Regulation of Phosducin-like Protein by Casein Kinase 2 and N-terminal Splicing*

Phosducin-like protein (PhLP) is a member of the phosducin family of G-protein βγ-regulators and exists in two splice variants. The long isoform PhLPLand the short isoform PhLPS differ by the presence or absence of an 83-amino acid N terminus. In isolated biochemical assay systems, PhLPL is the more potent Gβγ-inhibitor, whereas the functional role of PhLPS is still unclear. We now report that in intact HEK 293 cells, PhLPS inhibited Gβγ-induced inositol phosphate generation with ∼20-fold greater potency than PhLPL. Radiolabeling of transfected HEK 293 cells with [32P] revealed that PhLPL is constitutively phosphorylated, whereas PhLPS is not. Because PhLPL has several consensus sites for the constitutively active kinase casein kinase 2 (CK2) in its N terminus, we tested the phosphorylation of the recombinant proteins by either HEK cell cytosol in the presence or absence of kinase inhibitors or by purified CK2. PhLPL was a good CK2 substrate, whereas PhLPS and phosducin were not. Progressive truncation and serine/threonine to alanine mutations of the PhLPL N terminus identified a serine/threonine cluster (Ser-18/Thr-19/Ser-20) within a small N-terminal region of PhLPL (amino acids 5–28) as the site in which PhLPL function was modified in HEK 293 cells. In native tissue, PhLPL also seems to be regulated by phosphorylation because phosphorylated and non-phosphorylated forms of PhLPL were detected in mouse brain and adrenal gland. Moreover, the alternatively spliced isoform PhLPS was also found in adrenal tissue. Therefore, the physiological control of G-protein regulation by PhLP seems to involve phosphorylation by CK2 and alternative splicing of the regulator.

Phosducin-like protein (PhLP) is a member of the phosducin family of G-protein ␤␥-regulators and exists in two splice variants. The long isoform PhLP L and the short isoform PhLP S differ by the presence or absence of an 83-amino acid N terminus. In isolated biochemical assay systems, PhLP L is the more potent G␤␥-inhibitor, whereas the functional role of PhLP S is still unclear. We now report that in intact HEK 293 cells, PhLP S inhibited G␤␥-induced inositol phosphate generation with ϳ20fold greater potency than PhLP L . Radiolabeling of transfected HEK 293 cells with [ 32 P] revealed that PhLP L is constitutively phosphorylated, whereas PhLP S is not. Because PhLP L has several consensus sites for the constitutively active kinase casein kinase 2 (CK2) in its N terminus, we tested the phosphorylation of the recombinant proteins by either HEK cell cytosol in the presence or absence of kinase inhibitors or by purified CK2. PhLP L was a good CK2 substrate, whereas PhLP S and phosducin were not. Progressive truncation and serine/threonine to alanine mutations of the PhLP L N terminus identified a serine/threonine cluster (Ser-18/ Thr-19/Ser-20) within a small N-terminal region of PhLP L (amino acids 5-28) as the site in which PhLP L function was modified in HEK 293 cells. In native tissue, PhLP L also seems to be regulated by phosphorylation because phosphorylated and non-phosphorylated forms of PhLP L were detected in mouse brain and adrenal gland. Moreover, the alternatively spliced isoform PhLP S was also found in adrenal tissue. Therefore, the physiological control of G-protein regulation by PhLP seems to involve phosphorylation by CK2 and alternative splicing of the regulator.
Phosducin-like protein (PhLP) 1 was initially cloned as a product of an ethanol-responsive gene in NG108 -15 neuroblastoma x-glioma cells (1). PhLP has a high sequence homology to the G-protein ␤␥-subunit (G␤␥) regulator phosducin. Similarly, as phosducin, PhLP has been found to bind to G␤␥ with high affinity (2) and act as a G␤␥-binding protein in intact cells (3).
The PhLP mRNA exists in two major splice variants coding for two proteins, which differ by the presence (PhLP L ) or absence (PhLP S ) of the N-terminal 83 amino acids (1,4). Subsequent studies on diverse tissues demonstrated that PhLP L is expressed in many organs at high protein levels (e.g. brain, heart, and liver), whereas PhLP S protein levels could not or only hardly be detected in these organs (4,5). Direct comparison of the affinities toward G␤␥ showed a 15-fold weaker interaction for PhLP S compared with phosducin or PhLP L (5). Therefore, PhLP S had been believed to play only a minor role in G-protein regulation. Recently, it was reported that PhLP S could be purified from cultured bovine chromaffin cells where it might inhibit nicotine-stimulated exocytosis of catecholamines by a pathway involving a G␤␥⅐ADP-ribosylation factor 6 complex (6). These findings suggested that alternative splicing of PhLP might indeed occur in at least one organ or under specific conditions and therefore might play a role in differential functions of the PhLP isoforms.
Regulation of the homologous protein phosducin has been found to occur by phosphorylation via the serine/threonine kinases PKA, GRK, or CaM-dependent kinase II (7)(8)(9)(10)(11). Phosphorylation of phosducin reduces the affinity for G␤␥, and as a consequence, phosducin looses its ability to regulate G-protein function. In the case of PhLP, phosphorylation as a regulatory mechanism has been suggested but has never been demonstrated (10). Here we report that the G␤␥-regulatory function of PhLP L is inhibited in cells by N-terminal phosphorylation via casein kinase 2 (CK2), whereas PhLP S lacks such a regulatory mechanism. CK2 is a constitutively active and ubiquitously expressed serine-threonine kinase (12,13). Our findings further provide strong evidence that the extended N terminus of PhLP L becomes autoinhibitory upon phosphorylation by CK2. In contrast, alternative splicing leads to PhLP S , which is not a substrate for CK2 and thus escapes this inhibitory mechanism.

EXPERIMENTAL PROCEDURES
Materials-[ 32 P]ATP, [ 32 P]orthophosphoric acid, and myo-[2-3 H]inositol were purchased from PerkinElmer Life Sciences. The kinase inhibitors staurosporine, BAPTA-AM, and 5,6-dichloro-1-␤-D-ribofuranylbenzimidazole (DRB) were purchased from Calbiochem. Heparin sodium salt (average molecular weight 6,000) was from Sigma. Primary antibodies used were rabbit polyclonal anti-PhLP S as described previously (5) and rabbit polyclonal anti-PhLP-CT. This synthetic peptide was derived from the very C terminus of PhLP L and PhLP S (sequence: CHSEDSDLEID). After coupling to keyhole limpet hemocyanin (14,15), antibodies were raised in rabbits with two boost injections and collected as serum. Secondary goat anti-rabbit-horseradish peroxidase was from Dianova. Protein A-Sepharose was obtained from Amersham Biosciences. Phosphorimaging and quantification were done on a FLA-3000 from Fuji. ScanProsite was accessed at www.expasy.ch.
Construction of Expression Vectors-All of the cDNAs used in these studies were subcloned into pcDNA3. The cDNA for PhLP L was originally cloned from rat brain (2). The construction of deletion mutants and Ser/Thr to Ala mutants of PhLP L were done by a PCR-based strategy and confirmed by automated sequencing. * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Cell Culture, Transient Transfection, and Determination of Inositol Phosphates-Human embryonic kidney (HEK) 293 cells were grown in DMEM, 10% fetal calf serum and were kept in 7% CO 2 humidified atmosphere. Cells were transfected by using the CaPO 4 method (16) on 10-cm dishes at 70% confluency with a constant total amount of DNA. Transfection efficiency was usually between 60 and 80% and was equal for all of the cDNAs used as controlled by transfection of different green fluorescent protein-tagged constructs. For determination of inositol phosphates, cells were seeded in six-well plates and labeled with myo-[2-3 H]inositol (2 Ci/ml; specific activity 21 Ci/mmol) for 16 h in inositol-free RPMI 1640 medium containing 0.2% fetal calf serum. Unlabeled cells from the same transfection also seeded in six-well plates were used for Western blotting to control expression levels. After labeling, cells were washed once in incubation buffer (138 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 1.6 mM CaCl 2 , 1 g/liter glucose, 20 mM Na ϩ -HEPES, pH 7.3) and then incubated at 37°C for 20 min in incubation buffer containing 10 mM LiCl. Reactions were stopped, and inositol phosphates were extracted as described previously (17). Inositol determinations were performed in triplicates, and results were analyzed as the means Ϯ S.E. of at least three independent experiments. ANOVA and post-test comparison (Bonferroni) were performed as appropriate. For the control of expression levels, cells of one well were lysed in 250 l of hypotonic lysis buffer (1% Triton X-100, 20 mM Tris base, pH 10.5) for 15 min on ice and then centrifuged. Supernatants were analyzed by Western blots using polyvinylidene difluoride membranes (Millipore) and anti-PhLP-CT antibodies. Detection was performed with goat antirabbit-horseradish peroxidase antibodies and the Uptilight horseradish peroxidase-blotting kit (Interchim).
Phosphorylation of Phosducins-Phosducin, PhLP L , and PhLP S were purified from Escherichia coli as C-terminally His 6 -tagged proteins (5). PKA phosphorylation was performed as described (10), and phosphoproteins were visualized by exposure to Biomax MS films (Eastman Kodak Co.). CK2 kinase assays were performed with recombinant human CK2 enzyme (Roche Molecular Biochemicals) or cell extracts. In kinase assays using recombinant human CK2, a 50-l volume containing 400 nM recombinant PhLP L , PhLP S or phosducin, CK2 buffer (20 mM Tris-HCl, pH 7.5, 50 mM KCl, 10 mM MgCl, 5 mM dithiothreitol), 0.04 or 0.3 milliunit of CK2, and 200 M ATP (1 Ci of [ 32 P]ATP) was incubated at 30°C for 30 min or at 37°C for 120 min. The reaction was terminated by boiling samples in Laemmli buffer followed by SDS-PAGE, Coomassie Blue staining, and phosphorimaging. For kinase assays using HEK cell cytosol, a lysate was prepared by resuspending cells of one 15-cm dish in 5 ml of CK2 buffer (supplemented with 1 mM PMSF), disrupting the cells by sonication and clearing the lysate by centrifugation (20,000 ϫ g, 10 min). Protein content was determined in the supernatant by the Bradford method. Phosphorylation was then performed by adding 150 g of total protein of cell extract to a 250-l reaction with 400 nM His 6 -tagged protein and 10 Ci of [ 32 P]ATP and incubated at 30°C for 30 min. Reactions were then placed on ice, diluted with 5 volumes of ice-cold pull-down buffer (20 mM Tris-HCl, pH 8.0, 500 mM NaCl), supplemented with 50 l of Ni-NTA-agarose (Qiagen), and rotated at 4°C for 10 min. The beads were then washed, boiled, and analyzed by SDS-PAGE and phosphorimaging. Phosphorylation of PhLP in intact HEK 293 cells was done essentially as described with minor modifications (18). 48 h after transfection and 36 h after seeding in six-well plates, HEK 293 cells were washed in phosphateand serum-free DMEM supplemented with 20 mM Na ϩ -HEPES, pH 7.4, (labeling buffer) and then incubated in the same buffer for 30 min. Labeling of the intracellular ATP pool was then performed with 200 Ci/ml [ 32 P]orthophosphoric acid (specific activity 10 mCi/ml) for 2 h at 30°C. Where indicated, cells were stimulated with 1 mM carbachol for 5 min. Cells were immediately placed on ice and washed once with ice-cold buffer. The cells then were lysed on ice with 1 ml of lysis buffer (300 mM NaCl, 50 mM NaF, 5 mM Na 4 P 2 O 7 , 5 mM EDTA-Na 2 , 0.1 mM Na 3 VO 4 , 1% Triton X-100, 0.01% NaN 3 , 50 mM Tris-HCl, pH 7.2, freshly supplemented with 10 mM iodoacetamide and 1 mM PMSF). After 20 min, the Triton X-100-insoluble fractions were removed by centrifugation (20,000 ϫ g for 15 min) and PhLP was immunoprecipitated from the supernatants by the anti-PhLP S antibody precoupled to protein A-Sepharose for 2 h at 4°C. After washing, immunoprecipitates were subjected to SDS-PAGE, and radiolabeled PhLP was visualized and quantified by phosphorimaging. Expression of PhLP was controlled by Western blotting of an aliquot obtained from the centrifuged lysates (with the use of the anti-PhLP-CT antibody).
Tissue Preparation and Dephosphorylation-Mouse heart, brain, adrenal gland, and embryo (day 10.5 post-conception) were homogenized in five volumes of the same lysis buffer as described above (containing phosphatase inhibitors). After centrifugation and adjustment of total protein levels in supernatants as indicated, samples were subjected to SDS-PAGE and Western blotting. In the case of dephosphorylation experiments, lysis was performed in 40 mM Tris base, pH 10.5, 1% Triton X-100, and 1 mM PMSF (phosphatase inhibitors were omitted), samples were cleared by centrifugation (20,000 ϫ g for 10 min), and protein content was determined by the Bradford method. Dephosporylation was then performed by adding 50 g of HEK cell lysate or 100 g of brain lysate to a 50-l reaction with phosphatase buffer (50 mM Tris-HCl, pH 7.5, 2 mM MnCl 2 , 0.1 mM EDTA, 5 mM dithiothreitol, 0.01% Brij 35) and with or without 1000 units of -protein phosphatase (-PPase) (New England Biolabs) at 37°C for 20 min. The reaction was terminated by boiling in Laemmli buffer followed by SDS-PAGE and Western blotting.

Inhibition of Inositol Phosphate Signaling in Intact HEK 293
Cells-To investigate the functional role of PhLP L and PhLP S as G␤␥-regulators in living cells, we subcloned the appropriate cDNAs into the pcDNA3 expression vector and determined the inositol phosphate formation stimulated by the transient cotransfection of phospholipase C␤ 2 (PLC␤ 2 ) and G-protein subunits G␤ 1 and G␥ 2 as described previously (19,20). Without addition of any G␤␥-inhibiting proteins, the transfection of PLC␤ 2 and G␤ 1␥2 cDNA (3 g each) enhanced basal inositol phosphate accumulation ϳ5-fold compared with mock-transfected cells (data not shown). Co-transfection of the cDNA encoding PhLP L inhibited the G␤␥-stimulated production of inositol phosphates significantly (Fig. 1A). Unexpectedly, the short splice variant PhLP S turned out to be more effective than the long splice variant PhLP L , although PhLP S interacts with purified G␤␥ less efficiently than PhLP L (5). Co-transfection of 8 g of cDNA for each G␤␥-binding protein inhibited the total inositol phosphate formation by 32.2 Ϯ 8.4% (n ϭ 13) for PhLP L , 78.3 Ϯ 3.8% (n ϭ 7) for PhLP S , and 75.0 Ϯ 4.0% (n ϭ 3) for GRK2-K220R, which served as positive control (21). Because inositol phosphate levels of untransfected control cells were ϳ20% of the inositol phosphate levels of PLC␤ 2 , G␤ 1␥2 co-expressing cells (data not shown), inhibition by ϳ80%, can be considered full inhibition. Thus, inhibition by PhLP S was almost complete. The differences between control and PhLP L and between PhLP L and PhLP S were highly significant (p Ͻ 0.01). The different effects of PhLP L and PhLP S were not because of different protein expression as shown by Western blot analysis of the transfected cells (Fig. 1B). The blot with samples from three independent experiments shows equal expression of PhLP L and PhLP S within individual experiments and between different experiments.
To analyze further the different effects of PhLP L and PhLP S on G␤␥-function, dose-response experiments were performed. The cDNA of PhLP L or PhLP S was diluted with empty vector to maintain the total amount of transfected cDNA, and inositol phosphates were determined as before. As shown in Fig. 2A, the IC 50 of PhLP L was 20-fold higher than that of PhLP S (13.2 and 0.64 g of cDNA, respectively). Fig. 2, B and C, shows that the transfection of HEK 293 cells with increasing amounts of cDNA did indeed result in the expression of corresponding amounts of PhLP L or PhLP S , respectively.

Phosphorylation of PhLP L and PhLP S in HEK 293 Cells-To
analyze the role of phosphorylation in intact cells as a possible regulatory mechanism, we labeled HEK 293 cells with [ 32 P]orthophosphate and looked for basal-and receptor-stimulated phosphorylation. PhLP cDNAs were co-transfected with M 3 muscarinic receptor cDNA. The M 3 muscarinic receptor was chosen because of the prominent inositol phosphate signal that can be achieved by its stimulation (data not shown). 48 h after transfection, the receptors were stimulated with carbachol for 5 min (Fig. 3). PhLP L exhibited a markedly higher degree of basal phosphorylation compared with PhLP S . However, in both cases, stimulation of the M 3 receptor by carbachol did not enhance the extent of phosphorylation (Fig. 3A, left panel). Fig.  3B summarizes a series of similar experiments demonstrating that the basal phosphorylation of PhLP L was ϳ8-fold higher than that of PhLP S , but that carbachol did not significantly stimulate the phosphorylation of PhLP L or of PhLP S . Expression was controlled by Western blotting with the PhLP-CT antibody and showed that both proteins were equally expressed (Fig. 3A, right panel). Overexpression of PhLP S did not affect There is marked basal phosphorylation in PhLP L , whereas in PhLP S , phosphorylation is only low in HEK 293 cells. Expression levels of both constructs were comparable. B, shown is the 32 P incorporation in PhLP L and PhLP S with and without M 3 muscarinic receptor stimulation by carbachol (means Ϯ S.E.) of three to five independent experiments as the fold of basal PhLP L phosphorylation. Stimulation of PhLP L or PhLP S phosphorylation by carbachol treatment did not result in a significant increase. However, there was a 8-fold difference in the phosphorylation between basal PhLP L and basal PhLP S (***, p Ͻ 0.001). Data were determined by phosphorimaging as described under "Experimental Procedures." the phosphorylation state of PhLP L (data not shown). These findings demonstrate that PhLP L was constitutively phosphorylated in HEK 293 cells, whereas PhLP S only exhibited very low levels of phosphorylation.
Identification of CK2 as the Kinase Responsible for PhLP L Phosphorylation in HEK 293 Cells-To identify the responsible kinase for PhLP L phosphorylation, a computer analysis (Scan-Prosite) of the sequence of the N-terminal 83 amino acids of PhLP L revealed seven putative CK2 phosphorylation sites each within a classical CK2 phosphorylation motif ((S/T)-X-X-(D/E)) (see (Fig. 5A)). To test whether PhLP L could be a substrate of CK2 phosphorylation in vitro, we performed kinase assays with human recombinant CK2 and found that recombinant PhLP L was indeed a substrate, whereas PhLP S and phosducin were not (Fig. 4A). We used the potent but unspecific kinase inhibitor staurosporine (22) and the intracellular Ca 2ϩ chelator BAPTA-AM (23,24) to test their effects on recombinant CK2 and found no effect on PhLP L phosphorylation. On the other hand, we tested two substances that were known to inhibit CK2: the glycosaminoglycan heparin (EC 50 ϳ3 nM) (12,25) and the nucleoside derivative DRB (EC 50 ϳ7 M) (26,27). In the in vitro kinase assay, 3 M heparin completely abolished PhLP L phosphorylation by CK2 and 100 M DRB markedly inhibited PhLP L phosphorylation. To determine whether CK2 is the predominant PhLP L kinase in HEK 293 cells, we performed kinase assays using HEK cell lysate and recombinant PhLP L with and without the addition of kinase inhibitors and recombinant PhLP S as well as phosducin (Fig. 4B). Previous studies have shown that CK2 is abundantly present in mammalian cell lysates (12,28). Here we demonstrate that PhLP L but not PhLP S or phosducin was predominantly phosphorylated by HEK cell lysate. This phosphorylation was inhibited by heparin (3 M) and DRB (100 M), whereas staurosporine only weakly inhibited and BAPTA-AM weakly stimulated the phosphorylation of PhLP L . These results provide strong evidence that CK2 is the predominant kinase in HEK 293 cells that constitutively phosphorylates PhLP L . We then asked whether PhLP was also a substrate for PKA phosphorylation in vitro similar to PKAdependent phosphorylation of the homologous protein phosducin. PKA kinase assays were done with the same concentration (400 nM) of purified recombinant phosducin, PhLP L , and PhLP S (Fig. 4C). Phosducin was an excellent PKA substrate, whereas both PhLP variants were not. They roughly showed 10 -20-fold less radioactivity incorporated than phosducin. These results demonstrate that the different members of the phosducin family exhibit differential regulation by different kinases.
Identification of a Small Regulatory Region in the PhLP L N Terminus-We then asked whether any particular region of the PhLP L N terminus was responsible for the weaker functional effects of PhLP L compared with PhLP S as well as for the constitutive phosphorylation in HEK 293 cells. As depicted in Fig. 5A, the N terminus of PhLP L contains several putative phosphorylation sites that occur in clusters of serines and threonines. Seven of these sites could serve as CK2 phosphorylation sites. We constructed PhLP L mutants by stepwise shortening of the N terminus. The five PhLP L constructs were as follows: L5 (amino acids 5-301 with the loss of Thr-2 and Thr-3), L29 (amino acids 29 -301 with the additional loss of Ser-18, Thr-19, Ser-20 and Ser-25), L36 (amino acids 36 -301), L46 (amino acids 46 -301 with the additional loss of Ser-39, Ser-40, Ser-41, and Thr-42), and L58 (amino acids 58 -301 with the additional loss of Ser-54 and Thr-57). The determination of G␤␥-dependent inositol phosphate generation showed that a region within amino acids 5-28 caused the only major step in the gain of inhibitory function from PhLP L to PhLP S as demonstrated by the functional difference between L5 and L29 (Fig.   5A, lower panel). The removal of the first 28 amino acids of PhLP L (as in L29, L36, L46, and L58) was sufficient to gain the full G␤␥-inhibitory activity of PhLP S . Western blots showed that all of the constructs with the exception of L5 were stable and that the expression levels of the constructs were comparable (Fig. 5B). , and phosducin (11.62 g/ml) purified from E. coli were phosphorylated by either recombinant human casein kinase 2 or by HEK cell lysate or purified catalytic subunit of PKA. A, recombinant human CK2 (0.04 milliunit/ reaction) was used to phosphorylate recombinant PhLP L , PhLP S , and phosducin at 30°C for 30 min. After SDS-PAGE and Coomassie Blue staining (Coomassie), the gel was subjected to phosphorimaging ( 32 P). CK2 phosphorylated PhLP L efficiently, whereas PhLP S was only phosphorylated weakly (21% PhLP L ) and phosducin was not (2% PhLP L ). The concentration of the inhibitors used was 1 M for staurosporine (Stauro), 100 M for BAPTA-AM (BAPTA), 3 M for heparin, and 100 M for DRB. Me 2 SO (DMSO) (1% v/v) served as solvent control. B, HEK 293 cell lysate extracted as detailed under "Experimental Procedures" was used to phosphorylate recombinant PhLP L in the absence or presence of kinase inhibitors as well as PhLP S and phosducin at 30°C for 30 min. Recombinant proteins were pulled down by Ni-NTA-agarose and processed as described under "Experimental Procedures." Control denotes HEK cell lysate without any recombinant protein and reflects background phosphorylation. Shown is the gel as visualized by phosphorimaging ( 32 P) and by Coomassie Blue staining (Coomassie). PhLP L phosphorylation was inhibited by heparin and DRB (both known to inhibit CK2) but not by staurosporine or BAPTA, whereas PhLP S or phosducin was only weakly phosphorylated. C, recombinant phosducin, PhLP L , and PhLP S were phosphorylated by the catalytic subunit of PKA (100 units) at 30°C for 15 min. Reactions were stopped by Laemmli buffer, boiled, and separated on 12.5% SDS-PAGE. Shown is the film ( 32 P) and the corresponding gel (Coomassie) to demonstrate that PKA phosphorylation of PhLP isoforms is not significant compared with phosducin.

Effect of Mutation of Serine 18, Threonine 19, and Serine 20
to Alanines-We next asked whether the putative CK2 phosphorylation sites of the region compromising amino acids 5-28 might play a functional role in PhLP L regulation, and if so, which of the four candidates (Ser-18, Thr-19, Ser-20, and Ser-25) was involved. Two mutants were constructed, PhLP L A18 -20 in which Ser-18, Thr-19, and Ser-20 were changed to alanines, and PhLP L A25 in which Ser-25 was changed to an alanine. These constructs were tested for their ability to inhibit G␤␥-dependent inositol phosphate generation. PhLP L A18 -20 showed the same G␤␥-inhibitory capability as PhLP S and L29 (Fig. 6A), whereas PhLP L A25 exhibited the same functional capability as PhLP L A25. The differences be-tween the effects of PhLP L and PhLP L A25 on the one hand and PhLP S , L29, and PhLP L A18 -20 on the other hand were highly significant (Fig. 6A). Again, equal expression of the constructs was demonstrated in Western blot experiments (Fig. 6B). These data suggest that the cluster of Ser-18, Thr-19, and Ser-20 contains the phosphorylation sites responsible for the diminished G␤␥-inhibitory effects of PhLP L , whereas Ser-25 most probably was not involved. Overexpression of PhLP L in a 4-fold excess over PhLP S or PhLP L A18 -20 reduced the inhibition of the latter to the level exerted by PhLP L alone (Fig. 6C), suggesting that PhLP L competed with PhLP S for inhibiting G␤␥. A 4-fold overexpression of PhLP L over PhLP S was controlled by Western blots (data not shown).
In Vivo Relevance of PhLP Splicing and Phosphorylation-Finally, we asked whether the observed phenomenon of PhLP regulation via N-terminal phosphorylation or splicing might play a role in different mouse organs, which had been described to contain considerable amounts of PhLP L (heart and brain) (4, FIG. 5. Progressive truncation of the PhLP L N terminus revealed a short regulatory region between amino acids 5 and 28. A, the extended N terminus of PhLP L contains several clusters of putative phosphorylation sites (S, serine; T, threonine) as depicted in the upper part of the panel. Arrows indicate the possible CK2 phosphorylation sites, whereas the numbers over the arrowheads indicate the amino acids C-terminal of the cutting point of the truncated constructs of PhLP L . The removal of the first 28 amino acids of PhLP L (L29, L29) was sufficient to gain the same G␤␥-inhibitory effect as PhLP S , whereas removing the first four amino acids (L5) was not. Additional progressive truncation did not change this effect any further. The inhibition of the inositol phosphate signal was by 42 Ϯ 11% (n ϭ 5) for PhLP L , 52 Ϯ 9% (n ϭ 5) for L5, 84 Ϯ 1% (n ϭ 3) for L29, 85 Ϯ 2% (n ϭ 3) for L36, 84 Ϯ 1% (n ϭ 3) for L46, 85 Ϯ 2% (n ϭ 5) for L58, and 84 Ϯ 3% (n ϭ 5) for PhLP S (**, p Ͻ 0.01; ***, p Ͻ 0.001 versus PhLP L ). B, expression control of the five N-terminally truncated constructs in comparison to that of PhLP L and PhLP S . Transfections were done as in inositol phosphate measurements, and 20% of the lysate of one well of a 6-well plate/lane was used. Shown is one of two Western blots with similar results. Used was the polyclonal PhLP-CT antibody as described under "Experimental Procedures." 5) and which might also contain PhLP S (adrenal gland) (6). In addition, embryos were chosen because CK2 plays an essential role in developing cells and cell cycle progression (13). Therefore, we expected to detect a high level of phospho-PhLP L . PhLP L was detected by Western blotting in all three organs as well as in day 10.5 embryo and also in control (empty vector) transfected HEK 293 cells (Fig. 7A). In contrast, PhLP S was present in clearly detectable amounts only in the adrenal gland and in HEK 293 cells after transfection of the PhLP S cDNA (Fig. 7A). After longer exposure of the film, PhLP S seemed to be present also in heart and brain but not in HEK 293 cells or embryo (data not shown). In most instances, PhLP L ran as a doublet, one band corresponding to the position of recombinant PhLP L purified from E. coli (Fig. 7, A and B, lower arrowhead) and one to the position of cellular PhLP L (Fig. 7, A and B, upper  arrowhead). Because the presence of phosphate moieties is known to change the gel mobility, we wondered whether the mutation in Ser-18, Thr-19, and Ser-20 (PhLP L A18 -20) would lead to a similar gel mobility shift (Fig. 7B). Indeed PhLP L A18 -20 corresponded to the lower band in brain and adrenal gland as well as recombinant PhLP L purified from E. coli (Fig. 7B). HEK 293 cell lysate and mouse brain lysate were then treated with -PPase, which has been described to efficiently remove phosphates from serine, threonine, and tyrosine residues (29). This resulted in the loss of the PhLP L high molecular mass form as detected by Western blotting, whereas the low molecular mass form of PhLP L became detectable in HEK 293 cells and was enhanced in mouse brain (Fig. 7C). The contention that the upper band represents phosphorylated PhLP L was supported by the observation that phosphorylation of PhLP L by CK2 in vitro caused a reduction of gel mobility (Fig. 7D). Together, these data demonstrate that there is constitutive phosphorylation of PhLP L in different mouse tissues and that in brain and adrenal gland non-phosphorylated PhLP L seems to play a role. DISCUSSION G-protein function is regulated by different classes of proteins, most notably the RGS proteins and the members of the phosducin family. The members of the phosducin family bind to the G␤␥-subunits (30,31) and therefore inhibit G-protein function. This might result in the disruption of the G-protein cycle, (32,33) and/or in the inhibition of G␤␥-mediated effects like activation of PLC␤ isoforms and G-protein-coupled receptor kinase 2 (9,34,35). In this work, we used the activation of a G␤␥-dependent phospholipase C␤ 2 in HEK 293 cells as a functional readout of G␤␥-inhibition to demonstrate that PhLP inhibits G␤␥-functions in living cells. According to present knowledge, PhLP exists in two isoforms: 1) the long variant PhLP L with an additional N terminus of 83 amino acids and 2) the short variant PhLP S , which lacks this N terminus (Figs. 5A and 8) (1). However, the functional differences of these two isotypes in living cells remained to be elucidated, because (a) in vitro experiments showed that the short form, which lacks one G␤␥-binding region according to structural data (Fig. 8) (31), exhibited a markedly reduced binding affinity toward G␤␥subunits compared with PhLP L and phosducin (2,5) and (b) tissue levels of the PhLP isoforms appropriate for a physiological role in G␤␥-inhibition could only be demonstrated for the long form, PhLP L (4,5). This finding appeared to suggest that the physiological G␤␥-regulator should be the long form, PhLP L . Our present data show that in contrast to this hypothesis, PhLP S is an effective regulator of G␤␥-function in HEK 293 cells. The yeast phosducin homolog resembles PhLP S in size and domain structure (Fig. 8), suggesting that important G␤␥-regulating properties are contained in these proteins. Furthermore, our data suggest that the different forms of PhLP cooperate in inhibiting G␤␥ to a variable extent in that the phosphorylated form of PhLP L appears to have a partial dominant negative effect on the inhibitory function of PhLP S . PhLP S was found in mouse adrenal gland in amounts comparable with the levels of PhLP L in mouse brain and heart. In total protein of mouse heart, brain, adrenal gland, and embryo. Preparation of two different mice was performed in the presence of phosphatase inhibitors as detailed under "Experimental Procedures" and showed the same result. Adrenal gland has a prominent band of the size of PhLP S . Brain and adrenal gland show two bands: the upper one corresponding to the size of PhLP L in HEK 293 cells (transfected and endogenous) and the lower one corresponding to the size of PhLP L purified from E. coli. The PhLP S antibody has considerable cross-reactivity to the endogenous phosducin, which has an apparent gel size of 33 kDa. B, close-up view of a Western blot of 100 g of total protein of mouse tissue (brain, adrenal gland, and embryo), 10 g of total protein of HEK 293 cell lysate (PhLP L and PhLP L A18 -20), and PhLP L (0.1 pmol) purified from E. coli. In brain and adrenal gland, doublets are detectable, whereas in embryo, only a single band is visible. Mutation of the phosphorylation sites Ser-18, Thr-19, and Ser-20 to Ala caused a shift in mobility of PhLP L in SDS-PAGE from that of wild-type PhLP L -transfected HEK 293 cells to that of unphosphorylated recombinant (rec.) PhLP L and corresponded to the size of the lower band in mouse brain and adrenal gland. C, HEK cell lysate (50 g of total protein) and mouse brain lysate (100 g of total protein) were prepared as described under "Experimental Procedures" and subjected to dephosphorylation by 1000 units of -PPase at 37°C for 20 min. Controls were incubated without the phosphatase. In native probes, PhLP L could be dephosphorylated by -PPase and changed its gel mobility in a 15% SDS-PAGE. D, purified recombinant PhLP L (400 nM) was phosphorylated by CK2 (0.3 milliunit) in the absence or presence of 30 M heparin at 37°C for 120 min. SDS-PAGE was performed on 15% gel and subjected to phosphorimaging. Phosphorylation of recombinant PhLP L by CK2 caused a gel mobility shift similar to that seen in native tissue. the latter organs, tissue levels of PhLP L have been calculated to be sufficient for in vivo regulation of G-proteins (5). The detection of PhLP S in adrenal glands is in line with its purification from bovine chromaffin cells (6). These findings underline the contention that not only PhLP L but also PhLP S is a relevant physiological regulator of G-protein function.
In the case of phosducin, it has been shown that G-protein inhibition is subject to regulation by phosphorylation by several kinases, most notably PKA (7,8,10,11). A PKA consensus motif is also present in PhLP (amino acids 117-121 in PhLP L : GKMT*L) 2 but does not appear to serve as a PKA site in vitro or in vivo. However, in HEK 293 cells as well as in mouse organs, PhLP L is apparently constitutively phosphorylated in its N terminus at Ser-18, Thr-19, and Ser-20. CK2 is an ubiquitously expressed and constitutively active serine-threonine kinase (12,13), which has recently been found to constitutively phosphorylate arrestin-3, another protein of the G-protein-coupled receptor signaling pathway (36). Here, we report that recombinant PhLP L but not PhLP S or phosducin were substrates of CK2 phosphorylation, whereas PKA phosphorylated only phosducin. Therefore, we conclude that the different members of the phosducin family are differentially regulated by different kinases. Further evidence that CK2 is a regulatory kinase for PhLP L comes from the findings that (a) Ser-18, Thr-19, and Ser-20 contain the consensus motif for CK2, (b) mutation of the consensus sites into alanines resulted in a non-phosphorylated form in intact cells, (c) progressive N-terminal truncation or serine/threonine to alanine mutation at these sites both converted PhLP L into a protein with the same potency and efficacy as PhLP S , and (d) phosphorylation of PhLP L by kinase activity from HEK cell lysates was inhibited by two different CK2-specific inhibitors, whereas other kinase inhibitors did not result in inhibition.
Taken together, these findings indicate a regulatory function of the N terminus of PhLP L and underline the role of Nterminal regulation of the members of the phosducin family. Our data also show that PhLP isoforms differ in the mode of activity control and that the G␤␥-inhibitory effects of PhLP L are subject to regulation via the N-terminal domain. Phosphorylation of the N terminus, in vivo most probably by CK2, diminishes the ability of PhLP L to inhibit G␤␥-functions. The analysis of PhLP expression in mouse organs shows that both mechanisms are used in vivo.