234 , and Tyr 236 , to Phe was required to reproducibly decrease the H 2 O 2 /Na 3 VO 4 inducibility of FLAG-K protein phosphorylation, suggesting that all three tyrosine residues are phosphorylated in vivo in response to this treatment at the 5-min time point. Specific mapping is required to be certain which of the three sites are phosphorylated in vivo, but precise mapping may be difficult to do because these sites are clustered so closely together. Nonetheless, even with the three point mutations, there was constitutive and small inducible tyrosine phosphorylation of the FLAG-KY3F mutant, indicating that there are sites in addition to Tyr 230 , Tyr 234 , and Tyr 236 that are phosphorylated in vivo.
The slightly higher constitutive level of tyrosine phosphorylation of the FLAG-KY3F compared with wild-type FLAG-K (compare lanes 1 and 3) may, in part, be accounted for by the higher level of FLAG-KY3F expression (compare lanes 5 and 6). It is also possible that the FLAG-KY3F mutant structure is sufficiently altered to allow slightly increased constitutive tyrosine phosphorylation on another site(s).
The above experiments show that K protein is tyrosinephosphorylated in response to oxidative stress in cells grown in culture. Oxidative stress plays an important role in the pathogenesis of diseases of many organs as well as in aging (35).
Next we wished to test if oxidative stress alters tyrosine phosphorylation of K proteins in an intact organ. Portal vein injections into livers of anesthetized mice is a useful model to study the effects of growth factors and other agents in an intact organ. Portal veins were infused with either saline or H 2 O 2 / Na 3 VO 4 in six anesthetized mice, three animals in each treatment group (36). After 30 min, livers were harvested, and liver extracts were prepared (36). Immunoprecipitation of K protein (Fig. 2C, lanes 1-6) revealed that there is low constitutive tyrosine phosphorylation of K protein in the liver (Fig. 2C,  lanes 1-3), and the level of tyrosine phosphorylation is increased following oxidative stress induced by intrahepatic administration of H 2 O 2 /Na 3 VO 4 (Fig. 2C, compare lanes 4 -6 with lanes 1-3). The results in the mouse liver suggest that inducible tyrosine phosphorylation of K protein transmits a signal in intact organs in response to systemic changes. This observation is consistent with a recent report that K protein may participate in hepatic response to sepsis (4).
Tyrosine Phosphorylation of K Protein Stimulates Its Association with Lck-K protein interacts with a number of factors, including tyrosine kinases, which it can complex in vivo (14). The results from the above experiments suggest that tyrosine phosphorylation may stimulate the association of K protein with some of its partners that are also tyrosine-phosphorylated in response to changes in the extracellular environment (Fig.  2). Jurkat cells express high levels of the tyrosine kinase Lck (37) that may, in part, be responsible for the tyrosine phosphorylation of K protein seen in Jurkat cells ( Fig. 2A). We next tested if K protein associates with Lck and whether this inter-action is altered by tyrosine phosphorylation of K protein. Beads bearing GST-K protein were phosphorylated (2 h, 30°C) with or without baculovirus-expressed Lck and after extensive washes that removed Lck (see Amido Black staining in Fig. 7C, lane 8), beads were mixed with 35 S-Lck and then washed again. Proteins were eluted from beads by boiling in SDS-loading buffer, separated by SDS-PAGE, transferred to PVDF membrane, and analyzed by autoradiography and Amido Black staining (Fig. 3A). These results showed that 35 S-Lck binding to GST-K was greatly increased after tyrosine phosphorylation of K protein (Fig. 3A, Autoradiograph, compare lanes 2 and 3). Densitometry measurements revealed that 58% of 35 S-Lck used in the reaction bound to GST-K. Following tyrosine phosphorylation, the electrophoretic mobility of GST-K protein was slower, and the band became broader (Fig. 3A, Amido black, compare lanes 2 and 3). The unphosphorylated and the tyrosine-phosphorylated bands do not overlap, indicating high stoichiometry of Lck-mediated phosphorylation. This shift was not seen in Fig. 1, because those reactions were carried out for 20 min compared with the 2-h reaction carried out here (Fig.  3A).
To test if the in vivo interaction of K protein with Lck is also modulated by tyrosine phosphorylation, Jurkat cells were treated for 30 min with or without H 2 O 2 /Na 3 VO 4 , cell extracts were prepared, and Lck was immunoprecipitated with a monoclonal antibody (Fig. 3B, lanes 1 and 2). Monoclonal antibody against FLAG tag was used as a control (Fig. 3B, lanes 3 and 4). Immunoprecipitated proteins were analyzed by immunostaining with anti-K protein antibody. These results showed that the amount of K protein co-immunoprecipitated with Lck increased after treatment of cells with H 2 O 2 /Na 3 VO 4 (compare lanes 1 and 2) . These results suggest that the K protein-Lck association in vivo is modulated in response to changes in the extracellular environment and that the enhanced interaction may, in part, reflect an increased level of tyrosine phosphorylation of K protein ( Fig. 2A). We estimate by densitometry that approximately 3% of total K protein was immunoprecipitated with Lck from the extracts of H 2 O 2 /Na 3 VO 4 -treated Jurkat cells. The low apparent stoichiometry of the K protein-Lck complex formation is not unexpected, since K protein interacts with many protein partners. Nonetheless, this calculation is likely to be a significant underestimate of the true stoichiometry of the in vivo K protein-Lck interaction in the K protein microenvironment. First, the immunoprecipitation and the washes were carried out under high stringency conditions, where a total of 1.5% of detergent was used. Second, this calculation was performed under the assumption that the anti-Lck antibody immunoprecipitates all of the K protein-Lck complexes. Third, in the subcellular compartments, where both proteins are potentially co-localized, the stoichiometry of the interaction between K protein and Lck is likely to be significantly higher than the above estimate. This is supported by the efficient binding of Lck to K protein in vitro (Fig. 3A). The estimate of the K protein-Lck in vivo binding is similar to the apparent in vivo binding stoichiometry for interactions involving other scaffold or anchoring proteins. For example, similar stoichiometry of the protein phosphatase 2B A kinase anchor protein, AKAP79, has been reported in brain lysates (38) or the well described in vivo interaction between the endothelial nitric-oxide synthase with the scaffold protein, caveolin-1. As in the case of the in vivo K protein-Lck interaction, the co-immunoprecipitation of endothelial nitric-oxide synthase with caveolin-1 is stimulated by tyrosine phosphorylation of endothelial nitric-oxide synthase (39).
Mutation of Tyr 230 , Tyr 234 , and Tyr 236 to Phe, FLAG-Y3F mutant, decreased the level of tyrosine phosphorylation of K protein in response to treatment of cells with H 2 O 2 /Na 3 VO 4 (Fig. 2B). We used the FLAG-Y3F mutant to test if tyrosine phosphorylation of K protein regulates its interaction with Lck in vivo. The observation that Lck precipitated wild-type FLAG-K but not the FLAG-Y3F mutant (Fig. 2C) from extracts of H 2 O 2 /Na 3 VO 4 -treated Jurkat cells (Fig. 3C, compare lanes 1 and 2) supports the postulate that increased tyrosine phosphorylation is responsible for the enhanced binding of Lck to K protein in vivo and that Tyr 230 , Tyr 234 , and Tyr 236 play a role. The constitutive K protein interaction with the Src family of kinases is thought to be mediated by SH3 interactions (19). The enhanced binding (Fig. 3) that follows tyrosine phosphorylation (Fig. 2) may reflect the addition of SH2 interactions. Although Lck is primarily localized to the cytosolic side of the plasma membrane (40), K protein is equally distributed between the nuclear and cytosolic compartments (9). Since K protein is involved in mRNA translation (2,41), the interaction of Lck with K protein may regulate this and other cytosolic nucleic acid-dependent processes. In contrast, Src and other tyrosine kinases have a wide intracellular distribution and can be localized to the nucleus (42), where they could interact with nuclear K protein.
Tyrosine Phosphorylation of K Protein Stimulates Its Association with Vav-The proto-oncoprotein Vav (p95 vav ) is a key element of signaling through both B and T cell antigen receptors (43,44). Vav contains both SH2 and SH3 domains and is tyrosine-phosphorylated upon activation of cells by several cytokines and growth factors (45). The in vivo phosphorylation of Vav by tyrosine kinases may occur in the context of microenvironments of scaffold or docking proteins (14,46). K protein may play such a function because it can recruit both tyrosine kinases (Fig. 3) (14,19) and Vav (14,15). The constitutive binding of K protein and Vav is mediated by SH3 interactions (14,15). We tested if the K protein interaction with Vav in vitro is modulated by tyrosine phosphorylation. These results showed that 35 S-Vav binding to unphosphorylated GST-K beads was very weak, but the binding increased after K protein was phosphorylated by Lck (Fig. 4A, Autoradiograph, compare lanes 3 and 4). There was no 35 S-Vav binding to beads bearing GST (Fig. 4A, compare lanes 1 and 2), indicating that the increased binding is mediated by tyrosine-phosphorylated K protein and not by any residual Lck that might have bound to the GST beads. A similar increase in the binding of 35 S-Vav was also seen after GST-K protein was phosphorylated in vitro by Src (data not shown). Densitometry measurements revealed that 14 and 28% of 35 S-Vav used in the reaction bound to GST-K after Lck- (Fig. 4A, lane 4) or Src-mediated phosphorylation, respectively. The increased binding is probably the result of tyrosine phosphorylation of K protein and subsequent enhancement of SH2 interaction. It is also possible that the increased binding may, in addition, reflect binding of Vav to residual Lck that remained bound to K protein.
To test if K protein-Vav interaction is modulated by tyrosine phosphorylation in vivo, Jurkat cells were treated for 30 min with or without H 2 O 2 /Na 3 VO 4 , cell extracts were prepared, and co-immunoprecipitation experiments were carried out with an anti-Vav polyclonal antibody. Immunoprecipitated proteins were analyzed by immunostaining with anti-K protein antibody. Nonimmune rabbit serum was used as a control. These Western blots showed (Fig. 4B) that in untreated cells there was a very low level of co-immunoprecipitation of K protein with Vav, but the amount of co-immunoprecipitation greatly increased after treatment of cells with H 2 O 2 /Na 3 VO 4 (compare lanes 1 and 2) . These results suggest that like in vitro (Fig. 4A), the in vivo interaction of K protein-Vav is increased, at least in part, by K protein tyrosine phosphorylation, an effect that may reflect both SH3 and SH2 interactions. We estimate by densitometry that approximately 1% of total K protein in cell lysates from H 2 O 2 /Na 3 VO 4 -treated Jurkat cells complexed with Vav.  PKC␦ and other PKC isoforms (21). We next tested if tyrosine phosphorylation of K protein alters the ability of K protein to be phosphorylated on Ser 302 by PKC␦. Glutathione beads bearing either wild-type or mutated GST-K protein were first phosphorylated (2 h, 30°C) with or without Lck as described above (Fig.  3). After several washes to remove residual Lck, beads were resuspended in PKC phosphorylation buffer containing 200 g/ml of the tyrosine kinase inhibitor herbimycin A (47), and phosphorylation by baculovirus-expressed PKC␦ was carried out in the presence of [␥-32 P]ATP. PKC␦ phosphorylated GST-K protein, and the level of PKC␦-mediated phosphorylation was greatly augmented by prior tyrosine phosphorylation of K protein by Lck (Fig. 5A, compare lanes 1 and 2). The higher level of phosphorylation was not the result of any residual Lck that might have remained bound to the beads because (i) without PKC␦ in the second reaction, the 32 P signal was very low or undetectable (lanes 3, 6, 10, 13, and 15), (ii) no 32 P signal was detected at all in the GST-KS302A and GST-KS302E mutants (lanes 4 -10), and (iii) there was phosphorylation in the presence of herbimycin A (lanes 1 and 2 and lanes 11 and 12), which effectively blocked Lck-mediated GST-K protein phosphorylation (compare lanes 17 and 18).  12 with lanes 14 and 15), suggesting that phosphorylation of Tyr 236 is required for the increased ability of PKC␦ to phosphorylate Ser 302 in vitro. This suggests that in vivo phosphorylation on Tyr 236 may prime K protein to be phosphorylated on Ser 302 . These results also suggest that in vivo phosphorylation of K protein in response to changes in the extracellular environment may reflect a chain of phosphorylation reactions. In the case of oxidative stress, the priming may occur on tyrosine residues.
Next we investigated if the enhanced ability of PKC␦ to phosphorylate K protein is a result of the increased binding of these proteins to each other. 35 S-PKC␦ binding to unphosphorylated and tyrosine-phosphorylated GST-K protein was assessed in pull-down assays as described above (Fig. 3A). The autoradiograph of these gels (Fig. 5B) shows that phosphorylation of GST-K by Lck enhanced the binding of 35 S-PKC␦ to GST-K protein (compare lanes 1 and 2). Mn 2ϩ is required for Lck-mediated phosphorylation in vitro (data not shown). When Mn 2ϩ was omitted in the Lck-containing phosphorylation buffer, the enhanced binding was not observed (compare lanes 2 and 5), providing further evidence that the enhanced binding results from tyrosine phosphorylation of GST-K protein. PKC␦ can be tyrosine-phosphorylated by the Src family of kinases (48). Since the reticulocyte lysates that were used to synthesize 35 S-PKC␦ contain ATP, it is formally possible that the enhanced binding is the result, at least in part, of tyrosine phosphorylation of 35 S-PKC␦ by residual Lck that may have remained bound to the beads. To prevent any tyrosine phosphorylation of 35 S-PKC␦, we carried out the pull-down assay in the presence of the tyrosine kinase inhibitor, herbimycin A. These results show that herbimycin A did not alter the enhanced binding of 35 S-PKC␦ to tyrosine-phosphorylated GST-K (compare lanes 2 and 3), providing further evidence that this effect is the result of Lck-mediated tyrosine phosphorylation of K protein. The tyrosine phosphorylation-enhanced binding of PKC␦ to K protein (Fig. 5B, lanes 1-6) (21) is similar to the increased binding in vitro of Lck ( Fig. 2A) and of Vav (Fig.  3A) to tyrosine phosphorylated GST-K protein. However, the K protein tyrosine phosphorylation-mediated increase in binding for these factors does not apply for other partners, since the binding of TATA box-binding protein to K protein was not altered by tyrosine phosphorylation of K protein (Fig. 5B, lanes  6 -10).
Mutation of Tyr 236 to Phe abrogated the tyrosine phosphorylation-mediated increase in the ability of PKC␦ to phosphorylate GST-K protein on Ser 302 (Fig. 5A, lanes 11-16). To determine if the decreased ability of PKC␦ to phosphorylate the Tyr 236 to Phe mutant might be the result of decreased binding, we carried out pull-down assays of 35 S-PKC␦ using beads bearing either the wild-type GST-K or GST-KY236F fusion proteins after the beads were phosphorylated with or without Lck. Without Lck treatment, small amounts of 35 S-PKC␦ were pulled down with either wild-type or mutant fusion protein (Fig. 5C, lanes 2 and 3), but after Lck treatment there was a large increase in 35 S-PKC␦ that was brought down by the wild-type GST-K (lane 4), but a much smaller increase in binding was seen with the GST-KY236F mutant beads (lane 5). These results suggest that the increase in PKC␦-mediated phosphorylation of K protein that follows tyrosine phosphorylation may (Fig. 5A, lanes 1-2), in part, reflect the increase in K protein-PKC␦ association (Fig. 5C, lanes 1 and 4). Consistent with this model is the observation that mutation of Tyr 236 to Phe lowers both the level of PKC␦-mediated phosphorylation (Fig. 5A, compare lanes 12 and 15) and the level of 35 S-PKC␦-GST-K protein binding (Fig. 5C, compare lanes 4 and 5). These experiments provide evidence that phosphorylation of Tyr 236 may play a role in the regulation of K protein interaction with some of its partners in vivo.
Evidence that Ser 302 Regulates Tyrosine Phosphorylation of

FIG. 4. Tyrosine phosphorylation of K protein regulates its interaction with Vav in vitro and in vivo.
A, GST and GST-K proteins bound to glutathione beads were phosphorylated with or without 1 l of baculovirus-expressed Lck as before (Fig. 3). After phosphorylation, beads were washed three times with 1 ml of washing buffer and were resuspended in 100 l of HKMT buffer containing 2.5 l of 35 S-Vav in reticulocyte lysates. The suspension was mixed for 40 min (4°C), beads were washed three times with 1 ml of HKMT buffer, and proteins were eluted by boiling in SDS-loading buffer. Eluted proteins were resolved by SDS-PAGE, transferred to PVDF membrane, and visualized by autoradiography (Autoradiograph) and protein staining (Amido Black). 50% of load is shown in lane 5. B, Jurkat cells were treated and harvested as in Fig. 3. Half of each lysate was immunoprecipitated (IP) with 40 l of anti-Vav polyclonal antibody (␣Vav), and the other half was immunoprecipitated with 40 l of preimmune serum (pre). After sonication, 20 l of protein A/G beads were added to each sample. Western blotting was carried out with anti-K protein antibody (␣K) using chemiluminescence as in Fig. 3.
K Protein in Vivo-Activation of cells by a variety of ligands stimulates a cross-talk between serine/threonine and tyrosine kinases. For example, agents such as phorbol 12-myristate 13-acetate, that directly activate PKC, can indirectly activate protein tyrosine kinases (49). The cross-talk between these major classes of kinases could occur in the context of a microenvironment of docking proteins (1). The previous set of experiments provides evidence that tyrosine phosphorylation of K protein regulates the ability of PKC␦ to phosphorylate Ser 302 (Fig. 5A). Mutation of Ser 302 to either Ala or Glu did not detectably alter the ability of Lck to tyrosine-phosphorylate GST-K protein in vitro when the phosphorylation reaction was carried out for 2 h (30°C) (Fig. 5A, Coomassie; note the nearly complete electrophoretic shifts for GST-K (lane 2), GST-KS302A (lane 5), and GST-K302E (lane 8)). Although in vitro, under these conditions, Ser 302 did not appear to alter tyrosine phosphorylation of GST-K protein by a single enzyme, Ser 302 may play a role in the regulation of tyrosine phosphorylation of K protein in vivo, where many kinases are likely to target K protein. To test this possibility, Jurkat cells were transiently transfected with either wild-type FLAG-K fusion protein or FLAG-K fusion protein mutated from Ser 302 to Ala, FLAG-KS302A, or Glu, FLAG-KS302E. Twenty-four hours after transfections, cells were treated for different times with H 2 O 2 / Na 3 VO 4 . Cell extracts were prepared, and FLAG-K proteins were immunoprecipitated with anti-FLAG monoclonal anti-body. Immunoprecipitated proteins were analyzed by SDS-PAGE and Western blotting using anti-phosphotyrosine antibody (Fig. 6). After 5 min of treatment, there was a very low but equal increase in the level of tyrosine phosphorylation of all three FLAG-K fusion proteins (lanes 4 -6). For all three fusion proteins, there was further increase at the 10-min time point, but the level of tyrosine phosphorylation was clearly higher for the wild-type compared with either one of the Ser 302 mutants (compare lane 7 to lanes 8 and 9). After 20 min, there was a continued increase in the tyrosine phosphorylation level of the wild-type FLAG-K and an even greater increase in the level of tyrosine phosphorylation of the FLAG-KS302E mutant (compare lanes 10 and 11), but there was no further increase in the level of tyrosine phosphorylation of the FLAG-KS302A mutant (lane 12). At the 20-min time point in three separate transfection experiments of this type, twice the level of tyrosine phosphorylation of FLAG-KS302E was greater than that of FLAG-K and once the two signals were equal. In all three experiments at this time point, the signal of FLAG-KS302A was much lower than the signal with either FLAG-K or FLAG-KS302E. Since mutation of Ser to Glu can mimic phosphorylation of Ser (50), these results suggest that phosphorylation of Ser 302 , regulates in vivo tyrosine phosphorylation of K protein.
The identity of the 95-kDa band that co-immunoprecipitates with K protein and whose kinetics of tyrosine phosphorylation resembles that of K protein is not known. The size of the band

FIG. 5. Lck-mediated phosphorylation of K protein stimulates the ability of PKC␦ to bind and phosphorylate K protein in vitro.
A, glutathione beads bearing either wild-type GST-K (lanes 1-3, lanes 11-13, and lanes 17 and 18) or GST-KS302A (lanes 4 -6), GST-KS302E (lanes 7-10) and GST-KY236F (lanes 14 -16) were suspended in reaction buffer without (Ϫ) or with (ϩ) 1.5 l of baculovirus-expressed Lck. After phosphorylation for 2 h (30°C), beads were washed three times with 1.0 ml of 20 mM Hepes, pH 7.6, buffer and were resuspended in 50 l of 20 mM Hepes, pH 7.6, containing 200 g/ml herbimycin A. After incubation for 30 min (30°C), beads were washed with 1 ml of PKC␦ buffer, and phosphorylation by PKC␦ was carried out for 15 min (30°C) in 50 l of PKC␦ buffer containing 2 l of baculovirus-expressed PKC␦ and [␥-32 P]ATP. Beads were washed twice with 1.0 ml of binding buffer, and eluted proteins were resolved by SDS-PAGE. Gels stained with Coomassie (lower panel, Coomassie) were then autoradiographed (upper panels). Lanes 17 and 18 show GST-K eluted from beads that were phosphorylated by Lck in the absence of PKC␦ with (ϩ) or without (Ϫ) preincubation with herbimycin A as described for lanes 1-16. B, beads bearing GST-K protein suspended in Lck phosphorylation buffer were incubated without (ϪLck) or with (ϩLck) Lck for 2 h (30°C) in the absence (Ϫ) or presence (ϩ) of 10 mM MnCl 2 . Beads were washed four times with 1.0 ml of HKMT buffer and were then mixed for 30 min (4°C) with 100 l of HKMT buffer containing 2.0 l of reticulocyte lysate translational products of either 35 S-PKC␦ or 35 S-labeled TATA box-binding protein. After the binding reaction, beads were washed four times with HKMT buffer, bound proteins were eluted by boiling beads in loading buffer, and eluted proteins were analyzed by SDS-PAGE and autoradiography. C, beads bearing GST-K or GST-KY236F mutant were incubated without (ϪLck) or with (ϩLck) Lck for 2 h (30°C) as before. Beads were washed four times with HKMT buffer and were mixed with 100 l of HKMT buffer, containing 2 l of 35 S-PKC␦ translational product. After washing four times with 1.0 ml of HKMT buffer, bound proteins were analyzed by SDS-PAGE and autoradiography. matches that of Vav (45), a tyrosine-phosphorylated factor that binds K protein (Fig. 4) (14,15,51). p95 vav expression is expressed solely in hematopoietic cells (45). The fact that this tyrosine-phosphorylated band was present in Jurkat cells ( Figs. 2A and 6) but not in the liver (Fig. 2D) or HeLa cells (Fig.  2B) supports the possibility that this 95-kDa band is Vav.
Tyrosine Phosphorylation of K Protein Alters Its Binding to RNA-K protein binds selective RNA sequences via the three KH domains (52), an interaction that might be regulated by K protein phosphorylation (13). At sites of nucleic acid-dependent processes, inducible phosphorylation of K protein may regulate not only K protein-protein but also K protein-nucleic acid interaction. To determine if K protein-nucleic acid interactions are regulated by phosphorylation, we carried out a series of experiments to determine if tyrosine phosphorylation of K protein alters its ability to bind RNA (Fig. 7). In vitro, K protein strongly binds poly(C) RNA (22). Unphosphorylated and tyrosine-phosphorylated GST-K was incubated with beads bearing poly(C) RNA (Fig. 7A). Fig. 7A shows that unphosphorylated, but not Lck-phosphorylated, GST-K protein bound to the poly(C) RNA beads (compare lanes 3 and 4), suggesting that tyrosine phosphorylation of K protein inhibits its binding to poly(C) RNA in vitro. To test this further, we incubated 32 Plabeled poly(C) RNA with beads bearing unphosphorylated and tyrosine-phosphorylated GST-K protein (Fig. 7B). Fig. 7B shows that more 32 P-labeled poly(C) RNA bound to GST-K protein beads that were not treated with Lck (LckϪ) than to the beads that were phosphorylated with Lck (Lckϩ). Together these experiments indicate that tyrosine phosphorylation of K protein by Lck decreases its binding to poly(C) RNA in vitro.
To test if the tyrosine phosphorylation-mediated decrease in K protein poly(C) RNA binding is specific to this RNA or whether it is a more general effect, we compared binding of mRNAs to either unphosphorylated or tyrosine-phosphorylated GST-K protein (Fig. 7C). To assess K protein-mRNA interactions, we developed a facile cell-free translation-based assay (see "Materials and Methods"). Beads bearing GST-K protein were phosphorylated with or without Lck and then washed to remove Lck as before. The washes effectively removed the enzyme, since no 60-kDa band corresponding to Lck was detected by Amido Black staining (Fig. 7C, compare lanes 7 and  8). Total RNA prepared from Jurkat cells (RNA input, lane 2) was mixed with beads bearing unphosphorylated (Fig. 7C, lane  7) or Lck-phosphorylated (Fig. 7C, lane 8) GST-K protein.
Beads were spun, supernatant was saved (unbound, lanes 3  and 4), and, after washing, RNA bound to beads (bound, lanes 5 and 6) was eluted with RNA elution buffer. Eluted RNAs were phenol-deproteinized and translated in a cell-free system, and 35 S-labeled products were analyzed by SDS-PAGE and autoradiography. These results show that unphosphorylated, but not tyrosine-phosphorylated, GST-K protein beads (compare unbound, lanes 3 and 4, and bound, lanes 5 and 6) bound a large fraction of most of the mRNAs. These results show that Lck-mediated tyrosine phosphorylation of K protein dramatically decreased K protein-mRNA binding in vitro.
The above series of experiments suggests that, in vivo, K protein may interact with a large repertoire of RNAs, and these interactions may be modulated by K protein tyrosine phosphorylation (Fig. 7C). Anti-K protein antibody (number 54) precipitates 15-lipoxygenase mRNA from cell extracts with K protein (7). We used this antibody and the translation-based assay to determine if the interaction of K protein with native mRNAs is regulated by tyrosine phosphorylation. K protein was immunoprecipitated with anti-K antibody (number 54), and beads were extensively washed and then phosphorylated (10 min, 30°C) with or without baculovirus-expressed GST-Itk-KD, a catalytic domain of the Itk tyrosine kinase (27). We used the catalytic domain rather than the full-length enzyme to minimize binding of the kinase to K protein and therefore reduce the probability that binding of the kinase, rather than phosphorylation, induces changes in K protein-mRNA interaction. We used GST-Itk-KD because GST-Lck-KD was not active, while GST-Itk-KD was as effective as Lck in phosphorylating GST-K protein in vitro (Fig. 7D, lanes 1-4). After washing the beads, RNAs were eluted from the beads and were used in the translation-based assay using the [ 35 S]methionine label as before. The autoradiograph shown in Fig. 7D revealed a spectrum of translational products that were generated from mRNAs eluted from immunoprecipitated K protein that was not phosphorylated (lane 2), but only a few lower level bands were translated from mRNA bound to K protein that had been phosphorylated with GST-Itk-KD (lane 3). This indicates that K protein immunoprecipitates many mRNAs and that most of the mRNAs dissociate from K protein during Itk-mediated K protein phosphorylation. These results are consistent with the notion that increased tyrosine phosphorylation of K protein in response to extracellular signals might regulate K protein-mRNA interaction in vivo.

DISCUSSION
In this study, we show that K protein is tyrosine-phosphorylated by Src and Lck in vitro (Fig. 1), in cell culture (Fig. 2, A  and B), and in mouse liver (Fig. 2C) in response to oxidative stress. In addition to the multiple tyrosine-phosphorylated sites, K protein is constitutively and inducibly phosphorylated on multiple serine/threonine residues in vivo (23). For example, Ser 302 is located in the middle of the KI domain and is phosphorylated by PKC␦ and by other isoforms of this class of enzymes (21). The presence of multiple sites that are inducibly phosphorylated suggests that K protein phosphorylation in response to extracellular signals reflects a cascade of phosphorylation events in which different sites are sequentially phosphorylated by the same and/or different enzymes. In the case of oxidative stress (Fig. 2), phosphorylation of Tyr 230 , Tyr 234 , Tyr 236 , or a sequential combination of these tyrosines could be FIG. 6. Mutation of Ser 302 alters tyrosine phosphorylation of K protein in vivo. Jurkat cells were transiently transfected with FLAG tag mammalian expression plasmid containing either wild-type K protein, FLAG-K, or K protein with Ser 302 mutated to either Ala (FLAG-KS302A) or Glu (FLAG-KS302E). After overnight transfection with SuperFect, cells were treated with 3.5 mM H 2 O 2 , 0.1 mM Na 3 VO 4 . At given time points, cells were lysed with immunoprecipitation buffer; immunoprecipitation was carried out with 1 l anti-FLAG antibody, and Western blotting was carried out with anti-phosphotyrosine (␣P-Tyr) and chemiluminescence. the priming phosphorylation event(s). The priming role of Tyr 230 , Tyr 234 , or Tyr 236 is supported by the observation that in vivo phosphorylation of these tyrosine residues seems to occur rapidly (Fig. 2B) and by the fact that tyrosine phosphorylation of K protein greatly enhances the subsequent phosphorylation of Ser 302 by PKC␦ in vitro (Fig. 5A). In this in vitro situation, phosphorylation of Tyr 236 plays a key role (Fig. 5A, lanes 11-16). Results depicted in Fig. 6 show that the Ser 302 to Glu mutation, which mimics phosphoserine, stimulates tyrosine phosphorylation at a later time point following treatment. This suggests that phosphorylation of Ser 302 may regulate the subsequent phosphorylation of tyrosine sites. Although mapping of phosphorylation sites over a time course remains to be done, results presented here (Figs. 5 and 6) support the notion that phosphorylation of K protein in response to changes in the extracellular environment reflect a succession of phosphorylation events.
K is an abundant ubiquitously expressed protein that is constitutively bound to a repertoire of mRNAs (Fig. 7D) and probably other nucleic acids including single-and doublestranded DNA (53). K protein is inducibly phosphorylated in response to growth factors, acute phase reactants, and oxidative stress (14,21,23). Phosphorylation of K protein regulates its interaction with Lck, Vav, and PKC␦ (Figs. 3-5) (21) and nucleic acids (Fig. 7) (13). These observations suggest that one of the functions of K protein is to link signal transduction pathways at sites of nucleic acid-dependent processes such as transcription, translation, and RNA processing. There are several plausible mechanisms by which K protein could bridge signals to these distal nucleic acid-dependent processes. In one scenario, K protein might simply constitutively block access of effectors to, for example, RNA and/or DNA elements, and, upon inducible phosphorylation, K protein detachment from nucleic acids would allow such processes as transcription and translation to proceed. Consistent with this model is the observation that K protein is a known silencer of 15-lipoxygenase (41) and human pappillomavirus type 16 L2 mRNA translation (6). Another plausible and a simple scenario is one where following extracellular signal-triggered phosphorylation, kinases, and/or other factors, that by themselves can not bind nucleic acids, are recruited to RNA/DNA by K protein, where they can target effectors that regulate nucleic acid-dependent processes.
It is also plausible that the bridging action of K protein is more complex than what these two simple models describe. K protein can exist as an oligomer and is known to interact with many protein factors through several discrete domains (1). Thus, K protein has the potential to recruit many proteins and could serve as a docking platform that would concentrate ef- FIG. 7. Decreased binding of RNA to tyrosine-phosphorylated K protein. A, beads bearing GST-K protein were phosphorylated with (ϩ) (input, lane 2) or without (Ϫ) (input, lane 1) Lck. GST-K protein was purified from glutathione beads and then was incubated with beads bearing poly(C) RNA. After washing the beads, proteins were eluted from the poly(C) RNA beads, resolved by SDS-PAGE, and analyzed by anti-K protein blots (bound, lanes 3 and 4). B, GST-K bound to beads was phosphorylated with (ϩ) or without (Ϫ) Lck. After washing, the beads were mixed with 32 P-labeled poly(C) RNA. After several more washes, 32 P counts bound to the beads were measured (cpm). C, total RNA prepared from Jurkat cells (25 g) was incubated with beads bearing either unphosphorylated (GST-K, lane 7) or Lck tyrosine-phosphorylated (pTyr-GST-K, lane 8) GST-K fusion proteins. Beads were centrifuged, and the supernatant containing unbound RNA was saved. Beads were then washed, and bound RNA was eluted with RNA elution buffer. Eluted and unbound RNA was phenol-deproteinized and ethanol-precipitated. Total RNA (input, lane 2), RNA that did not bind to the beads (unbound, lanes [3][4], and RNA eluted from the beads (bound, lanes 5-6) was translated in reticulocyte cell-free system in the presence of [ 35 S]methionine (Flexi Rabbit Reticulocyte Lysate System; Promega). Translational products were analyzed by SDS-PAGE and autoradiography. Translation was done without added RNA (lane 1). D, beads bearing GST-K protein were phosphorylated for 2 h with (ϩ) (lanes 2 and 4) or without (Ϫ) (lanes 1 and 3) 1 ml of baculovirus GST-Itk-KD, the catalytic domain of the tyrosine kinases, Itk (27). Beads were boiled in SDS-loading buffer, and proteins were separated by SDS-PAGE and transferred to PVDF membrane. The membrane was first immunostained colorimetrically with anti-phosphotyrosine antibody (␣-pTyr, lanes 1 and 2), and then the same membrane was stained with Amido Black (lanes 3 and 4). Jurkat cells (100 ϫ 10 6 ) were lysed in 0.5 ml of immunoprecipitation (IP) buffer containing RNase inhibitor (400 units) as well as phosphatase and protease inhibitors for 30 min on ice. After centrifugation (20 min, 4°C, 14,000 ϫ g), supernatant was added to 200 l of protein A/G beads with 5 l of preimmune serum (antibody 54). After mixing (1 h, 4°C) and centrifugation, supernatants were added to 20 l of protein A/G beads and 2 l of anti-K protein serum (antibody 54). After mixing (1 h, 4°C), beads were divided into two aliquots and were washed four times with 1 ml of immunoprecipitation buffer and once with Lck phosphorylation buffer. The phosphorylation reaction was carried out in 50 l of reaction buffer without (Ϫ) (lane 5) or with (ϩ) (lane 6) 1 l of GST-Itk-KD for 10 min at 30°C in the presence of 40 units of an RNase inhibitor. After centrifugation, RNA bound to K protein was extracted from beads using 100 l of RNA elution buffer. RNA extracted from the beads was phenol-purified and ethanol-precipitated. RNA pellets were dissolved in 10 ml of water and were translated in vitro using a rabbit reticulocyte cell-free system, and translational products were analyzed by SDS-PAGE and autoradiography as in C. fectors in its microenvironment, thus facilitating molecular cross-talk. For example, in response to extracellular signals, the Src family of kinases phosphorylate and then bind to K protein (Figs. 1-3). Tyrosine phosphorylation of K protein then leads to increased binding of PKC␦ to K protein (21). By recruitment of serine/threonine and tyrosine kinases, K protein would provide an opportunity for cross-talk between the two classes of kinases, leading to their interdependent activation and sequential phosphorylation of additional K protein sites. Tyrosine phosphorylation of K protein would concurrently release the K protein-kinase complex from RNA/DNA (Fig. 7). Activated Src and/or PKC␦ would then be in position to target effectors of nucleic acid-dependent processes that are present in the microenvironment of K protein. The target effectors of the nucleic acid-dependent processes could include transcription, translation, and/or other factors.
K protein interaction with several of its protein partners is blocked by cognate nucleic acids (17,21). For example, the native K protein-PKC␦ complex can be dissociated by poly(C) RNA (21). Based on these types of observations, where binding of protein partners and phosphorylation of K protein is altered by cognate RNA/DNA, we have previously suggested that K protein is a nucleic acid-regulated docking platform (1). This notion is conceptually consistent with the postulated model of K protein action. The inducible phosphorylation initiated by extracellular signals is transient ( Fig. 2A), suggesting that the constitutively bound K protein dissociates from nucleic acids, as the level of tyrosine phosphorylation increases, and then reassociates with RNA/DNA following dephosphorylation. The newly bound nucleic acid could halt the cross-talk and cause the release of kinases and other factors docked on K protein, thus shutting down the chain of events that were initiated by K protein tyrosine phosphorylation and bringing K protein back to the constitutive state. This is but one hypothetical model for K protein action, and many of the proposed steps remain to be tested.
In summary, we show that K protein is tyrosine-phosphorylated in vivo and in vitro, a modification that alters K protein interaction with protein and nucleic acid partners. The results are consistent with a model where K protein, acting as docking platform, anchor, or a scaffold, bridges signal transduction pathways to sites of active transcription, translation, and/or other nucleic acid-dependent processes. Regardless of which, if any, of the postulated models of K protein action is correct, the interaction of K protein with nucleic acids and protein partners is likely to be complex and tightly regulated by a chain of phosphorylation and dephosphorylation events on multiple serine/threonine and tyrosine residues.