EXPERIMENTAL and MOLECULAR MEDICINE, Vol. 36, No. 2, 172-178, April 2004 Phosphorylation of phospholipase D1 and the modulation of its interaction with RhoA by cAMP-dependent protein kinase Min-Jung Jang1, Min-Jung Lee1 phorylation of RhoA rather than by the phosphor- Hae-Young Park1, Yoe-Sik Bae1 ylation of PLD1. Do Sik Min2, Sung Ho Ryu3 1,4 Keywords: Phospholipase D; RhoA; cAMP; Protein and Jong-Young Kwak kinase A; U87 cells 1Medical Research Center for Cancer Molecular Therapy and Introduction Department of Biochemistry College of Medicine, Dong-A University Phospholipase D (PLD) catalyzes the hydrolysis of Busan 602-714, Korea phosphatidylcholine to generate free choline and 2 Department of Physiology phosphatidic acid (PA). PA can be further metabolized College of Medicine, The Catholic University of Korea by PA phosphohydrolase to form diacylglycerol and Seoul 137-701, Korea and by phospholipase A2 to form lysophosphatidic acid. 3 Department of Life Science Moreover, PA and its metabolites act as second Pohang University of Science and Technology messengers in the regulation of secretion, mitogene- Pohang 790-784, Korea sis, and cytoskeletal reorganization (Exton, 1999; 4 Corresponding author: Tel, 82-51-240-2928; Liscovitch et al., 2000) Fax, 82-51-241-6940; E-mail, [email protected] PLD is tightly regulated by a variety of hormones, growth factors, cytokines, and other agonists (Singer Accepted 3 March 2004 et al., 1997; Frohman et al., 1999). Two types of PLD have been cloned, PLD1 (~120 kDa) and PLD2 (~105 Abbreviations: dbcAMP, dibutyryl cAMP; dbcGMP, dibutyryl cGMP; kDa), which differ both in activation mechanism and PKA, cAMP-dependent protein kinase; PKG, cGMP-dependent subcellular localization (Hammond et al., 1995; Colley protein kinase; PLD, phospholipase D et al., 1997). PLD1 is constitutively inactive and is activated by ADP-ribosylation factor (ARF), RhoA, and protein kinase C (PKC) (Hammond et al., 1995). In contrast, PLD2 is constitutively active but insensitive Abstract to ARF and RhoA (Colley et al., 1997). Previous Agents that elevate cellular cAMP are known to studies have demonstrated that the effector region inhibit the activation of phospholipase D (PLD). (switch I) of RhoA interacts with the C-terminus of We investigated whether PLD can be phosphor- human PLD1 (Bae et al., 1998; Yamazaki et al., ylated by cAMP-dependent protein kinase (PKA) 1999; Cai and Exton, 2001). Moreover, in vivo studies and PKA-mediated phosphorylation affects the in- have shown that PLD can be activated by the direct teraction between PLD and RhoA, a membrane binding of RhoA (Du et al., 2000; Xie et al., 2002). regulator of PLD. PLD1, but not PLD2 was found The intracellular accumulation of cAMP leads to functional inhibition in various cells. Increased cAMP to be phosphorylated in vivo by the treatment of in neutrophils lead to the subsequent inhibition of dibutyryl cAMP (dbcAMP) and in vitro by PKA. PLD, suggesting that cross-talk is present between PKA inhibitor (KT5720) abolished the dbcAMP-in- the PLD and the cAMP-dependent pathways (Agwu duced phosphorylation of PLD1, but dibutyryl et al., 1991; Tyagi et al., 1991). Our previous study cGMP (dbcGMP) failed to phosphorylate PLD1. The showed that PLD activation was inhibited by cAMP- association between PLD1 and Val14RhoA in an dependent protein kinase (PKA) in a cell-free system immunoprecipitation assay was abolished by both of neutrophils (Kwak and Uhlinger, 2000). A com- dbcAMP and dbcGMP. Moreover, RhoA but not ponent of the plasma membrane has been identified PLD1 was dissociated from the membrane to the as the target of inhibition by PKA, and the dissoc- cytosolic fraction in dbcAMP-treated cells. These iation of the phosphorylated active form of RhoA from results suggest that both PLD1 and RhoA are the plasma membrane may be involved in the down- phosphorylated by PKA and the interaction bet- regulation of PLD activity (Kwak and Uhlinger, 2000). ween PLD1 and RhoA is inhibited by the phos- However, no prior evidence of the direct phosphor- Inhibition of interaction between PLD1 and RhoA by PKA 173 ylation of PLD by PKA in vivo and in vitro has been (Val14RhoA). Briefly, cells (2×106) in 60 mm culture presented. Moreover, direct involvement of the phos- dish were incubated with FuGENE6 transfection rea- phorylations of RhoA and/or PLD in the inhibition of gent (20 µl) and empty vector or pcDNA3.1 containing their association is not clearly understood although the cDNA of Val14RhoA plasmid (4 µg) for 24 h. RhoA phosphorylation by PKA was shown to enhance or inhibit the association between RhoA and its Immunoprecipitation and Western blot analysis binding proteins, such as Rho GDP dissociation 7 inhibitor (RhoGDI) and Rho kinase (Lang et al., 1996; Cells (2×10 ) were treated with or without 300 µM o Dong et al., 1998; Forget et al., 2002; Ellerbroek et of dbcAMP or dbcGMP for 10 min at 37 C, washed al., 2003; Qiao et al., 2003). once with cold phosphate buffered saline (PBS), and We report here that PLD1 is a substrate of PKA, resuspended in cold lysis buffer (20 mM Tris-HCl, pH and that interaction between PLD1 and RhoA is 7.4, 50 mM NaCl, 1% Triton X-100, 1% deoxycholate, inhibited by RhoA phosphorylation rather than PLD 1 µg/ml leupeptin, 1 µg/ml pepstatin A, 1 µg/ml apro- phosphorylation. tinin, and 1 mM phenylmethylsulfonyl fluoride). The samples were rotated for 1 h at 4oC and cell lysates were separated from insoluble material by centrifu- o Materials and Methods gation at 13,000 g for 5 min at 4 C. The samples were precleared using protein A-agarose for 1 h at o Reagents 4 C and incubated with anti-PLD antibodies for 1 h at 25oC. The immune complex obtained was then Dibutyryl cAMP (dbcAMP), dibutyryl cGMP (dbcGMP), transferred to 40 µl of 1:1 slurry of protein A- agarose protein A-agarose, and the catalytic subunit of PKA and mixed for 1 h. The beads were pelleted and (cPKA) were purchased from Sigma (St. Louis, MO); washed six times with PBS containing 1% Triton [γ-32P]ATP (3,000 Ci/mmol) and [γ-32P]Pi (3,000 X-100. The immunoprecipitated protein complex was Ci/mmol) from NEN Life Science Products (Boston, subjected to 10% sodium dodecyl sulfate-polyacr- MA); Dulbecco's modified Eagle's medium (DMEM) ylamide gel electrophoresis (SDS-PAGE) and trans- and geneticin (G418) from Gibco-BRL (Grand Island, ferred to nitrocellulose membranes. Membranes were NY); KT5720 and KT5823 from Calbiochem (La Jolla, blocked for 1 h at 25oC with a blocking buffer (25 CA); anti-Myc antibody from Santa Cruz Biotechnol- mM Tris-HCl, 0.15 M NaCl, and 0.2% Tween 20), ogy (Santa Cruz, CA); antibody directed against incubated with primary antibodies (1:1,000 in dilution) phospho-(Ser/Thr) PKA substrate from Cell Signaling overnight at 4oC, washed, and incubated for 1 h with Technology; the enhanced chemiluminescence (ECL) secondary antibodies (1:5,000 in dilution) conjugated kit from Amersham Bioscience (Beverly, MA); and to horseradish peroxidase. Signals were detected by FuGENE6 transfection reagent from Roche Applied chemiluminescence. Science (Manheim, Germany). Antibodies against PLD1 and PLD2 were generated as previously described (Kim et al., 2000). 32P incorporation into PLD The phosphorylation of PLD was determined from the 32 Cell culture and transfection amount of [ P]ATP incorporated into the enzyme after immunoprecipitation with specific PLD antibo- U87 human glioma cells were maintained in DMEM dies. Cells (5×106 cells/ml) were prelabeled with 0.2 medium supplemented with 10% fetal bovine serum mCi/ml [32P]orthophosphate for 90 min and incubated (FBS), 100 U/ml penicillin, and 100 µg/ml streptomy- with or without 300 µM of dbcAMP or dbcGMP for cin at 37oC in a humidified atmosphere containing 5% 10 min at 37oC. Phosphorylation of cell extracts with CO . pcDNA3.1 containing cDNA for PLD1 or PLD2 2 cPKA in vitro was carried out for 30 min at 30oC in were used as expression vectors (Min et al., 2001). a 100 µl final volume of phosphorylation buffer (50 U87 cells were transfected either with the empty mM Tris-HCl, pH 7.5, 10 mM MgCl , 1 mM dithi- vector as a control, or with the PLD-expression vector 2 othreitol, protease inhibitor mixture, 100 U cPKA, 0.1 using LipofectAMINE (Gibco-BRL), following the pro- mM cold ATP, and 2 µCi [32P]ATP). Cell lysates were cedure recommended by the manufacturer. Trans- immunoprecipitated using PLD antibodies. Samples fected cells were subsequently grown in selection were subjected to SDS-PAGE and followed by medium containing 500 µg/ml G418. After 3 weeks, autoradiography of dried gel. antibiotic-resistant clones were isolated and expanded for further analysis under selected conditions. Cells over-expressing PLD1 (U87/PLD1) and PLD2 Preparation of membranes and cytosolic fraction (U87/PLD2) were co-expressed with the Myc dbcAMP-treated or -untreated cells were centrifuged epitope-tagged constitutively active form of RhoA and resuspended at 2×107 cells/ml in ice-cold rela- 174 Exp. Mol. Med. Vol. 36(2), 172-178, 2004 xation buffer (50 mM Hepes, pH 7.4, 100 mM KCl, rate of cell proliferation was found not to be signi- 5 mM NaCl, 1 mM MgCl2, 0.5 mM EGTA, and pro- ficantly different between U87/PLD1 or U87/PLD2 tease inhibitor mixtures). Cell suspensions were cells and control U87 cells (data not shown). This sonicated for 20 s and centrifuged for 8 min at 700 finding suggests that these stable transfectants may g. The supernatants were ultracentrifuged at 100,000 be a suitable model for studying interactions between rpm for 45 min in a Beckman TL-100 ultracentrifuge PLD and its binding proteins, including RhoA.