Coordinated Movement of RACK1 with Activated βIIPKC*

Protein kinase C (PKC) isozymes move upon activation from one intracellular site to another. PKC-binding proteins, such as receptors for activated C kinase (RACKs), play an important role in regulating the localization and diverse functions of PKC isozymes. RACK1, the receptor for activated βIIPKC, determines the localization and functional activity of βIIPKC. However, the mechanism by which RACK1 localizes activated βIIPKC is not known. Here, we provide evidence that the intracellular localization of RACK1 changes in response to PKC activation. In Chinese hamster ovary cells transfected with the dopamine D2L receptor and in NG108-15 cells, PKC activation by either phorbol ester or a dopamine D2 receptor agonist caused the movement of RACK1. Moreover, PKC activation resulted in thein situ association and movement of RACK1 and βIIPKC to the same intracellular sites. Time course studies indicate that PKC activation induces the association of the two proteins prior to their co-movement. We further show that association of RACK1 and βIIPKC is required for the movement of both proteins. Our results suggest that RACK1 is a PKC shuttling protein that moves βIIPKC from one intracellular site to another.

Specific intracellular localization of signaling proteins such as PKC 1 is important for the regulation of complex signal transduction cascades (1). PKC is a family of 10 isozymes that are localized to specific intracellular sites in unstimulated cells. Upon activation, each PKC isozyme moves to a different intracellular site (2). Localization of inactive or activated PKC isozymes is mediated, at least in part, by interaction with anchoring proteins (3,4). For example, inactive PKC isozymes appear to be localized by binding to the scaffolding proteins AKAP-79 and gravin (5,6). In contrast, activated PKC isozymes are localized by binding to receptors for activated C kinase (RACKs). RACK1 specifically binds the active form of ␤IIPKC (7,8) thereby regulating PKC function (8 -12). In vitro, RACK1 binds PKC only in the presence of PKC activators and increases PKC kinase activity, presumably by stabilizing its active conformation (13). The RACK1 binding site on PKC is within the C2 region of the regulatory domain providing a direct protein-protein interaction (8). Indeed, RACK1 belongs to the WD40 family of proteins, and the WD40 motif is implicated in mediating protein-protein interactions (14). Furthermore, peptides derived from either PKC and/or RACK1 can alter PKC activity in vitro and in vivo (8,12,15,16).
Although RACK1 binds activated PKC and is clearly important for PKC function, the mechanism by which RACK1 localizes ␤IIPKC to its site after activation is not understood. One prediction is that the anchoring protein RACK1 should always be localized to the same site that accepts ␤IIPKC after translocation. We therefore used confocal microscopy to determine whether RACK1 is co-localized with activated ␤IIPKC, whether RACK1 is localized to a specific organelle, and whether the intracellular localization of RACK1 changes in response to PKC activation. Here, we provide evidence that RACK1 is localized to different sites in unstimulated and stimulated cells and that PKC activation leads to movement of RACK1. Furthermore, PKC activation induces the association and co-localization of RACK1 with ␤IIPKC. Based on these results, we propose that RACK1 is a shuttling protein that localizes ␤IIPKC upon activation by shuttling the kinase to its appropriate subcellular site.

Cell Culture
Chinese hamster ovary (CHO) cells stably expressing the long form of the dopamine D2 receptor (D2L) (17) were seeded and grown in Ham's F-12 medium containing 10% FBS and 2 mM glutamine. After 48 h, media were replaced with Ham's F-12 medium containing 5% serum and 25 mM HEPES (pH 7.4), 2 mM glutamine, 50 g/ml human transferrin, 10 g/ml oleic acid (complexed with 2 mg/ml fatty acid-free bovine serum albumin), 25 g/ml bovine insulin, and trace elements at the following concentrations: 0.5 nM MnCl 2 , 0.5 nM (NH 4 ) 6 1. PKC activation induces RACK1 movement. CHO/D2L cells were treated with either 100 nM PMA for 10 min at 37°C or with a 500 nM concentration of a D2 receptor agonist, NPA. NG108-15 cells were treated with 100 nM PMA. Cells were washed, fixed, and blocked as described under "Experimental Procedures." RACK1 localization was assayed by immunostaining with monoclonal anti-RACK1 antibodies (Transduction Laboratories, 1:100). Cells were scanned using a confocal microscope and viewed with a ϫ 60 lens for CHO/D2L and a ϫ 40 for NG108-15 cells. RACK1 staining is represented by false color; intensity is represented by the false color bar on the left, with red indicating the areas with the most intense staining. No change in staining intensity or localization was observed when unstimulated cells were incubated with fresh medium containing 0.001% Me 2 SO in the medium or with media containing 0.01% ascorbic for 10 min to 1 h. a, control CHO/D2L cells; b, CHO/D2L cells treated with 100 nM PMA for 10 min; c, CHO/D2L cells treated with 500 nM NPA for 30 min; d, unstimulated CHO/D2L cells stained with anti-RACK1 antibodies (1:100) that were preabsorbed overnight at 4°C with 5 g of the recombinant fusion protein maltose-binding protein RACK1. e, unstimulated NG108-15 cells; f, NG108-15 cells treated with 100 nM PMA for 10 min. The images are representative of more than 10 individual experiments for CHO/D2L and 3 experiments for NG108-15. The images shown are individual middle sections of the projected Z-series.
FIG. 2. PKC activation induces ␤IIPKC movement to the Golgi apparatus in CHO cells. CHO/D2L cells were treated with 100 nM PMA for 10 min at 37°C. Cells were washed, fixed, and blocked as described under "Experimental Procedures." Cells were then stained with both polyclonal anti-␤IIPKC antibody (1:100) (a-d) and monoclonal anti-mannosidase antibodies (1:5000) (e) and visualized with fluorescein isothiocyanate-conjugated goat anti-rabbit antibodies (1:500) for ␤IIPKC and Texas Red-conjugated goat anti-mouse (1:500) antibodies for mannosidase. Cells were scanned using a confocal microscope and viewed with a ϫ 40 lens. No change in staining intensity or localization was observed when unstimulated cells were incubated with fresh media, with 0.001% Me 2 SO in the medium (because PMA stock solution is dissolved in Me 2 SO), or with 0.01% ascorbic acid in the medium (because dilution of NPA is done in medium containing 0.01% ascorbic acid) for 10

Immunocytochemistry and Confocal Microscopy
CHO/D2L or NG108-15 cells were treated with different reagents as described in the figure legends. The cells were then washed with cold phosphate-buffered saline (PBS), fixed with ice-cold methanol for 3 min, and then washed twice with cold PBS. Cells were incubated for 2 h with 1% normal goat serum in PBS containing 0.1% Triton X-100, followed by overnight incubation at 4°C with the appropriate primary antibody (diluted in PBS containing 0.1% Triton X-100 and 2 mg/ml bovine serum albumin). Cells were then washed three times with PBS containing 0.1% Triton and incubated for 1.5 h with fluorescein isothiocyanateconjugated anti-rabbit IgG antibody (1:500, Cappel), Texas Red, or Cy5-conjugated anti-mouse (IgM) antibodies (1:500, Cappel). Cells were washed an additional three times with cold PBS containing 0.1% Triton. Slides were mounted using Vectashield and viewed with a Bio-Rad MRC-1024 laser scanning confocal microscope. The confocal images were processed using the computer programs NIH Image, version 1.61 (National Institutes of Health), and Adobe Photoshop (Adobe Systems Inc.). All the images shown are individual middle sections of projected Z-series.

Image Analysis
Quantification of Co-localization-Co-localization of the pairs RACK1 and ␤IIPKC or mannosidase and ␤IIPKC was determined using the NIH Image program, version 1.61. Threshold was used to separate immunofluoresence pixels from background and to create binary images. The intensity values of each image were recorded. Pairs of binary images of RACK1 and ␤IIPKC or mannosidase and ␤IIPKC were multiplied and divided by 255 to generate a final merged image that could be visualized with an 8-bit gray scale. The intensity value of the final merged image was recorded. The percentage of RACK1 staining merged with ␤IIPKC staining was calculated using the following equation.
The number of pixels with intensity value Ͼ0 in the final merged image The number of pixels with intensity ϫ100 Each image contained 20 -100 cells.
Scoring Movement-Movement of RACK1 and ␤IIPKC was scored by counting at least three random fields of cells (total of at least 100 cells per treatment) for staining of RACK1 and/or ␤IIPKC. The percentage of translocated (moved) RACK1 and/or ␤IIPKC was calculated using the following equation.

The number of cells in which RACK1
or ␤IIPKC stained at the Golgi Total number of cells ϫ 100 Each image contained 20 -100 cells.
Co-Immunoprecipitation 10 7 cells were incubated for 30 min at 37°C with either fresh media, 100 nM PMA, or 500 nM NPA. Cells were washed once with cold PBS and lysed in 20 mM Tris-HCl (pH 7.5) containing 10 mM EGTA, 2 mM EDTA, 0.25 M sucrose, 1% Triton X-100, and 10 g/ml of the following protease inhibitors: soybean trypsin inhibitor (Sigma), aprotinin, phenylmethylsulfonyl fluoride (Sigma), and leupeptin. Lysed cells were centrifuged at 14,000 ϫ g at 4°C for 10 min. The Triton-soluble material (supernatant) was precleared by incubation with 50 l of protein G agarose (Life Technologies, Inc.) for 2 h at 4°C. The samples were centrifuged and protein quantity was determined using BCA reagent (Pierce). Immunoprecipitation was performed with 5 g of anti-␤IIPKC antibody or anti-RACK1 antibody, together with approximately 0.5 mg of protein diluted in an equal volume of 2ϫ immunoprecipitation buffer (1ϫ ϭ 1% Triton X-100, 150 mM NaCl, 10 mM Tris HCl, pH 7.4, 1 mM EDTA, 1 mM EGTA, 0.2 mM sodium vanadate, and 10 g/ml each of soybean trypsin inhibitor, aprotinin, phenylmethylsulfonyl fluoride, and leupeptin) and water to a total volume of 1 ml. After overnight incubation at 4°C, 50 l of protein G agarose was added and the mixture was incubated at 4°C for 2 h. The agarose resin was then washed three times with 1ϫ immunoprecipitation buffer and twice with ice-cold PBS. The sample was centrifuged and sample buffer was added to the pellet fraction. The sample was resolved on a 10% SDS-polyacrylamide gel electrophoresis gel and transferred to a nitrocellulose membrane. The membrane was cut at approximately 50 kDa and probed with monoclonal anti-RACK1 antibodies (lower part) (1:500) and polyclonal anti-␤IIPKC antibodies (upper part) (1:250). Immunoreactivity was detected using a chemiluminescent reaction (2.5 mM luminol and 400 M p-coumaric acid).

RESULTS
In order to determine whether RACK1 is co-localized with activated translocated (moved) ␤IIPKC, we first established the localization of RACK1 in CHO cells using confocal microscopy. As shown in Fig. 1a, RACK1 was localized to a perinuclear structure in unstimulated cells. Next, we determined the localization of ␤IIPKC in CHO cells. As shown in Fig. 2a, ␤IIPKC is localized to the cytoplasm in unstimulated cells. Activation by a phorbol ester (PMA) induced ␤IIPKC to move from the cytoplasm (Fig. 2a) to a site that resembles the Golgi apparatus (Fig. 2b) and not the perinuclear structure where RACK1 is found (Fig. 1a). This structure was identified as the Golgi apparatus by double-staining of cells with anti-␤IIPKC antibodies (Fig. 2d) and anti-mannosidase antibodies (Golgi marker) (Fig. 2e). Colocalization of ␤IIPKC with the Golgi marker was confirmed by performing image analysis on the merged image as described under "Experimental Procedures." ␤IIPKC ( Fig. 2d) co-localized with mannosidase (Fig. 2e) as shown in the merged image (Fig. 2f). Because RACK1 in un- stimulated cells is localized to the perinucleus (Fig. 1a) and is not localized to the Golgi apparatus, where translocated (moved) activated ␤IIPKC is found (Fig. 2b), RACK1 is not always localized to the same site as ␤IIPKC.
We next determined whether RACK1 is always localized to a specific organelle. Specifically, we asked whether RACK1 is always localized to perinuclear structures in CHO cells, regardless of the activation state of the cell. CHO cells that stably express the dopamine D2L receptor (CHO/D2L) were treated with either PMA or with the dopamine D2 agonist NPA, and RACK1 localization was determined. PKC activation either directly with PMA or by NPA activation of the D2 receptor induced the movement of RACK1 from perinuclear structures (Fig. 1a) to a different site (Fig. 1b and c). In another cell line, NG108-15 neuroblastoma ϫ glioma cells, RACK1 was also localized to different intracellular sites before (Fig. 1e) and after PKC activation (Fig. 1f). In NG108-15 cells PKC activation induced RACK1 to move to yet unidentified cytosolic structures and to neurites (Fig. 1f). Therefore, RACK1 is not localized to a specific organelle, and activation of PKC leads to movement of RACK1.
If RACK1 movement is dependent on PKC activation via phosphatidylinositol-derived second messengers, then a PLC inhibitor should inhibit RACK1 movement induced by activation of the D2L receptor. Indeed, movement of both RACK1 (Fig. 3a) and ␤IIPKC (Fig. 3b) was inhibited when CHO/D2L cells were pretreated with the PLC inhibitor ET18OCH 3 prior to activation with NPA. As expected, ET18OCH 3 did not inhibit PMA-induced movement of ␤IIPKC or of RACK1 (Fig. 3), because phorbol esters bypass PLC signaling and directly activate PKC. Therefore, RACK1 movement is dependent on PLC activation, suggesting that under physiologic conditions, the generation of second messengers is required not only for the movement of ␤IIPKC but also for movement of RACK1.
PKC activation induced movement of both RACK1 and ␤IIPKC (Figs. 1 and 2). We therefore determined whether RACK1 and ␤IIPKC become co-localized to the same site after PKC activation. CHO/D2L cells were incubated in the absence and presence of PMA or NPA and stained for both RACK1 and ␤IIPKC. The images were merged in order to detect co-local-ization, and image analysis was performed. As shown in Fig. 4, a and d, in unstimulated cells, approximately 60% of RACK1 and ␤IIPKC were not co-localized. Upon PKC activation, RACK1 and ␤IIPKC moved to the same site, and their intracellular staining patterns merged to more than 70% (Fig. 4,  b-d). In CHO cells, PKC activation induced ␤IIPKC to move to the Golgi apparatus (Fig. 2, d-f). To confirm that RACK1 also localized to the Golgi apparatus upon PKC activation, we stained cells with anti-RACK1 antibodies together with the specific Golgi marker BODIPY FL C 5 -ceramide. As shown in Fig. 4, e and f, RACK1 co-localized with the Golgi marker after PKC activation. Thus, PKC activation induces the co-localization of RACK1 and ␤IIPK and the movement of both proteins from different sites to the same locations.
The co-localization of RACK1 and ␤IIPKC after PKC activation, observed by immunofluoresence (Fig. 4), suggests that the two proteins associate with each other in cells. To explore this possibility, we determined whether the two proteins can be co-immunoprecipitated and whether PKC activation is required for their association. ␤IIPKC was immunoprecipitated from unstimulated, PMA-treated, or NPA-treated cells using anti-␤IIPKC antibodies, and we determined whether RACK1 was co-immunoprecipitated. Anti-␤IIPKC antibodies co-immunoprecipitated RACK1 in CHO-D2L cells (Fig. 5a, lanes 5 and  6) and in NG108-15 cells (Fig. 5b, lane 2), and anti-RACK1 antibodies also co-immunoprecipitated ␤IIPKC (data not shown), indicating that RACK1 and ␤IIPKC do associate in cells. Furthermore, the association between RACK1 and ␤IIPKC was increased by PKC activation with PMA or NPA (Fig. 5a, lanes 5 and 6 compared with lane 7 for CHO/D2L cells and Fig. 5b, lane 2 compared with lane 1 for NG108-15 cells). Anti-␤IPKC antibodies, which were used as control antibodies, did not immunoprecipitate RACK1 (data not shown), indicating that the association between ␤IIPKC and RACK1 is specific. Western blot analysis of RACK1 (30 kDa) and ␤IIPKC (80 kDa) show that the amount of the detected protein does not significantly change with the experimental conditions (Fig. 5,  a, lanes 1-3, and b, lanes 5 and 6), and no cross-reactivity with either antibody was observed (data not shown). Taken to- gether, our data indicate that activation of PKC causes RACK1 and ␤IIPKC to associate with each other.
We next determined whether RACK1 and ␤IIPKC move together or whether the movement of one precedes the other. We therefore compared RACK1 and ␤IIPKC movement (Fig. 6a) and co-localization (Fig. 6, b and c) as a function of time. As shown in Fig. 6a, the time courses of movement for both RACK1 and ␤IIPKC were very similar, indicating that it is unlikely that one protein moves prior to the other. In contrast, the time course of co-localization (Fig. 6, b and c) indicates that the two proteins co-localize prior to their movement. At 1 and 5 min, more than 70% of RACK1 was co-localized with ␤IIPKC, but only 25% of both proteins had reached the Golgi apparatus at that time (Fig. 6, a and b). Because co-localization was detected before RACK1 and ␤IIPKC reached the Golgi apparatus (Fig. 6c, compared with Figs. 1, 2, and 4), it is possible that the two proteins associate prior to their movement.
If prior association of ␤IIPKC and RACK1 is required for movement, then inhibition of RACK1 and ␤IIPKC association should prevent movement. Recently, the PKC inhibitor dequalinium (DECA) has been shown to inhibit PKC movement by interacting with the RACK1 binding site on PKC (1). We determined whether DECA inhibits the interaction of ␤IIPKC with RACK1 and compared the results with the effect of other PKC inhibitors. The regulatory domain inhibitor calphostin C and the kinase domain inhibitors bisindolylmaleimide HCl and chelerythrin chloride were used at concentrations equal to their IC 50 values. Fig. 7a presents an overlay assay of ␤IIPKC binding to immobilized RACK1 in the presence of activators and in the presence of DECA, calphostin C, and bisindolylmaleimide HCl. DECA, as well as calphostin C, reduced the binding of ␤IIPKC to RACK1 (Fig. 7a). On the other hand, bisindolylmaleimide HCl did not affect the interaction between ␤IIPKC to RACK1 (Fig. 7a). Similar effects were obtain with chelerythrine (data not shown due to high background). We next determined whether PKC inhibitors would inhibit the movement of both RACK1 and ␤IIPKC. All inhibitors were used at concentrations equal to their IC 50 values. DECA inhibited movement of ␤IIPKC (Fig. 7b) in CHO/D2L cells. Interestingly, DECA also inhibited the movement of RACK1 (Fig. 7b). These data suggest that activation-induced binding of ␤IIPKC to RACK1 is a prerequisite for the movement of both proteins. Furthermore, the regulatory domain inhibitor calphostin C inhibited NPA-induced movement and, to a lesser degree, PMA-induced movement (Fig. 7b). Because calphostin C is a competitive inhibitor for the DAG binding site, these results are another indication that suggests that generation of second messengers is required for the interaction and movement of both proteins. On the other hand, the kinase domain inhibitors bisindolylmaleimide HCl and chelerythrin did not significantly FIG. 6. NPA-induced RACK1 and ␤IIPKC movement and co-localization as function of time. a, CHO/D2L cells were treated with 500 nM NPA at 37°C. Cells were washed, fixed, and blocked at different time points as described under "Experimental Procedures." Cells were then stained, visualized, and scored as described under "Experimental Procedures." Results are presented as mean Ϯ S.D. of three experiments. The percentage of movement was defined as (the number of cells in which RACK1 or ␤IIPKC was translocated (moved)/ the total number of cells) ϫ 100. Data are the mean Ϯ S.D. of three experiments. b, CHO/D2L cells were treated as in a, and the data were analyzed using the computer program NIH Image 1.61 as described under "Experimental Procedures." The percentage of ␤IIPKC and RACK1 co-localization is defined as (the pixels value of the merged image/the pixel value of RACK1 staining) ϫ 100. Data are the mean Ϯ S.D. of three experiments. c, CHO/D2L cells were treated with 500 nM NPA for 30 s to 5 min as described in Fig.  1. RACK1 staining is red, ␤IIPKC staining is green, and a merged image of the two colors is yellow. Images were viewed with a ϫ 60 lens. Image is a representative of three different experiments. The images shown are individual middle sections of the projected Z-series.
inhibit the movement of ␤IIPKC and RACK1, indicating that PKC kinase activity is not involved in the movement of both proteins (Fig. 7b). DISCUSSION PKC anchoring proteins determine the localization of different activated PKC isozymes (3,4). However, the mechanism by which PKC anchoring proteins localize PKC isozymes to specific sites after movement is not well understood. Here, we provide evidence that RACK1, the anchoring protein for activated ␤IIPKC, also moves upon activation of PKC. RACK1 moves in response to PKC activation and localizes to the same sites as activated ␤IIPKC. The PLC inhibitor ET18OCH 3 blocked dopamine D2 receptor-induced movement of RACK1, and calphostin C, an inhibitor that competes with DAG, interfered with the interaction and movement of RACK1 and ␤IIPKC. These results suggest that generation of second messengers needed for the activation of PKC is also necessary for movement of RACK1. On the other hand, our results with PKC kinase inhibitors suggest that PKC kinase activity per se is not involved in the binding or is required for the movement of the two proteins. These findings are in line with previous data showing that RACK1 itself is not a substrate for PKC (13).
We further show that RACK1 and ␤IIPKC co-localize prior to their movement and that the association of the two proteins appears to be required for their simultaneous movement. Based on these findings, we propose that RACK1 is a PKC shuttling protein. When ␤IIPKC is activated, it binds to RACK1. RACK1 then moves together with ␤IIPKC to bring the enzyme in close proximity to its appropriate substrate.
The association between activated ␤IIPKC and RACK1 in situ was detected by co-immunoprecipitation. These results are in agreement with previous in vitro studies showing that the association of RACK1 and ␤IIPKC occurs only in the presence of the PKC activators phosphatidylserine, DAG, and calcium (13). The early time points of co-localization between RACK1 and ␤IIPKC (30 s to 5 min) could be detected with confocal microscopy but could not be confirmed by immunoprecipitation. Labeling each protein with a different fluoresence tag may allow us to follow the movement and co-localization of RACK1 and ␤IIPKC at early time points in live cells.
Although it is possible that movement of ␤IIPKC is responsible for movement of RACK1, we consider this possibility unlikely for several reasons; RACK1 belongs to the WD40 family of proteins that regulate (via protein-protein interaction) the localization and/or activity of various signaling proteins. For example, the ␤-adrenergic receptor kinase is localized by the WD40-containing protein G␤ (the ␤ subunit of GTP-binding protein) (19); the transforming growth factor-␤ receptors interact with a subunit of phosphatase 2A (a WD40containing protein) (20); cytosolic phospholipase A2 binds to the WD40-containing protein PLAP (21), and ⑀PKC is localized by yet another WD40-containing RACK, RACK2 (22). Furthermore, PKC-mediated functions are inhibited when the association between RACK1 and PKC is disrupted by peptides (8,16). Therefore, it is most likely that RACK1 is directing activated ␤IIPKC to a specific site.
Furthermore, PKC activation induces ␤IIPKC to move to different sites in different cells. For example, in NIH3T3 cells, activated ␤IIPKC is found in cytoskeletal elements (23); in cardiac myocytes, activated ␤IIPKC is localized to perinuclear structures; and in human leukemic cell lines, ␤IIPKC moves to the nuclear membrane, where it phosphorylates lamin B (24). Activation induces ␤IIPKC to move to cytoplasmic filaments (25) in human endothelial cells, and to the plasma membrane in HEK 293 cells (26). In addition, different stimuli cause ␤IIPKC to move to different intracellular sites in the same cell (2,25). Therefore, it is not surprising that we detected ␤IIPKC movement to the perinulcear structures in CHO cells and to neurites in NG108 cells. We suggest that RACK1 can localize activated ␤IIPKC to different sites because it is a mobile rather than a fixed protein. This is consistent with our finding that RACK1 is not associated to a specific organelle and with other reports that RACK1 is localized at different sites (10,27). Indeed, sequence analysis reveals that RACK1 does not contain consensus sequence motifs that could anchor it to a particular subcellular site. Thus, the mobility of RACK1 enables it to shuttle ␤IIPKC to different sites in different cells. These observations also suggest that RACK1 movement may be affected by other signaling cascades. Indeed, we have found that treatment with ethanol induces RACK1 to move to the nucleus, whereas ␤IIPKC localization is unchanged in three different cell lines (NG108-15, CHO, and C6), as well as in certain brain areas of mice. 2 Furthermore, we found that forskolin (an activator of adenylate cyclase) also induces the nuclear movement of RACK1 but not ␤IIPKC. 3 Taken together, our studies indicate that different stimuli induce the recruitment of RACK1 to different sites. These results also suggest that the intracellular localization of RACK1 does not depend exclusively upon PKC activation, whereas the movement of ␤IIPKC is directed by RACK1.
RACK1 may represent a new class of mobile targeting proteins. It is conceivable that other anchoring, scaffolding, or adaptor proteins may also shuttle signaling proteins between intracellular sites. One possible candidate is the adaptor protein 14-3-3 that has recently been shown to bind both inactive Raf in the cytosol and active Raf at the plasma membrane (28). 14-3-3 protein could be a Raf shuttling protein that is responsible for movement of Raf from the cytosol to the plasma membrane. Other candidates are members of the AKAP family of proteins that are redistributed in response to stimuli (29,30).
What mediates the localization of RACK1? It is possible that RACK1 localization is determined by interaction with organelle specific proteins. One intriguing possibility is that RACK1 targets membranes of organelles via binding to the pleckstrin homology (PH) domains that bind both phospholipids and proteins. Indeed, WD40-containing proteins have been found to interact with PH domain-containing proteins (31). The WD40 motif of G␤ binds to the PH domain of ␤-adrenergic receptor kinase (31,32), and RACK1 itself was found to bind PH domains in vitro. 4 Another possibility is that RACK1 is associated with a PKC substrate after movement. Some of the PH-containing proteins, such as pleckstrin, are PKC substrates, and RACK1 associates with the cytoplasmic domain of ␤-integrins only in the presence of PMA (10). Based on the translocating properties of RACK1, it is possible that activation of PKC causes movement of RACK1 together with ␤IIPKC to the plasma membrane, where RACK1 binds to the cytoplasmic tail of ␤-integrin, allowing PKC to phosphorylate either ␤-integrin or neighboring proteins.
In summary, our data show that RACK1 localization is regulated by PKC activation and suggest that RACK1 is a PKC shuttling protein. The shuttling properties of RACK1 and other members of its class may add another dimension to our understanding of how PKC isozymes are localized to different sites after activation and movement.