CVAK104 Is a Novel Poly-l-lysine-stimulated Kinase That Targets the β2-Subunit of AP2*

Isolated clathrin adaptor protein (AP) preparations are known to co-fractionate with endogenous kinase activities, including poly-l-lysine-stimulated kinases that target various constituents of the clathrin coat. We have identified CVAK104 (a coated vesicle-associated kinase of 104 kDa) using a mass spectroscopic analysis of adaptor protein preparations. CVAK104 is a novel serine/threonine kinase that belongs to the SCY1-like family of protein kinases, previously thought to be catalytically inactive. We found that CVAK104 co-fractionates with adaptor protein preparations extracted from clathrin-coated vesicles and directly binds to both clathrin and the plasma membrane adaptor, AP2. CVAK104 binds ATP, and kinase assays indicate that it functions in vitro as a poly-l-lysine-stimulated kinase that is capable of autophosphorylation and phosphorylating the β2-adaptin subunit of AP2.

Isolated clathrin adaptor protein (AP) preparations are known to co-fractionate with endogenous kinase activities, including poly-L-lysine-stimulated kinases that target various constituents of the clathrin coat. We have identified CVAK104 (a coated vesicle-associated kinase of 104 kDa) using a mass spectroscopic analysis of adaptor protein preparations. CVAK104 is a novel serine/ threonine kinase that belongs to the SCY1-like family of protein kinases, previously thought to be catalytically inactive. We found that CVAK104 co-fractionates with adaptor protein preparations extracted from clathrincoated vesicles and directly binds to both clathrin and the plasma membrane adaptor, AP2. CVAK104 binds ATP, and kinase assays indicate that it functions in vitro as a poly-L-lysine-stimulated kinase that is capable of autophosphorylation and phosphorylating the ␤2-adaptin subunit of AP2.
Endocytosis involves the invagination of specialized regions of the plasma membrane, which pinch off to form cargo-containing vesicles that are transported into the cell. Although several diverse endocytic pathways are present in eukaryotic cells, the clathrin-mediated pathway is ubiquitous and the most efficient mechanism. This internalization pathway is key to a variety of biological processes that range from the downregulation of activated signaling receptors to synaptic vesicle recycling (1).
Clathrin-mediated endocytosis is a dynamic process that is tightly regulated both spatially and temporally and involves the coordinated effort of a host of cytosolic and membraneassociated proteins (2,3). Central to endocytic vesicle formation are two protein complexes, clathrin and AP2, 1 which make up the major coat constituents. The formation of progressively curved clathrin lattices on the cytoplasmic face of the plasma membrane is believed to be a driving force in the generation of coated pits and coated vesicles. However, clathrin, being incapable of directly binding to the plasma membrane or to cargo, requires the action of AP2. AP2 is a multifunctional complex consisting of two large peptide chains (␣ and ␤2), a medium (2), and a small chain (2) that coordinates the recruitment and assembly of clathrin and links it to cargo that is marked for internalization. AP2 also acts as a platform to recruit other functionally relevant endocytic factors, such as amphiphysin, eps15, epsin, AP180, AAK1, etc. (4 -6).
Although the mechanisms involved in regulating the clathrin-mediated endocytic pathway are not fully understood, accumulating evidence suggests that phosphorylation cycles may be a key step (7)(8)(9). Indeed, phosphorylation of the clathrin heavy chain (10) and the large and medium subunits of AP2 (11)(12)(13) have been demonstrated to be important to the clathrin-mediated internalization of growth factors and nutrients, respectively. However, the phosphorylation of endocytic factors is not limited to the major coat constituents. Many other components of the endocytic machinery (e.g. dynamin 1, amphiphysin 1 and 2, synaptojanin, AP180, epsin, and eps15), collectively known as the dephosphins, are phosphorylated in resting neurons and then coordinately dephosphorylated following nerve terminal stimulation (14).
Although it is clear that phosphorylation plays an important role in clathrin-dependent internalization, a more detailed understanding of the endocytic regulatory mechanisms will require the identification of the respective kinases that target endocytic components. Multiple kinases have been defined that target the major constituents of clathrin-coated vesicles. Although Src kinase (10) and casein kinase II (15) phosphorylate the clathrin heavy and light chains, respectively, and two distantly related kinases, AAK1 (5) and cyclin G-associated kinase/auxilin 2 (16,17) phosphorylate the 2-subunit of AP2, the ␣and ␤2-adaptin kinases remain to be determined.
We have taken a directed proteomic approach to identify novel kinases associated with adaptor protein preparations. Here we report the identification and initial characterization of a novel serine/threonine kinase that we have named CVAK104 (coated vesicle-associated kinase of 104 kDa).
CVAK104 Antibody and Protein Production-For antibody production, the KIAA1360 cDNA insert (encoding amino acids 145-933 of CVAK104) was subcloned into the EcoRI site of pGEX-4T3 and transformed into BL21(DE3) cells. Expression of the fusion protein was induced by the addition of 0.1 mM isopropyl 1-thio-␤-D-galactopyranoside to the cell cultures for 3 h at 25°C. Bacterial cells were lysed by nitrogen cavitation, cellular debris was removed by centrifugation, and the soluble fraction containing the expressed fusion protein was isolated using glutathione-agarose beads. The bound protein was eluted from beads using 20 mM reduced glutathione in TBST (50 mM Tris, pH 7.5, 150 mM NaCl, 0.05% Tween 20, and 0.5% bovine serum albumin), which was subsequently used as an immunogen for antibody production in rabbit (number 6369) using established protocols (21). Antibody specificity was tested by immunoblot analysis with preimmune sera showing no immunoreactivity to bovine brain coat constituents (data not shown). The CVAK104-GST fusion construct in pVL1393-GST (see "CVAK104 Cloning") was used for insect cell-mediated protein expression using the BaculoGold expression system following the manufacturer's protocols (Pharmingen). After 48 h of infection with recombinant baculovirus and expression in TN5 cells, the cells were pelleted, processed, and fusion protein was isolated following the identical procedures as those followed for bacterially expressed GST fusion proteins, with one exception. Following GST fusion protein binding to glutathione beads, bound protein was washed three times with 20 bead volumes of PBS containing 0.1% Tween 20 (PBST) containing 0.5 M NaCl before elution with the reduced glutathione.
Nucleotide Binding Assays-To assay nucleotide binding, 2 g of isolated CVAK-or AAK1-GST fusion protein was incubated in kinase buffer (see "Kinase Assays") and 50 M AB11 (Affinity Labeling Technologies, Inc., Lexington, KY), a biotinylated, UV light, cross-linkable ATP analogue, in the absence or presence of unlabeled ATP at various concentrations. Following AB11 addition, samples were incubated for 5 min at room temperature and then exposed to UV light at 254 nm using a handheld UV lamp for 2.5 min. The binding assay was terminated by the addition of protein sample buffer and boiling. Samples were resolved by SDS-PAGE, transferred to nitrocellulose, probed for bound biotin using alkaline-phosphatase conjugated to avidin, and developed as described previously (22).
Kinase Assays-Kinase assays were performed essentially as described previously (5), with minor modification. Briefly, isolated baculovirus-expressed CVAK104-GST was incubated in kinase buffer (150 mM KCl, 5 mM MgCl 2 , 100 M [␥ 32 P]ATP) in the presence or absence of 50 g/ml poly-L-lysine (30 -70 kDa, Sigma) with isolated adaptor proteins for 45 min at room temperature. The kinase reaction was stopped by the addition of protein sample buffer and boiled for 2 min at 100°C. The proteins were resolved on a 10% polyacrylamide gel (Bio-Rad), transferred to nitrocellulose, exposed to a phosphorimaging plate, and analyzed using the ImageQuant Software package (Amersham Biosciences). The endogenous kinase activities of adaptor and other protein preparations were inactivated by pre-incubation with 1 mM FSBA (fluorosulphonylbenzoyladenosine, Sigma), for 1 h on ice. Unbound FSBA was removed from protein preparations by gel filtration using a G25 mini-spin column (Amersham Biosciences) before use in the kinase assays.
Adaptor Protein and Clathrin Interaction Tests-Isolated CVAK104-GST fusion protein was incubated with glutathione beads (Amersham Biosciences) for 1 h at room temperature. The beads were washed three times with 20 bead volumes of PBST. CVAK104-GST-beads (10 l at ϳ0.5 g/l) were incubated with either 5 g of isolated AP1, AP2, or clathrin triskelia in 250 l of PBST for 1 h at room temperature. The beads were washed three times with 20 bead volumes of PBST and then analyzed for adaptor protein and clathrin binding by SDS-PAGE and immunoblot analysis.
Electrophoresis and Immunoblot Analysis-SDS-PAGE and immunoblot analysis were performed using established protocols (22). Two-dimensional gel analysis of adaptor proteins incorporating urea/SDS followed by SDS-PAGE was performed as described previously (23). For immunoblots, CVAK104 polyclonal antibody was diluted 1:2,000. Preimmune sera and secondary antibody controls showed no immunoreactivity to bovine brain samples (data not shown).

RESULTS
CVAK104 Isolation and Cloning-Isolated adaptor protein preparations co-fractionate with endogenous kinase activities that target various subunits of the AP2 complex. We previously identified AAK1, a 2 kinase present in these preparations, using a phage display protein interaction screen (5). To identify other endogenous kinase activities, isolated adaptor protein preparations derived from bovine brain (19) were analyzed using mass spectroscopy. This analysis revealed the presence of AAK1, as expected, and led to the identification of a previously uncharacterized kinase encoded by human cDNA KIAA1360 (GenBank TM accession number AB037781). Sequence analysis indicated that KIAA1360 lacks an initiating methionine, whereas domain analysis using SMART (24) suggested the presence of a kinase domain.
To identify the full-length cDNA, the 5Ј nucleotide sequence of KIAA1360 was used to search the human expressed sequence tag data base (NCBI). This resulted in the identification of the initiating methionine and full-length cDNA (other fulllength sequences have now been identified, see accession number NP_060458) that encodes a kinase we now call CVAK104. CVAK104 has a predicted molecular mass of ϳ104 kDa and belongs to a class of phosphotransferases, whose specificity for serine/threonine or tyrosine is ambiguous (Fig. 1). Interestingly, although CVAK104 has other kinase signature motifs, it lacks the highly conserved aspartic acid residue used for nucleotide catalysis in other kinases. Thus, CVAK104 belongs to the SCY1-like family of protein kinases (referred to as SCYL2) (25) that are thought to be catalytically inactive. CVAK104 contains two weak clathrin binding motifs, DLL (26) and an NPF motif shown to support interaction with EPS15 homology domain-containing proteins, such as eps15 and intersectin (27).
Phylogenic analysis indicates that CVAK104 is highly conserved in mammals, sharing Ͼ93% identity with the full-length amino acid sequences of rat (GenBank TM accession number XM_2350507) and mouse (accession number NM_198021) and 98% identity within their predicted kinase domains (not shown).
CVAK104 Co-fractionates with Adaptor Protein Preparations-To analyze the distribution of CVAK104, we raised polyclonal antibodies against the C-terminal two-thirds of the protein (see "Experimental Procedures"). In crude bovine brain lysates, CVAK antisera specifically recognized three protein species (104-, 116-, and 118-kDa) ( Fig. 2A). Of these, the 104-kDa species had a size consistent with the predicted molecular weight of the full-length cDNA. This species was predominantly membrane-associated and co-fractionated with isolated clathrin-coated vesicles ( Fig. 2A). The larger two CVAK species are found in both cytosolic and membrane fractions, but neither is enriched along with clathrin-coated vesicles. Whether they represent alternate splice forms of CVAK104, as we have observed for AAK1 (5), or another antigenically related protein remains to be determined.
Further fractionation of Tris-extracted clathrin coat proteins, resolved by gel filtration and hydroxyapatite chromatography (19,20), revealed that CVAK104 elutes from the hydroxyapatite column at low salt in fractions enriched with AP1, as detected by immunoblot analysis using an anti-␥-adaptin antibody (Fig. 2B). However, we do not believe that this cofractionation reflects a direct interaction between AP1 and CVAK104 (see next paragraph). Together, these results confirm our mass spectroscopy data and those of others (28), indi-cating the presence of CVAK104 in AP preparations, and suggest that CVAK104 may directly interact with clathrin coat constituents to regulate their function.
CVAK104 Specifically Interacts with Endocytic Vesicle Coat Components-We postulated that, because CVAK104 co-fractionates with clathrin-coated vesicles, CVAK104 might also directly interact with at least one of the major coat proteins. To explore this possibility, baculovirus-expressed CVAK104-GST fusion protein was immobilized on glutathione beads and incubated with the major coat proteins, AP1, AP2, and clathrin. CVAK104-GST specifically bound clathrin triskelia and isolated AP2 complexes (Fig. 3), whereas no interaction was observed following incubation of these coat protein complexes with immobilized GST. In contrast, we were unable to detect interaction between CVAK104 and AP1 under our experimental conditions (Fig. 3), although minor contaminating amounts of AP2 in our AP1 preparations were detected and found to interact with CVAK104. These data would suggest that CVAK104 might be preferentially associated with clathrincoated vesicles derived from the plasma membrane rather than from the trans-Golgi network or endosomes. However, we were unable to test this prediction, because our antibodies were not effective reagents for immunofluorescence localization studies.
The C-terminal ear domains of ␣and ␤2-adaptin are regions of the AP2 complex known to support interaction with a variety of endocytic accessory factors. However, binding studies (5) indicate that the AP2-CVAK104 interaction is not supported through the ear domain of ␣-adaptin; interaction between the isolated ␣-adaptin ear domain and CVAK104-GST fusion protein was not detectable under conditions known to support binding to other factors, such as AAK1 (data not shown). Attempts to directly test a potential interaction between the ear domain of ␤2-adaptin and CVAK104 proved inconclusive, as both proteins were found to nonspecifically bind to agarose beads (data not shown). These results suggest that CVAK104 binds specifically to AP2 at a site distinct from the appendage domain and by a means distinct from other AP2-interacting proteins. This is consistent with the lack of AP2-interacting motifs in its sequence. These AP2-CVAK104 interactions are likely disrupted by the 0.5 M Tris used to extract coated vesicles, and the co-elution of CVAK104 with AP1 during hydroxyapatite chromatography most likely reflects its interactions with hydroxyapatite medium rather than direct interactions with AP1.
CVAK104 Is a Serine/Threonine Kinase That Phosphorylates the ␤2-Adaptin Subunit of AP2-We have shown that CVAK104 encodes a putative kinase that is present in AP FIG. 1. The predicted amino acid sequence of human coated vesicle-associated kinase, CVAK104. The N-terminal serine/threonine kinase domain of CVAK104 is shaded, whereas a predicted coiled-coil (C-C) domain is underlined (amino acids 664 -699). The arrowhead indicates the position of the asparagine in CVAK104 that, in other kinases, is an aspartic acid used for nucleotide catalysis. CVAK104 also contains two conserved protein interaction motifs (boxed) that are thought to support binding to clathrin (DLL) and EH domain-containing proteins (NPF). Amino acids 145-933 were used for the production of polyclonal antibodies (see "Experimental Procedures").
preparations, which contain endogenous kinase activities. Therefore, we tested the ability of CVAK104 to function as a kinase. Initially, we analyzed the ability of baculovirus-expressed CVAK104-GST fusion protein to bind ATP. As a positive control, we used the AAK1-GST fusion protein that, as expected, was capable of binding a UV light cross-linkable, biotinylated ATP analogue. This binding was effectively competed by incubation with unmodified ATP (Fig. 4). Similarly, we found that CVAK104 bound the ATP analogue, which also could be competed with unlabeled ATP. The identity of the nucleotide binding species was confirmed by immunoblot analysis. Basic polypeptide chains, such as poly-L-lysine, stimulate endogenous kinases associated with clathrin-coated vesicles; therefore, we tested their effect on CVAK104 nucleotide binding. However, the presence or absence of poly-L-lysine did not alter the ability of CVAK104 to bind nucleotide (Fig. 4).
CVAK104 directly interacts with AP2, is present in enriched adaptor protein preparations, and binds nucleotide. Thus, we tested its ability to function as a phosphotransferase and to target adaptor proteins. The activity of endogenous kinases that co-fractionate with adaptor proteins from bovine brain was readily visualized following the incubation of samples with [␥ 32 P]ATP (Fig. 5A, lanes labeled APs). The major phosphorylated product, which results from the action of AAK1 (5), was the 2-subunit of AP2. The addition of poly-L-lysine to these kinase reactions resulted in the stimulation of an endogenous kinase(s) leading to an increase in the phosphorylation of the large adaptor protein subunits, the adaptins. When isolated CVAK104 was incubated with [␥ 32 P]ATP, we did not observe any significant kinase activity. However, when reactions were supplemented with poly-L-lysine, CVAK104 was capable of autophosphorylation (Fig. 5A, lanes labeled CVAK). To assay the ability of CVAK104 to target adaptor proteins, we inactivated endogenous kinases by pre-incubation with the ATP analogue FSBA, which irreversibly binds the active site and inhibits kinase activity (5,12). The addition of baculovirus-expressed CVAK104-GST fusion protein and poly-L-lysine to FSBA-inactivated AP1 samples did not reveal any significant CVAK104mediated phosphorylation beyond CVAK104 autophosphorylation (Fig. 5A, lanes labeled FSBA-AP1), consistent with the inability of CVAK104 to directly interact with AP1 (Fig. 3). In contrast, incubation of CVAK104 and poly-L-lysine with AP2 resulted in the phosphorylation of proteins at a molecular weight consistent with that of the large adaptin subunits (Fig.  5A, lanes labeled FSBA-AP2). As was observed for autophosphorylation, CVAK104-mediated phosphorylation of the AP2associated adaptins required the presence of poly-L-lysine.
To determine which adaptin subunit was being phosphorylated following CVAK104 addition, the subunits were resolved using two-dimensional electrophoresis employing a urea- SDS   FIG. 3. CVAK104 binds endocytic vesicle coat components. Isolated clathrin triskelia, AP1, and AP2 (ϳ5 g of each) were incubated with either baculovirus-expressed CVAK104-GST fusion protein or GST immobilized on glutathione beads, as indicated. Coat protein binding to immobilized CVAK104 was detected by immunoblot analysis following SDS-PAGE. A 10% fraction of the amount of coat protein that was incubated with immobilized CVAK104-GST or GST is shown. The monoclonal antibody TD.1 that recognizes the clathrin heavy chain and 100/1 that recognizes the ␤1and ␤2-adaptin subunits of AP1 and AP2, respectively, were used to identify interacting proteins. Note that minor contaminating amounts of AP2 in the AP1 preparation are detected with the monoclonal antibody 100/1 .   FIG. 4. CVAK104 binds ATP in a poly-L-lysine-independent manner. AAK1-and CVAK104-GST fusion proteins were incubated with the biotinylated, UV light cross-linkable ATP analogue AB11 (50 M, 2-azidoadenosine 5Ј-triphosphate 2Ј,3Ј-biotin-long chain-hydrazone) in the presence or absence of poly-L-lysine (50 g/ml) and/or unmodified ATP at the indicated concentrations. Protein load was verified by immunoblot using polyclonal antibodies against AAK1 or CVAK104 (bottom split panel).
FIG. 5. CVAK104 is a poly-L-lysine-stimulated kinase that targets the ␤2-subunit of AP2. A, crude untreated adaptor proteins (APs) or highly enriched AP1 or AP2 fractions treated with FSBA to inactive endogenous kinases (see "Experimental Procedures") were incubated with [␥ 32 p]ATP in the presence or absence of CVAK104-GST fusion protein and/or 50 g/ml poly-L-lysine, as indicated. Kinase assays incorporating CVAK104-GST fusion protein, 50 g/ml poly-L-lysine, and [␥ 32 p]ATP were performed in the absence (B and C) or presence (D-G) of AP2 complexes. Protein samples were then resolved in two dimensions, the first employing urea/SDS-PAGE, followed by standard SDS-PAGE for the second dimension (23). Following transfer to nitrocellulose, phosphorylated proteins were detected by autoradiography (B, D, and F) and immunoblot analysis (C, E, and G). The identity of the resolved proteins was determined using polyclonal antibodies against CVAK104 in combination with the monoclonal antibody 100/1 that recognizes ␤2-subunit (E) or polyclonal antibodies raised against a brain-specific isoform of the ␣-subunit of AP2 that was fused to GST (G, see "Experimental Procedures"). Note that the ␣-adaptin-specific antibodies also recognize the GST moiety of the CVAK104-GST fusion protein. Arrows are labeled in the figure. Circles indicate corresponding regions of the autoradiogram and immunoblot. system in the first dimension, followed by standard SDS-PAGE in the second dimension. This approach, developed by Ahle et al. (23), separates the large adaptin subunits, as these proteins fail to resolve under isoelectric focusing conditions (23,29). When kinase reactions containing CVAK104, poly-L-lysine, and AP2 were resolved under this system, immunoblot analysis indicated that the ␤2-subunit of AP2 was selectively phosphorylated (Fig. 5, D and E) with no detectable phosphorylation of the ␣-adaptin subunit (Fig. 5, F and G).
In addition to assessing CVAK104 activity toward isolated adaptor proteins, other potential substrates were tested. Poly-L-lysine-stimulated CVAK104 was also found to phosphorylate an ϳ60-kDa protein from FSBA-inactivated cytosol (Fig. 6). Interestingly, when CVAK104 is incubated with Tris-extracted proteins from isolated microsomes that contain clathrin coat proteins, we observed a phosphorylated band at ϳ90 kDa and no detectable phosphorylation of the ␤2-subunit of AP2. Additionally, we tested CVAK104 activity toward poly(Glu:Tyr) (4:1) peptides, a general substrate for protein tyrosine kinases. However, the only phosphorylated product detected was CVAK autophosphorylation (Fig. 6). Collectively, these observations demonstrate that CVAK104, like many kinases, has multiple targets in different cellular locations.
To further characterize CVAK104 activity, we analyzed its sensitivity to an array of kinase inhibitors. Kinase reactions incorporating staurosporine (2.7 nM), a broad-spectrum serine/ threonine kinase inhibitor, showed a marked reduction in CVAK104 autophosphorylation and activity toward the ␤2subunit of AP2. In contrast, CVAK104 was relatively insensitive to the cAMP-dependent kinase inhibitor H89 (100 M), heparin (5 g/ml), a casein kinase II inhibitor, cAMP-dependent protein kinase inhibitor peptides (600 nM), or protein kinase C/calmodulin kinase inhibitor mixtures ( Fig. 7A; see "Experimental Procedures"). Moreover, CVAK104 activity was not perturbed by incubation with genistein (10 M) at concentrations known to inhibit a range of protein tyrosine kinases, consistent with its lack of activity toward tyrosine kinase substrates (poly(Glu:Tyr) peptides) (Fig. 6). Treatment of synaptosomes with reagents that inhibit cdk5 activity alters clathrinmediated synaptic vesicle recycling, although with differentially reported effects (30,31). Given that CVAK104 targets the ␤2subunit of AP2, a core component of the endocytic machinery, we examined CVAK104 sensitivity to the cdk5 inhibitors olomoucine and roscovitine. Olomoucine treatment had little effect on CVAK104 activity toward ␤2-adaptin (Fig. 7B), whereas roscovitine showed an inhibitory effect at 10 M, a concentration 10-fold greater than that required to inhibit cdk5. The differential activities of roscovitine versus olomoucine may, in part, account for their differential effects on synaptic vesicle recycling in neurons (30,31). Collectively, these observations suggest that CVAK104 is a poly-L-lysine-stimulated serine/threonine kinase whose activity is responsible for the phosphorylation of the ␤2-adaptin subunit of AP2 in isolated adaptor protein preparations. DISCUSSION We have identified a novel coated vesicle-associated kinase, CVAK104, which belongs to the SCY1-like family of protein kinases, previously thought to be catalytically inactive, as they lack the highly conserved catalytic aspartic acid (25). Although we do not currently understand the mechanism of catalysis, here we provide evidence that CVAK104 is a poly-L-lysinestimulated serine/threonine kinase that has autocatalytic activity and targets the large ␤2-subunit AP2. The isolated CVAK104-GST fusion protein used in our studies derives from insect cell overexpression; thus, we cannot eliminate the possibility that our protein samples are contaminated with trace amounts of another tightly associated kinase. However, we do not favor this scenario for several reasons. 1) It is unlikely that a contaminating baculovirus-derived kinase would share the same properties as that of the endogenous activity found in adaptor protein preparations. CVAK104 is not only present in adaptor protein preparations, its activity, similar to that of the observed endogenous activity, is stimulated by the presence of basic polypeptides and appears to be specific for the ␤2-subunit. 2) Only a single population of nucleotide-binding protein, corresponding to that of isolated CVAK104, was observed in our ATP binding studies. 3) During the course of this study, KIAA1360 (CVAK104) was also found in an extensive mass spectroscopic analysis of highly purified clathrin-coated vesi-FIG. 6. CVAK104 has multiple substrates. Isolated cell fractions were treated with FSBA to inactivate endogenous kinase activities. Samples were then incubated with 50 g/ml poly-L-lysine and [␥ 32 p]ATP in the presence or absence of isolated CVAK-GST fusion protein as indicated. Poly(Glu:Tyr) (4:1) peptides were also analyzed as a potential substrate for CVAK-GST. CVAK-GST-specific phosphorylated protein products are indicated by the arrows. Phosphorylated proteins were visualized by autoradiography using a phosphorimaging plate following SDS-PAGE.
FIG. 7. CVAK104 is sensitive to a general serine/threonine kinase inhibitor. CVAK104 sensitivity to kinase inhibitors was assayed by incubating CVAK104-GST fusion protein, 50 g/ml poly-L-lysine, and [␥ 32 p]ATP with isolated AP2, in the presence or absence of the indicated inhibitor (A, upper panel). Immunoblot analysis was performed to verify equal gel loading using polyclonal antibodies against CVAK104 and the 100/1 monoclonal antibody that recognizes ␤2-adaptin (A, lower panel). B, the serine/threonine kinase activity of CVAK was further explored by testing CVAK sensitivity to olomoucine and roscovitine (cyclin-dependent kinase inhibitors) at the indicated concentrations. All samples were resolved by SDS-PAGE and phosphorylated proteins were detected by phosphorimaging; PKA, cAMP-dependent protein kinase; PKC, protein kinase C. CamK, Calmodulin kinase. cles along with AAK1 and GAK, 2-adaptin kinases, and phosphofructokinase (28,32). However, in neither our analysis nor that of others were any other kinases identified. Thus, it is likely that CVAK104 corresponds to the previously unidentified poly-L-lysine-stimulated ␤-adaptin kinase associated with clathrin-coated vesicles and coat protein preparations.
A second poly-L-lysine-stimulated, casein kinase II-like activity also associates with clathrin-coated vesicles, and following stimulation, it phosphorylates a broad range of coat protein constituents, including the clathrin light chain (16). The activity we observed following CVAK104 addition to kinase assays cannot result from casein kinase II contamination, because unlike casein kinase II, which is potently inhibited by heparin (33), CVAK104 activity is heparin-insensitive. Therefore, we believe that CVAK104 is indeed a catalytically active kinase, although future studies incorporating site-directed mutagenesis will be needed to definitively demonstrate its activity and define the catalytic mechanism.
Although ␤2-adaptin appears to be the only substrate targeted in coated vesicle extracts, CVAK104 can phosphorylate a 60-kDa protein in cytosolic fractions and a 90-kDa protein extracted from microsomal membranes. The identity of these other substrates remains to be determined, but these data suggest that CVAK104, similar to other kinases, may control multiple protein activities in vivo. Whether these other substrates are also involved in vesicular trafficking or function in other biological processes also remains to be determined.
What is the functional consequence of CVAK104-mediated phosphorylation of ␤2-adaptin? The ␤2-subunit is known to support interaction with AP180, eps15, epsin, amphiphysin, and clathrin (6,34). These protein interactions occur over a relatively small region of the protein, and thus it is unlikely that ␤2 can support the simultaneous interaction with each protein. Thus, it is feasible that phosphorylation may regulate the ability of ␤2 to interact with other endocytic factors. Previous observations suggest that the poly-L-lysine-dependent phosphorylation of ␤2, at multiple locations, prevents its recruitment to and sedimentation with preassembled clathrin cages (13). However, our attempts to repeat this observation were unsuccessful; the addition of poly-L-lysine, which is required for ␤2-adaptin kinase activity, cross-links clathrin coat constituents (35). Hence, in our hands, ␤2-adaptin was consistently found associated with clathrin cages, following the addition of poly-L-lysine and high speed centrifugation, independent of its phosphorylation state (not shown).
␤2-adaptin phosphorylation may also influence AP2 localization. Previous studies (11) have established that ␤2 is phosphorylated in vivo by a staurosporine-sensitive kinase whose function is balanced by the constitutive activity of protein phosphatase 2A. Lauritsen et al. (11) observed that treatment of Jurkat cells with agents that block protein phosphatase 2A function perturb AP2 localization at the plasma membrane and disrupt transferrin internalization. Whether CVAK104 is responsible for this phosphorylation reaction in vivo or whether ␤2-adaptins are also substrates for other kinases, remains to be determined.
In summary, we have identified CVAK104 as an endogenous poly-L-lysine-stimulated kinase that phosphorylates the ␤2subunit of the AP2 complex. To gain a better understanding of CVAK104 function, it will be important to identify other interacting partners and isolate and characterize its other substrates.