Phosphorylation of Serine 1105 by Protein Kinase A Inhibits Phospholipase Cβ3 Stimulation by Gαq *

The mechanism by which protein kinase A (PKA) inhibits Gαq-stimulated phospholipase C activity of the β subclass (PLCβ) is unknown. We present evidence that phosphorylation of PLCβ3 by PKA results in inhibition of Gαq-stimulated PLCβ3 activity, and we identify the site of phosphorylation. Two-dimensional phosphoamino acid analysis of in vitro phosphorylated PLCβ3revealed a single phosphoserine as the putative PKA site, and peptide mapping yielded one phosphopeptide. The residue was identified as Ser1105 by direct sequencing of reverse-phase high pressure liquid chromatography-isolated phosphopeptide and by site-directed mutagenesis. Overexpression of Gαq with PLCβ3 or PLCβ3 (Ser1105 → Ala) mutant in COSM6 cells resulted in a 5-fold increase in [3H]phosphatidylinositol 1,4,5-trisphosphate formation compared with expression of Gαq, PLCβ3, or PLCβ3 (Ser1105 → Ala) mutant alone. Whereas Gαq-stimulated PLCβ3 activity was inhibited by 58–71% by overexpression of PKA catalytic subunit, Gαq-stimulated PLCβ3 (Ser1105→ Ala) mutant activity was not affected. Furthermore, phosphatidylinositide turnover stimulated by presumably Gαq-coupled M1 muscarinic and oxytocin receptors was completely inhibited by pretreating cells with 8-[4-chlorophenythio]-cAMP in RBL-2H3 cells expressing only PLCβ3. These data establish that direct phosphorylation by PKA of Ser1105 in the putative G-box of PLCβ3 inhibits Gαq-stimulated PLCβ3 activity. This can at least partially explain the inhibitory effect of PKA on Gαq-stimulated phosphatidylinositide turnover observed in a variety of cells and tissues.

Ligand stimulation of seven transmembrane domain receptors coupled to G␣ proteins of the G␣ q or G␣ i subfamilies results in the activation of the respective heterotrimeric G␣␤␥ protein complexes. Free G␣ q or G␤␥ subunits activate PLC␤ 1 isoforms to catalyze the production of IP 3 and diacylglycerol from phosphatidylinositide 4,5-bisphosphate (1)(2)(3). PLC␤ [1][2][3][4] comprise the currently known mammalian phosphatidylinositide-specific PLC␤ subfamily. Although all PLC␤s are activated by G␣ q , PLC␤ 2 and PLC␤ 3 are also stimulated by G␤␥, primarily released from G␣ i (1).
Cross-talk between the G protein-PLC␤ pathway and PKA has been documented in numerous studies (4 -13). Although it is generally agreed that G protein-activated PLC␤ activity can be inhibited by PKA (4 -11), PKA can enhance the G protein-PLC␤ pathway in some cases (12,13). Because PKA can inhibit phosphatidylinositide (PI) turnover activated by both G␣ q (4 -8) and G␣ i (9 -11) coupled receptors, it may inhibit the stimulation of both G␣ q -and G␤␥-stimulated PLC␤ activity. This notion is further supported by studies with the G protein activators GTP␥S and AlF 4 Ϫ . These two compounds nonselectively activate all heterotrimeric G proteins and generate free G␣ and G␤␥ subunits that can stimulate PLC␤s. PKA inhibition of PI turnover initiated by GTP␥S or AlF 4 Ϫ (5,8,14,15) is consistent with the inhibition of G␣q-as well as G␤␥-stimulated PLC␤ activity. In addition, this phenomenon also suggests that the PKA effect is distal to receptors.
Recently, the mechanism for PKA inhibition of G␤␥-stimulated PI turnover has been elucidated. Phosphorylation of PLC␤ 2 by PKA resulted in inhibition of G␤␥-stimulated PI turnover (10). However, in the same study, PKA apparently did not inhibit G␣ 15 -and G␣ 16 -stimulated endogenous PLC␤ (␤ 1 and ␤ 3 ) activity. More recently, Ali et al. (11) have reported phosphorylation of PLC␤ 3 in response to CPT-cAMP treatment in RBL-2H3 cells expressing only PLC␤ 3 . CPT-cAMP inhibited G␤␥-stimulated PLC␤ 3 activated by the G␣ i -coupled formylmethionylleucylphenylalanine receptor but had no effect on PAFstimulated PLC␤ 3 activity, presumably mediated by G␣ q . These studies led to the conclusions that phosphorylation of PLC␤ 2 and PLC␤ 3 by PKA could explain the inhibition of G␤␥-stimulated PI turnover by cAMP (10,11). However, a biochemical mechanism for the inhibition by PKA of G␣ q -stimulated PLC␤ activity observed in several systems remains to be clarified. In this study, we present evidence that phosphorylation of PLC␤ 3 Ser 1105 by PKA results in direct inhibition of G␣ q -stimulated PLC␤ 3 activity. Cloning, Site-directed Mutagenesis, and Protein Purification-PLC␤ 3 and PLC␤ 3 (His) 6 in pCR3.1 vector (Invitrogen, San Diego, CA) and PLC␤ 3 (His) 6 in baculovirus (Pharmingen, San Diego, CA) were constructed from the PLC␤ 3 cDNA plasmid (17). Site-directed mutation of Ser 1105 to Ala was achieved with the mutagenic primer (5Ј-AGCGC-CATAACGCCATCTCGGAGG-3Ј) using the GeneEditor kit (Promega, Madison, WI). All plasmid sequences were confirmed by DNA sequencing. PLC␤ 3 (His) 6 was purified essentially as described for PLC␤ 1 (18) from the membrane fraction from Sf9 cells and was 99% pure as judged by SDS-PAGE.

Materials-PLC␤
In Vitro Phosphorylation, Phosphoamino Acid Analysis, Peptide Mapping, and Sequencing-0.5, 1.5, or 2.5 M purified recombinant PLC␤ 3 (His) 6 was incubated with PKA catalytic subunit at molar ratios of 20:1 or 50:1 in the presence of 1-10 Ci of [␥-32 P]ATP and 100 M ATP in a total volume of 10 l of PKA buffer (10 mM Tris, pH 7.0, 5 mM MgCl 2 ) for 10 min at 30°C. For the time course study, 1.3 M PLC␤ 3 (His) 6 was incubated with PKA at a ratio of 10:1. Reactions were terminated by addition of an equal volume of 2ϫ SDS sample buffer (15) and boiling for 5 min. Proteins were separated on SDS-PAGE gels, and the phosphorylated band was localized by autoradiography.
Two-dimensional phosphoamino acid analysis and peptide mapping of in vitro 32 P-labeled PLC␤ 3 (His) 6 bound to PVDF membranes were carried out with a Hunter thin layer electrophoresis system (CBS Scientific Company, Del Mar, CA) according to the manufacturer's instructions. For two-dimensional peptide mapping, the membrane bound samples were digested with Lys-C (3 g) for 24 h at 35°C. For peptide sequencing, 150 pmol of [ 32 P]PLC␤ 3 (His) 6 was digested with Lys-C (1 g). The phosphopeptide separated by reverse-phase HPLC was sequenced at the microsequencing facility at Baylor College of Medicine (Houston, TX).
In Vivo 32 P Labeling and Immunoprecipitation-Nearly confluent PHM1-41 immortalized myometrial cells (10-cm dish) were labeled with [ 32 P]orthophosphate (0.33 mCi/ml) in phosphate-free DMEM containing 10% dialyzed fetal calf serum for 4 h. After the treatments indicated in the figure legends, cells were lysed in 1 ml of ice-cold lysis buffer containing a mixture of protease and phosphatase inhibitors (11) and centrifuged at 15,000 ϫ g for 5 min at 4°C. Phosphorylated proteins immunoprecipitated with 4 g of anti-PLC␤ 3 antibody were separated on a 7.5% SDS-PAGE gel, transferred to a PVDF membrane, and analyzed by autoradiography. PLC␤ 3 was visualized by Western blot using anti-PLC␤ 3 antibody (1:1000) to normalize for sample loading.
Cell Culture, Transfection, and PI Turnover-COSM6 and RBL-2H3 cells were cultured and transfected as described (4,11) with the following modifications. COSM6 cells were transfected with a total of 1.25 (see Fig. 4A) or 1.5 g (see Fig. 4B) of plasmid DNA (using empty vector rcCMV as necessary) and 6 l of LipofectAMINE in 0.75 ml of DMEM/ well in 6-well plates, whereas 1.0 g of total plasmid DNA and 5 l of LipofectAMINE in 0.5 ml of DMEM were used to transfect RBL-2H3 cells. An equal volume of culture medium (4) containing 16% fetal calf serum was added 5 h later. The following day, cells were labeled with 6 Ci/well [ 3 H]inositol in 1 ml of culture medium for 24 h at 37°C. ZnSO 4 (60 M) was also included in the labeling medium to stimulate PKA catalytic subunit expression in COSM6 cells. After incubating with 10 mM LiCl for 10 (RBL-2H3) or 45 (COSM6) min, cells were treated as indicated in the figure legends and lysed by addition of ice-cold 10% trichloroacetic acid. The accumulation of [ 3 H]IP 3 was determined as described elsewhere (4).

RESULTS AND DISCUSSION
In Vitro and in Vivo Phosphorylation of PLC␤ 3 by PKA-We have determined previously that the PKA inhibitory effect is distal to receptor and most likely affects the coupling between G␣ q and PLC␤ 1 or PLC␤ 3 isoforms in pregnant human myometrial (PHM1-41) and COSM6 cell lines (4). As shown in Fig.  1A, when incubated with PKA, C-terminal (His) 6 -tagged PLC␤ 3 (PLC␤ 3 (His) 6 ) purified from Sf9 cells was clearly a substrate for PKA in vitro. Similar results were also obtained with immunoprecipitation-purified recombinant PLC␤ 3 (data not shown). The phosphorylation of PLC␤ 3 was quite specific; neither highly purified recombinant G␣ q nor recombinant PLC␤ 1 was phosphorylated by PKA under similar conditions (data not shown), confirming previous observations (19,20) To quantify PKA-stimulated 32 P incorporation from [␥-32 P]ATP, PLC␤ 3 (His) 6 was phosphorylated by PKA in vitro. Fig. 1C shows that the time course of 32 P incorporation approached a plateau after 15 min. A maximum ratio of 0.65 mol phosphate/mol PLC␤ 3 was determined by filter binding assay at the 60-min incubation point. This is consistent with a single PKA phosphorylation site in PLC␤ 3 .
To examine whether PLC␤ 3 could be phosphorylated in vivo, PHM1-41 myometrial cells were labeled with [ 32 P]orthophosphate, and PKA was activated with the cell-permeable cAMP analogue CPT-cAMP or relaxin, a hormone that increases myometrial cell cAMP (21). Fig. 1B shows that PLC␤ 3 immunoprecipitated from cells exposed to CPT-cAMP or relaxin exhibited increased phosphorylation. After normalizing for the amount of PLC␤ 3 loaded, the treatments resulted in a 2-fold increase in PLC␤ 3 phosphorylation. A similar fold increase in PLC␤ 3 phosphorylation was recently reported in RBL-2H3 cells treated with CPT-cAMP (22). These data indicate that endogenous PLC␤ 3 can be phosphorylated in cells in response to elevated cAMP.
Identification of the PKA Phosphorylation Site-Two-dimensional phosphoamino acid analysis with in vitro phosphorylated PLC␤ 3 (His) 6 revealed that only serine was phosphorylated by PKA ( Fig. 2A). A similar result was also obtained with non-His-tagged recombinant PLC␤ 3 phosphorylated by PKA in COSM6 cells (data not shown). Peptide mapping of in vitro phosphorylated PLC␤ 3 (His) 6 digested with Lys-C revealed one major phosphorylated peptide (Fig. 2B). Importantly, an increase in phosphorylation of the same peptide (indicated by the arrow) was also detected in endogenous PLC␤ 3 in PHM1 (Fig.  2, D versus C) and COSM6 cells (Fig. 2, F versus E) treated with CPT-cAMP as well as in overexpressed PLC␤ 3 in COSM6 cells coexpressing PKA catalytic subunit (Fig. 2, H versus G). In the case of overexpressed PLC and PKA (Panels G, H), there appears to be phosphorylation of another site in the basal state that decreases when the PKA site is phosphorylated. We are in  6 . The inset shows the PLC␤ 3 phosphorylation autoradiographs (Auto) and Coomassie Blue staining (Stain) at the indicated times. The densitometric analysis at each time point after normalizing for PLC␤ 3 loading is shown in the plot. the process of determining the residue phosphorylated and the functional significance of this site. Importantly, in all three cases, activation of PKA increases phosphorylation on the site that is phosphorylated in vitro by PKA.
To identify the sequence of the phosphopeptide, PLC␤ 3 (His) 6 was phosphorylated in vitro and subjected to Lys-C digestion. A fraction containing more than 60% of the incorporated 32 P was isolated by reverse-phase HPLC and sequenced. The 32 P-labeled peptide was identified as RHNS 1105 ISEAK (Fig. 3A), in which more than 80% of Ser 1105 was labeled, and no label was present in Ser 1107 . ϳ30% of the 32 P found in the HPLC flowthrough appeared to be free phosphate, as judged by phosphoamino acid analysis (data not shown).
To confirm the phosphorylation site and avoid contamination with endogenous PLC␤ 3 , His-tagged PLC␤ 3 (Ser 1105 3 Ala) mutant was constructed. This mutant was overexpressed in COSM6 cells, purified on a Ni-NTA column and phosphorylated by PKA in vitro. As shown in Fig. 3B, mutation of Ser 1105 to Ala reduced PKA phosphorylation of PLC␤ 3 by ϳ90%. The small residual phosphorylation probably represents background, because it was also seen in extracts from cells transfected with empty vector and processed similarly (data not shown). We conclude from these studies that PKA phosphorylates PLC␤ 3 Ser 1105 both in vivo and in vitro. Notably, this PKA phosphorylation site is not present in the corresponding sequences (20) of PLC␤ 1 or PLC␤ 2 (Fig. 3A).
Inhibition of G␣ q -stimulated PLC␤ 3 Activity by PKA-The C terminus of PLC␤ 1 is critical for activation by G␣ q (23). Deletion studies have identified a P-box (Thr 903 to Gln 1030 ) and a G-box (Lys 1031 to Leu 1142 ) in this region. The P-box is essential for both PLC␤ 1 association with the cell membrane and its activation by G␣ q , whereas the G-box is involved in association with G␣ q subunit (24). Ser 1105 of PLC␤ 3 falls in a region analogous to the G-box of PLC␤ 1 . We therefore hypothesized that phosphorylation of Ser 1105 by PKA might cause interference with G␣ q -PLC␤ 3 association and thereby inhibit G␣ q -stimulated PLC␤ 3 activity. To test this, PI turnover was studied in COSM6 cells transfected with G␣ q and PLC␤ 3 or PLC␤ 3 (Ser 1105 3 Ala) mutant in the absence and presence of PKA. As shown in Fig. 4A, transfection of empty vector (rcCMV), G␣ q , PLC␤ 3 , or PLC␤ 3 (Ser 1105 3 Ala) mutant alone had no effect on basal PI turnover, suggesting that the proteins are primarily in their inactive forms under these conditions (25). Cotransfection of G␣ q with PLC␤ 3 produced a 5-fold increase in [ 3 H]IP 3 , presumably because of the increased activation of PLC␤ 3 by G␣ q as reported previously (26). Notably, the PLC␤ 3 (Ser 1105 3 Ala) mutant was as effective as wild type at stimulating PI turnover. This indicates that the substitution of Ala for Ser 1105 did not have a major effect on catalytic activity or G protein coupling. Importantly, when PKA catalytic subunit was also coexpressed, G␣ q -stimulated [ 3 H]IP 3 formation associated with wild type PLC␤ 3 was inhibited by ϳ58%, whereas no inhibition was observed with the PLC␤ 3 (Ser 1105 3 Ala) mutant. Increas- ing the amount of PKA catalytic subunit plasmid (0.5, 0.65, and 0.75 g) resulted in a trend toward greater inhibition (58, 62, and 71%, respectively), although these values were not statistically different from each other (Fig. 4B). These data demonstrate both that PKA indeed inhibits G␣ q -stimulated PLC␤ 3 activity and that this effect requires the phosphorylation of Ser 1105 .
More evidence in support of this contention was obtained in the RBL-2H3 cell line expressing only PLC␤ 3 (11). RBL-2H3 cells were transfected with the M1 muscarinic and oxytocin receptors shown to couple in other cell types to PLC␤ through G␣ q proteins (26 -29). Stimulation with the respective ligands resulted in a 4-fold increase in [ 3 H]IP 3 , presumably through the coupling of the receptors to endogenous G␣ q and PLC␤ 3 . Importantly, pretreating cells with CPT-cAMP, previously shown to activate endogenous PKA and result in PLC␤ 3 phosphorylation (11), completely inhibited M1 and oxytocin receptor-stimulated PI turnover (Fig. 4C). These data are consistent with our previous data in COSM6 and PHM1-41 cells (4). In myometrial membranes, oxytocin-stimulated PI turnover was determined to be essentially completely G␣ q -mediated (28).
Based on the position of Ser 1105 in the enzyme, we hypothesize that its phosphorylation by PKA may perturb the association of PLC␤ 3 with G␣ q . However, we do not know at present how Ser 1105 phosphorylation affects the kinetic properties of G␣ q /PLC␤ 3 coupling. It is also not yet clear what relationship this phosphorylation has to the reported inhibition of G␤␥stimulated PLC␤ 3 activity by PKA (11). These questions are under study. Interestingly, Ser 954 , one of the two putative PKA phosphorylation sites in PLC␤ 2 , is located in the P-box. It has been suggested that PKA phosphorylation of that site may interfere with the membrane association of PLC␤ 2 (10). The close proximity of Ser 1105 to the P-box may allow PKA to affect the membrane association of PLC␤ 3 as well.
Based on the ubiquitous expression of PLC␤ 3 (20), our observations could explain the inhibition of G␣ q -stimulated PI turnover by PKA observed in a variety of cells and tissues (4 -11). However, the basis for the complete inhibition of G␣ qstimulated PI turnover by PKA in cells expressing both PLC␤ 1 and PLC␤ 3 (4, 15) cannot be adequately addressed without knowing the cellular localization and relative contributions of these two isoenzymes to total PI turnover. Our data apparently contradict the reported inability of PKA to inhibit G␣ q -coupled PAF receptor-stimulated PLC␤ 3 activity in RBL-2H3 cells (11). The reason for this discrepancy is unknown at present but may reflect differences in the nature of specific receptor/G protein coupling in that cell line, differences in experimental design, or some other as yet unknown factor. Consistent with the findings reported here, we have found that coexpression of PKA catalytic subunit inhibits carbachol-stimulated PI turnover in COSM6 cells cotransfected with M1 muscarinic receptor and G␣ q (4). In contrast, G␣ 15 -and G␣ 16 -stimulated endogenous PLC␤ (␤ 1 and ␤ 3 ) activity was not inhibited by PKA in COS7 cells (10). It is unclear how effectively the various G proteins stimulated PLC␤ 3 versus PLC␤ 1 in these cells. In reconstitution assays, G␣ 16 appears to be as effective as G␣ q in stimulating PLC␤ 1 but less effective than G␣ q in stimulating PLC␤ 3 (30). The relative contribution of PLC␤ 3 versus PLC␤ 1 to PI turnover and possible preferential coupling of G␣ q subfamily isoforms to PLC␤ 3 may account for some of the observed differences.
In summary, the data presented here establish a direct relationship between PKA-stimulated phosphorylation of Ser 1105 and inhibition of PLC␤ 3 activity. This can at least partially explain the inhibitory effect of PKA on G␣ q -coupled receptorstimulated PI turnover observed in a variety of cells and tissues.