INTRODUCTION

Top Abstract Introduction Materials and methods Results Discussion References

The initiation of signal transduction through antigen receptors in lymphocytes is associated with the rapid sequential activation of a number of tyrosine kinases, including well-studied members of the Src and the ZAP-70/Syk families (35). However, in B cells, a third nonreceptor tyrosine kinase, Btk, is also important for signaling from antigen and other lymphocyte cell surface receptors (4). Mutations in btk result in a severe B-cell immunodeficiency, X-linked agammaglobulinemia (XLA), and the murine counterpart X-linked immunodeficiency (xid), characterized by decreased mature B-cell and immunoglobulin levels. Btk kinase activity and tyrosine phosphorylation have both been shown to increase upon cross-linking or stimulation of surface immunoglobulin M and the interleukin 5 (IL-5) and IL-6 receptors in B cells, as well as cross-linking of the high-affinity immunoglobulin E receptor (FcRII) in mast cells (15, 24, 29, 30). Another Btk family kinase, Itk, which is expressed in T cells, is tyrosine phosphorylated and activated by engagement of either the T-cell receptor (TCR) or CD28 in T-cell lines and by FcRII stimulation in mast cells (1, 6, 16).

The Btk family kinases have a modular structure including the kinase and SH2 and SH3 protein interaction domains, as well as a proline-rich region and a pleckstrin homology (PH) domain. Examination of lymphocyte signaling suggests that Btk family members function downstream of Src family kinases. Src family kinases have been shown to phosphorylate Btk family members on a tyrosine in the activation loop of the kinase domain leading to kinase activation and, for Btk, autophosphorylation of a site in the SH3 domain (10, 22, 26). The activation of Btk and Itk by Src family kinases is both potentiated by and dependent on an interaction between the PH domains of the Btk kinases and the products of phosphatidylinositol (PI) 3-kinase (2, 19), demonstrating the importance of the PH domain in these signaling pathways.

rlk/TXK, originally isolated in screens for genes encoding tyrosine kinases from both human peripheral blood and murine embryonic thymus, encodes a kinase that closely resembles members of the Btk/Tec family (9, 12, 32). Rlk shares with Btk kinases the proline-rich motif and homologous SH3, SH2, and kinase domains, including the tyrosine phosphorylation sites in the SH3 and kinase sequences (Fig. 1). However, unlike the other Btk family members, Rlk lacks a PH domain and instead possesses a distinctive cysteine string motif, suggesting that Rlk has a unique role in TCR signaling. Like Itk, Rlk is specifically expressed in developing and mature T cells as well as mast cells. However, little is known about Rlk protein function in the T cell. Evidence from our and other laboratories suggests that Rlk and Itk together have important roles in signaling from the TCRtargeted mutations of either kinase cause only mild defects in T-cell numbers and decreased cytokine production (20, 21). However, mutation of both kinases leads to profound defects in TCR-based signaling (unpublished data).

View larger version (14K): [in this window] [in a new window]   FIG. 1.   Comparison of the predicted organizations of Rlk and Btk. Percent amino acid identities are indicated at the bottom of the figure.

In order to characterize Rlk/Txk biochemically, we examined properties of endogenous Rlk in T cells and Rlk-green fluorescent protein (GFP) constructs expressed in the Jurkat T-cell lymphoma cell line and in heterologous cells. We demonstrate here that Rlk consists of two protein species generated by alternative initiation of translation. These two species have distinct properties and subcellular localizations that are dictated by the presence of a cysteine string motif, which is an essential requirement for palmitoylation of the full-length protein. Coexpression of Rlk and Src family members leads to the tyrosine phosphorylation and kinase activation of Rlk. However, unlike for Btk and Itk, phosphorylation of Rlk does not appear to require PI 3-kinase activity, consistent with the lack of a PH domain. Activation of the Jurkat T-cell lymphoma line through the TCR increases tyrosine phosphorylation of transiently transfected Rlk and can lead to alterations in Rlk subcellular localization, suggesting that the nuclear localization of Rlk may play an important role in its normal function. Thus, despite lacking the PH domain, Rlk can act as a substrate for Src family kinases and may participate in novel signal transduction pathways downstream of the TCR.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and methods Results Discussion References

Cell culture and transfection. 293T cells were a generous gift from the laboratory of D. Baltimore (Massachusetts Institute of Technology). 293T and HeLa cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, penicillin, streptomycin, and 10 mM glutamine. Jurkat cells expressing simian virus 40 large T antigen (Tag) were a generous gift from Gerald Crabtree (Stanford University). Jurkat and Jurkat Tag cell lines were maintained in RPMI 1640 supplemented with 10% fetal calf serum, penicillin, streptomycin, glutamine, and 50 µM 2-mercaptoethanol.

Antibodies and reagents. Anti-murine Rlk was raised in rabbits against a glutathione S-transferase (GST) fusion protein containing the predicted amino terminus and SH3 domain of full-length murine Rlk (residues 1 to 138). DNA sequences encoding residues 1 to 138 of full-length Rlk were amplified by PCR and fused in frame to pGEX-4T (Pharmacia). GST fusion proteins were grown in bacterial strain BL21Plys5 at 37°C and induced with 100 mM IPTG (isopropyl--D-thiogalactopyranoside) at 25°C for 2 h. Proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), lightly stained with Coomassie blue, eluted from the gel into 0.1% SDS-0.1 M Tris (pH 8.0), and used to immunize rabbits (Assay Research Co., College Park, Md.). The antisera were affinity purified by incubation of 1-ml aliquots of whole serum with nitrocellulose strips containing the GST fusion protein overnight in 50 ml of phosphate-buffered saline (PBS), elution with 1 ml of 25 mM glycine (pH 3.3), and resuspension in 5 ml of Tris-HCl (pH 8.0) with 1% bovine serum albumin (fraction V; Sigma).

Constructs and mutagenesis. A fragment of mouse Rlk cDNA was subcloned into pCI (Invitrogen) by using an MscI site near the 5' end of the cDNA and an XhoI site at the 3' end. An myc epitope tag was introduced by inserting an oligonucleotide encoding the 15-amino-acid Myc tag from the BstEII site at the stop codon of Rlk to the downstream AflII site. The myc-tagged cDNA was also transferred into pcDNA3 with an NheI-to-XhoI fragment. The amino-terminal deletion mutant 54 was generated by deleting from the NheI site to the BstXI site in pCIRlk. Rlk-GFP was generated by inserting a linker containing a BstEII site from the SacI site to the NheI site of pFRED25 (a generous gift of G. Gaitanairis and G. Pavlakis, NCI-ABL, Frederick, Md.). A BstEII-to-XbaI fragment of the modified pFRED25 containing the GFP coding sequences was subcloned onto the BstEII site at the stop codon of Rlk subclone pCIR (12). pCEF was a gift of Silvio Gutkind (NIH). pCEF-Rlk was generated by subcloning Rlk-GFP into pCEF by three-fragment ligation with the EcoRI, Asp718, and NotI sites.

Transfections. Constructs were introduced into 293T cells by calcium phosphate transfection in media containing 25 µM chloroquine (Sigma) (25). Three to five micrograms of each plasmid was used for each transfection (2 × 10⁶ cells); the cells were harvested 24 to 36 h later. Alternatively, cells were split 24 h after infection and treated 24 h later with either Wortmannin (1 µm), Ly294002 (100 µM), or carrier (dimethyl sulfoxide) for 30 min before harvest. Transfections into HeLa cells were performed identically, without the addition of chloroquine.

In vitro translation. In vitro translation employed the TnT coupled reticulocyte lysate system (Promega); 1 µg of each construct was used for each reaction at 30°C for 1 h with 20 µCi of [³⁵S]methionine (1,000 Ci/mmol) (Redivue; Amersham). Translation products were separated by SDS-10% PAGE, soaked in Enlighten (NEN), dried, and analyzed by fluorography.

Western blotting and immunoprecipitation. Cells were harvested 24 h following transfection. Cells were washed once with ice-cold PBS and lysed in ice-cold Nonidet P-40 (NP-40) lysis buffer (0.5% NP-40, 50 mM HEPES [pH 7.4], 5 mM EDTA, 50 mM NaCl, 10 mM NaPO₄, 50 mM NaF, 1 mM sodium orthovanadate, 1 mM AEBSF, 2 U of aprotinin per ml) for 10 min on ice. Total protein extracts were mixed directly with 2× SDS protein sample buffer (1× SDS protein sample buffer is 50 mM Tris [pH 6.8], 2% SDS, 10% glycerol, and 0.1% bromphenol blue plus 5% -mercaptoethanol). Alternatively, lysates were clarified by centrifugation and detergent-soluble protein was analyzed. Equivalent amounts of protein were boiled in SDS sample buffer and separated by SDS-10% PAGE and transferred to nitrocellulose (Schleicher and Schuell). The membranes were blocked in TBST (10 mM Tris-HCl [pH 8.0], 150 mM NaCl, 0.1% Tween 20) with 5% (wt/vol) nonfat dry milk and incubated with the appropriate primary antibody at a 1:2,000 (anti-Rlk and 4G10 [antiphosphotyrosine]) or 1:1,000 (anti-Fyn) dilution overnight at 4°C for 1 to 2 h at room temperature. Membranes were washed and incubated with the appropriate secondary antibody at a 1:10,000 dilution (Boehringer Mannheim) for 1 h, and protein was visualized by enhanced chemiluminescence (Amersham).

In vitro kinase assays. Cell extracts were prepared in NP-40 lysis buffer and immunoprecipitated as described above. Complexes were recovered with protein A-Sepharose, washed twice with ice-cold NP-40 wash buffer, and washed twice with ice-cold Rlk kinase buffer (30 mM HEPES [pH 7.4], 150 mM NaCl, 5 mM MgCl₂, 5 mM MnCl₂, 100 µM sodium orthovanadate). The kinase reaction was carried out in kinase buffer supplemented with 1 µg of acid-denatured enolase and 10 µCi of [-³²P]ATP (Redivue; Amersham) for 5 min at room temperature. The reaction was quenched with SDS sample buffer, boiled, and separated by SDS-10% PAGE. Phosphorylated proteins were detected by autoradiography or quantified on a PhosphorImager. The level of Rlk protein was measured by Western blotting. For in vitro kinase assay of Rlk derived from murine thymocytes, 10⁷ cells were lysed in 1 ml of NP-40 lysis buffer and kinase assays were performed in a buffer containing 50 mM HEPES (pH 7.0), 150 mM NaCl, 10 mM dithiothreitol (DTT), 0.01% Brij 35, and 1 mM sodium orthovanadate. Assays were initiated by adding 50 mM MgCl₂ and 133 µCi of [-³²P]ATP (ICN), and mixtures were incubated for 20 min at room temperature. Assays were terminated by two washes with ice-cold NP-40 wash buffer plus 0.1% SDS and resuspension in 1× SDS protein sample buffer.

Subcellular fractionation of lymphocytes. Thymocytes (6 × 10⁷) from C57BL/6 mice were swollen and dounced 20 times in homogenization buffer (10 mM HEPES [pH 7.0], 10 mM KCl, 1.5 mM MgCl₂, 1 mM EDTA, 1 mM DTT, and 1 mM Na₃VO₄ plus aprotinin, leupeptin, and AEBSF). The crude nuclear pellet was obtained by spinning at 600 × g for 10 min at 4°C. A crude mitochondrial pellet was obtained from supernatant spun at 15,000 × g for 10 min at 4°C. The supernatant was spun at 100,000 × g for 60 min and was used as the cytoplasmic compartment. The pellets were resuspended in NP-40 lysis buffer at 37°C for 5 min and then placed at 4°C for 15 min and cleared at 15,000 × g for 10 min at 4°C. The S100 fraction was adjusted to 1× NP-40 lysis buffer, and immunoprecipitations and kinase assays were performed as described above.

Immunofluorescence. HeLa or 293T cells were transfected with Rlk expression constructs, passed onto coverslips 20 h later, and, 12 to 24 h later, fixed with 1% paraformaldehyde for 20 min at room temperature. Cells were washed in PBS and permeabilized with 0.1% Triton in PBS for 5 min at room temperature. Cells were then washed in PBS followed by H₂O, counterstained with DAPI (4',6-diamidino-2-phenylindole) or propidium iodide, and visualized with a charge-coupled device camera on a Zeiss Axioplan microscope. Alternatively, Jurkat Tag cells were electroporated with pCEF Rlk-GFP and 24 h later divided into equal aliquots and left unstimulated or stimulated with either OKT3 or phorbol myristate acetate (PMA) at 20 ng/ml and ionomycin at 1 µg/ml. Cells were either examined live at sequential time points after stimulation or harvested, fixed, and counterstained with propidium iodide.

Metabolic labeling with [³H]palmitate. Cells were transfected with Rlk expression constructs and incubated 36 h later in 1.5 ml of medium containing 5% dialyzed serum and 400 to 500 µCi of [³H]palmitate per 6-cm dish. One to 3 h later, cells were lysed and Rlk was immunoprecipitated as described above. Low-reducing protein sample buffer with 20 mM DTT was used in these experiments, and samples were incubated at 65°C for 3 min prior to loading on SDS-10% polyacrylamide gels. Gels were soaked in En³Hance (NEN), dried, and visualized by fluorography.

	RESULTS

Top Abstract Introduction Materials and methods Results Discussion References

rlk encodes two protein species. The rlk gene is expressed in the T-cell lineage, where its mRNA is down-regulated upon activation (12, 32). However, very little is known about the Rlk protein. To facilitate studies of Rlk, we generated antisera directed against a GST fusion protein containing the predicted amino-terminal 138 amino acids of Rlk, including the SH3 domain. We also extended the 3' end of the rlk cDNA with a 45-nucleotide sequence encoding an myc epitope for expression studies in vitro and in cell culture.

View larger version (55K): [in this window] [in a new window]   FIG. 2.   Rlk encodes two proteins. (A) In vitro protein kinase assay of Rlk protein immunoprecipitated with anti-Rlk antiserum from lysates of murine lymph node cells (lane 1), splenocytes (lane 2), or thymocytes (lane 3). Immune complexes were labeled with [-32P]ATP, resolved by SDS-PAGE, and visualized by autoradiography. (B) (Lanes 1 to 4) In vitro protein kinase assay of Rlk protein immunoprecipitated with anti-Rlk antiserum from lysates of 293T cells transfected with vector alone (lane 1) or murine myc-tagged rlk cDNA (lane 2) compared with immunoprecipitates from murine thymocytes (lane 3). Lane 4 contains immunoprecipitates with preimmune sera from thymocyte lysates. The two major protein species generated from the Rlk-myc construct migrate slightly slower than endogenous Rlk in T cells, as expected from the 15-amino-acid tag (lanes 3 and 2). Versions of the rlk cDNA lacking the myc insert generated proteins that comigrated with the endogenous forms (data not shown). (Lanes 5 to 7) Immunoblot analysis with anti-Rlk of 293T cell lysates from cells that were mock transfected (lane 5) or transfected with murine rlk cDNA (lane 6) compared to in vitro transcription-translation products (lane 7) generated from rlk cDNA. (C) Immunoblot analysis with anti-human TXK antisera of 293T cell lysates from cells that were mock transfected (lane 1) or transfected with human Txk cDNA (lane 2). The expected TXK products are indicated.

Murine and human rlk/txk genes contain internal initiation ATG codons. The production of two proteins from a single mouse or human rlk/txk cDNA in vivo and in vitro suggests that a posttranscriptional mechanism, such as protein processing or alternative translational initiation, is responsible for production of the second species. The results with in vitro translation, however, argue against proteolytic processing of the 58- to the 52-kDa form. Furthermore, pulse-chase analyses were also inconsistent with a precursor-product relationship for the 58- and 52-kDa species (data not shown). Examination of the rlk open reading frame revealed downstream in-frame ATGs that are embedded within efficient translation initiation consensus sequences in both the mouse and human cDNAs (17) and would be predicted to generate the shorter species (52 and 55 kDa in mice and humans, respectively [Fig. 3A]). Interestingly, these shorter Rlk open reading frames are predicted to lack the cysteine string motif that is unique to this tyrosine kinase.

View larger version (26K): [in this window] [in a new window]   FIG. 3.   Schematic of Rlk mutants. (A) Rlk protein organization and locations of the cysteine string motif and putative NLS in the amino terminus of Rlk. Residues changed in the mutants of the cysteine string motif and NLS are marked by asterisks. The relative positions of the putative translational initiation sites, as well as the nucleotide sequences of the various start site mutant derivatives (italicized), are shown below the schematic. SH3, Src homology 3; SH2, Src homology 2. (B) Schematic organization of the two forms of Rlk.

Mutational analysis of putative translation start sites in murine rlk. To address whether alternative initiation codons were used, we generated mutations affecting the first ATG, the second ATG, or both (Fig. 3). The resulting mutant cDNAs were assayed in vitro by using rabbit reticulocyte lysates and in transfected 293T cells.

View larger version (29K): [in this window] [in a new window]   FIG. 4.   Distinct protein isoforms of rlk are generated by the utilization of alternative translational start codons. (A) In vitro transcription-translation products generated from WT and mutant rlk cDNAs, resolved by SDS-PAGE, and visualized by fluorography. (B) Expression of rlk WT and mutant cDNAs in 293T cells. Equal amounts of protein from total cellular lysates of cells transfected with the indicated constructs were resolved by SDS-PAGE and visualized by immunoblotting with anti-Rlk serum (lower). Mutants are described in Fig. 3.

The two RLK isoforms localize to different subcellular compartments. Analyses of extracts of cells that expressed the individual forms of the Rlk proteins suggested that there were differences in the detergent solubility of the two forms. In particular, the larger (58-kDa) form was relatively insoluble in nonionic detergents, suggesting that the two forms may localize to different subcellular compartments when produced separately.

View larger version (51K): [in this window] [in a new window]   FIG. 5.   Subcellular localization of the two forms of WT and mutant Rlk-GFP fusions. HeLa cells were transiently transfected with the following constructs, fixed, and stained with DAPI. (A) WT GFP-Rlk. (B) AUG2Rlk-GFP expressing only the long form. (C) AUG1Rlk-GFP expressing only the short form. (D) NLS mutant of the short form of Rlk-GFP (NLS-AUG1Rlk-GFP). (E and F) Cysteine mutant of the long form of Rlk (CMRlk-GFP) with fluorescein channel only (E) or DAPI only (F).

The two forms colocalize when coexpressed. Despite the nuclear localization of the p77^Rlk protein when expressed in isolation, nuclear staining was only weakly observed in HeLa cells expressing both the 83- and 77-kDa proteins from the WT Rlk-GFP construct. The retention of the 77-kDa protein in the cytoplasmic compartment when both forms are coexpressed suggests that they may interact and thereby colocalize. To pursue this possibility, we differentially labeled the two forms by tagging the large form (ATG2-Rlk) at its carboxy terminus with blue fluorescent protein (BFP), a modified version of GFP. Because the fluorescent intensity of the BFP fusion protein is significantly lower than the GFP fusions, we performed these analyses with 293T cells, in which we obtain higher levels of protein expression than in HeLa cells.

View larger version (41K): [in this window] [in a new window]   FIG. 6.   The two forms of Rlk colocalize when coexpressed. 293T cells were transiently transfected with Rlk-GFP expression constructs, fixed, and counterstained with propidium iodide. (A) AUG1Rlk-GFP (short form). (B) AUG2Rlk-BFP (long form). (C and D) AUG2Rlk-BFP (long form) and AUG1Rlk-GFP (short form) cotransfected and visualized for GFP (C) or visualized for BFP (D). The colocalizations of the two forms of Rlk are indicated by arrows.

The two forms display distinct localizations in T cells. T cells express rlk mRNA throughout development. We therefore asked whether we could find the two forms of the active enzyme in similar cell compartments in freshly harvested thymocytes. We fractionated thymocytes and examined immunoprecipitated Rlk protein by kinase assay. In these cells, we normally observed two forms of the Rlk kinase (Fig. 2A), which we also found in the cytoplasmic fraction (Fig. 7, lane 2). In contrast, the nuclear fraction is greatly enriched in the short form of the Rlk protein (Fig. 7, lane 1). These results suggest that endogenous Rlk can localize to different subcellular fractions and that at least a portion of the shorter species can be found in the nucleus. Furthermore, the detection of the short form in the nucleus in the absence of the long form strongly argues that the shorter phosphorylated protein observed in lymphocytes is also Rlk.

View larger version (50K): [in this window] [in a new window]   FIG. 7.   Subcellular fractionation of murine thymocytes. Cells were fractionated as described in the text, and the nuclear and cytoplasmic fractions from 6 × 107 cells were immunoprecipitated and analyzed by kinase assay. Nucl, nuclear pellet; Cyto, S100 supernatant.

Determinants of protein localization. Inspection of the predicted amino acid sequence of Rlk revealed a cryptic bipartite nuclear localization signal (NLS) at residues 57 to 71 that is present in both species of RLK protein (Fig. 3) (9). To determine if this NLS causes nuclear localization of the shorter protein, we mutated the motif in the shorter GFP fusion construct (ATG1NLS-RlkGFP). The mutant protein failed to accumulate in the nucleus of either HeLa or 293T cells, instead being found in a granular pattern in the cytoplasm (Fig. 5D). Thus, this NLS is required for the nuclear localization of the shorter isoform of Rlk.

The cysteine motif mediates palmitoylation of Rlk. Cysteine string motifs have been found to be palmitoylated, a fatty acid modification that causes the association of proteins with certain membranous compartments (7, 8). We therefore examined whether the cysteines of Rlk could undergo fatty acid modification. 293T cells expressing WT Rlk-GFP, the cysteine mutant (CM) described above, or Fyn (an Src family member known to be palmitoylated) were metabolically labeled with [³H]palmitate, and the appropriate proteins were immunoprecipitated from cell lysates. Incorporation of [³H]palmitate, but not [³H]myristate, is observed with the WT Rlk-GFP protein (Fig. 8A, lane 3, and data not shown). However, mutation of 3 of the 6 cysteines in the cysteine string motif (CM) dramatically decreased labeling of Rlk-GFP, consistent with alteration of the site of fatty acid labeling (Fig. 8A, lane 2). Boiling in 2-mercaptoethanol also reduced labeling of the WT protein (data not shown), suggesting that the mechanism of incorporation of fatty acid is consistent with an S acetylation such as palmitoylation. Thus, the cysteine string appears to mediate palmitoylation, and this modification may account for the ability of this motif to prevent nuclear localization of Rlk.

View larger version (36K): [in this window] [in a new window]   FIG. 8.   WT Rlk is modified by palmitoylation. Mutation of the cysteine string abolishes [3H]palmitate labeling. 293T cells were transfected with cDNAs encoding Fyn (positive control, lane 1), a cysteine mutant of Rlk-GFP (lane 2), WT Rlk (lane 3), or control vector alone (lane 4). Twenty-four hours after transfection, cells were labeled with [3H]palmitate for 2 h, washed, and lysed, and Rlk or Fyn was immunoprecipitated as described in the text. (Top) Samples were analyzed by SDS-10% PAGE and visualized by fluorography. (Bottom) Immunoblot (IB) for Rlk.

Rlk is activated by Src family members. The above data suggest that Rlk, unlike the other Btk family kinases, has a distinct regulation of localization, consistent with its lack of a PH domain. Nonetheless, other features of Rlk closely resemble the Btk kinases. In particular, Rlk shares the common SH2, SH3, and kinase domains, as well as a proline-rich sequence. This proline-rich region has been shown in vitro to mediate an interaction with the SH3 domains of certain Src family kinases in vitro. Similarly, we have observed an in vitro interaction with the SH3 domains of the Src family members Fyn, Hck, and Lyn (3a).

View larger version (40K): [in this window] [in a new window]   FIG. 9.   Tyrosine phosphorylation and activation of Rlk kinase upon coexpression with Fyn in 293T cells. (A) 293T cells were transfected with Rlk-GFP fusion and FynT expression constructs, followed by cell lysis at 24 h. Equal amounts of detergent-soluble lysates were resolved by SDS-PAGE and subjected to antiphosphotyrosine (4G10) (top) or anti-Rlk (bottom) immunoblotting (IB). (B) Immune complex kinase assays. Cells expressing Rlk-GFP fusion proteins alone or with Fyn were lysed and immunoprecipitated with anti-Rlk. Immune complexes were washed and subjected to an in vitro kinase assay with [-32P]ATP and acid-denatured enolase as an exogenous substrate. Phosphorylated proteins were resolved by SDS-PAGE and detected by autoradiography. (Top) In vitro kinase activity on enolase. (Bottom) anti-Rlk. Equivalent expression of Fyn was confirmed by immunoblotting with anti-Fyn (data not shown). WT, wild type (kinase active); KI, kinase inactive.

Activation by Src family kinases is independent of PI 3-kinase activity. It has been suggested that activation of Itk and Btk by Src kinases requires a translocation to the membrane mediated by the interaction between the PH domain and the products of PI 3-kinase (2). The products of PI 3-kinase have been found to interact with PH domains of Btk and Itk, and it has recently been demonstrated that phosphorylation and activation of Itk by Src family kinases is dependent on the activity of PI 3-kinase. Either inhibitors of PI 3-kinase activity or mutation of the PH domain will decrease activation of Itk by Src family kinases (2). Furthermore, activation of BTK by Src can be potentiated by overexpression of PI 3-kinase (19).

View larger version (49K): [in this window] [in a new window]   FIG. 10.   Rlk activation by Fyn is not affected by inhibitors of PI 3-kinase. (Top) Cells were transfected with constructs expressing either Rlk alone or Rlk plus Fyn. Twenty-four hours after transfection, cells were split equally and treated 24 h later with either dimethyl sulfoxide (DMSO) (lanes 1 and 2), 1 mM Wortmannin (lane 3), or 100 mM Ly294002 (lane 4). After 30 min, cells were lysed, resolved by SDS-PAGE, and examined by Western blotting with antiphosphotyrosine (top row) or anti-Rlk (bottom row). (Bottom) Cells were transfected with constructs expressing either Itk-FLAG alone or Itk-FLAG plus Fyn and treated as above with either DMSO (lanes 5 and 6), 1 mM Wortmannin (lane 7), or 100 mM Ly294002 (lane 8). Western blotting was performed with antiphosphotyrosine (top row) or anti-FLAG (bottom row).

Rlk can be activated by TCR signaling in Jurkat cells. The ability of Src family kinases to activate Rlk suggests that Rlk may be activated downstream of Src family kinases by TCR signaling, in a fashion analogous to Btk in B-cell-receptor signaling. To examine Rlk activation in T-cell signaling, we again took advantage of our Rlk-GFP constructs and expressed them in the Jurkat T-cell lymphoma line, a strategy that allowed us to bypass technical difficulties associated with comigration of Rlk with the heavy chain of our anti-Rlk antiserum. Activation of Jurkat cells by stimulation with the OKT3 antibody directed against the CD3 component of the TCR induced rapid tyrosine phosphorylation of Rlk-GFP (Fig. 11, lanes 1 to 4). In contrast, the shorter form of Rlk was not efficiently phosphorylated in response to OKT3 (Fig. 11, lanes 5 to 7), again supporting hypothetically different roles for the two forms of Rlk. Thus, we conclude the longer form of Rlk is most likely to be the principle form involved in signaling from cell surface receptors in T cells.

View larger version (24K): [in this window] [in a new window]   FIG. 11.   Tyrosine phosphorylation of Rlk-GFP in Jurkat cells. Cells were electroporated with either WT Rlk-GFP or AUG2Rlk-GFP expressing only the short form of Rlk. Twenty-four hours later, cells were stimulated with anti-CD3 antibody OKT3 for the indicated times (in minutes). Cells were lysed and Rlk was immunoprecipitated. Immunoprecipitated Rlk was immunoblotted with antiphosphotyrosine (4G10) (top) or anti-Rlk (bottom).

Activation of Jurkat cells alters the subcellular localization of Rlk. To examine whether T-cell activation affects the subcellular distribution of Rlk, we again expressed Rlk-GFP transiently in Jurkat cells and stimulated them either through the TCR with the OKT3 antibody or with PMA and ionomycin. Full-length Rlk-GFP gave a vesicular type of localization outside the nucleus in Jurkat cells, consistent with that observed in HeLa cells (Fig. 12A). Localization of the short form of Rlk was less clear in Jurkat cells, suggesting both a nuclear and nonnuclear localization (data not shown). However, long-term activation with either PMA and ionomycin or with anti CD3 antibodies led to an apparent localization of either full-length Rlk-GFP or the short form (AUG1Rlk-GFP) in the nucleus (Fig. 12B and C and data not shown). To investigate this subcellular trafficking in further detail, we examined live Jurkat cells expressing the Rlk-GFP constructs at a range of times post-TCR stimulation. While initial stimulation did not appreciably alter the pattern of localization, stimulation for 10 to 15 min revealed two new distinct compartments. A subpopulation of Rlk was observed to migrate to the plasma cell membrane, consistent with activation by the TCR. Additionally, a second population of the protein was observed in the nucleus. Over the next 30 to 60 min the membrane fraction disappeared, while the nuclear component remained. While the function of this nuclear fraction remains unclear, these results further demonstrate a nuclear localization for Rlk in lymphocytes, which may be important for unique signaling pathways involving Rlk.

View larger version (63K): [in this window] [in a new window]   FIG. 12.   Activation of Jurkat cells changes the subcellular localization of Rlk. Cells were electroporated with WT Rlk-GFP and stimulated 24 h later with either OKT3 or PMA and ionomycin, counterstained with propidium iodide, and examined by confocal microscopy. (A) Unstimulated cells. (B) Stimulated cells, fluorescein channel. (C) Same as panel B, with rhodamine channel for visualizing propidium iodide. (D) Time course of localization of Rlk-GFP after TCR stimulation. Cells were visualized live at the indicated times (in minutes) after treatment with anti-CD3.

	DISCUSSION

Top Abstract Introduction Materials and methods Results Discussion References

By several criteria, Rlk appears to be a member of the Btk family of kinases, yet significant differences exist between Rlk and the other kinases in this family. In this report, we show that the distinct molecular features of Rlk confer unique biological properties on the kinase that may be important for its normal cellular function. We demonstrate here that rlk/txk encodes two proteins generated by the alternative initiation of translation from a single mRNA. In the case of murine Rlk, most or all of the smaller isoform is generated from an internal start site located 55 residues downstream of the upstream AUG. This smaller form contains the catalytic, SH2, and SH3 domains, as well as the proline-rich motif found in common with other BTK family kinases; however, it lacks much of the unique amino-terminal sequence, including a distinct cysteine string motif, present in the larger 58-kDa isoform. We further demonstrate that the two Rlk isoforms have distinct properties, including subcellular locations and fatty acid modification, and that the cysteine string motif helps dictate both palmitoylation and subcellular localization.

Determinants of the subcellular localization of Rlk. Alternative initiation sites can affect the subcellular location and modification of proteins. For example, the 59-kDa isoform of Hck is both palmitoylated and myristoylated and is associated with cellular membranes, most notably caveolae, whereas the 61-kDa isoform is not palmitoylated and not associated with caveolae (28). In Rlk, two motifs help determine its subcellular location. The cysteine string motif is important for perinuclear localization and can override a bipartite NLS, which is required for nuclear targeting of the short form. Hence, lacking the cysteine-rich motif, the 52-kDa form is located in the nucleus when expressed in isolation.

rlk^/ itk^/

Activation of Rlk. We demonstrate here that activation of the TCR via OKT3 can lead to increased phosphorylation of Rlk transfected into Jurkat T cells. Coexpression of Rlk with Src family kinases in heterologous cells also leads to increased phosphorylation of the full-length Rlk and activation of the Rlk kinase, suggesting that Rlk may be a downstream effector of Src family kinases in TCR signaling. However, while Rlk is phosphorylated and activated in a fashion similar to the other Btk kinases, there are marked differences between these kinases, most notably the lack of a PH domain in Rlk. A recent model suggests that membrane association of Itk and Btk results from the interaction of their PH domains with products of PI 3-kinase, thereby allowing activation by Src family kinases. Accordingly, activation of Itk by Src kinases is inhibited by Wortmannin and Ly294002 (2, 19). This is not observed for Rlk and suggests that interactions of Rlk with Src family kinases is distinct from the other Btk kinases. It is of interest that the full-length form of Rlk has a higher basal level of tyrosine phosphorylation and kinase activity than Itk when expressed in heterologous cells (Fig. 10), which could be the result of different upstream interactions and regulation of localization. Moreover, we can reproducibly coimmunoprecipitate Rlk and Fyn, whereas we and others have been unable to coprecipitate Itk and Btk with members of the Src family (data not shown), implying that there is a direct interaction between Rlk and the Src kinases.