Characterization of p87PIKAP, a Novel Regulatory Subunit of Phosphoinositide 3-Kinase γ That Is Highly Expressed in Heart and Interacts with PDE3B*

Phosphoinositide 3-kinase (PI3K) γ has been implicated in a vast array of physiological settings including the activation of different leukocyte species and the regulation of myocardial contractility. Activation of PI3Kγ is primarily mediated by Gβγ subunits of heterotrimeric G proteins, which are recognized by a p101 regulatory subunit. Here, we describe the identification and characterization of a novel regulatory subunit of PI3Kγ, which we termed p87PIKAP (PI3Kγ adapter protein of 87 kDa). It is homologous to p101 in areas that we have recently shown that they mediate binding to the catalytic p110γ subunit and to Gβγ. Like p101, p87PIKAP binds to both p110γ and Gβγ and mediates activation of p110γ downstream of G protein-coupled receptors. In contrast to p101, p87PIKAP is highly expressed in heart and may therefore be crucial to PI3Kγ cardiac function. Moreover, p87PIKAP and p101 are both expressed in dendritic cells, macrophages, and neutrophils, raising the possibility of regulatory subunit-dependent differences in PI3Kγ signaling within the same cell type. We further provide evidence that p87PIKAP physically interacts with phosphodiesterase (PDE) 3B, suggesting that p87PIKAP is also involved in the recently described noncatalytic scaffolding interaction of p110γ with PDE3B. However, coexpression of PDE3B and PI3Kγ subunits was not sufficient to reconstitute the regulatory effect of PI3Kγ on PDE3B activity observed in heart, implying further molecules to be present in the complex regulating PDE3B in heart.

Receptor-regulated class I phosphoinositide 3-kinases (PI3K) 2 are lipid kinases that produce the 3Ј-phosphorylated inositol lipid phosphatidylinositol 3,4,5-trisphosphate (PtdIns 3,4,5-P 3 ). It acts as a lipid second messenger by recruiting proteins containing pleckstrin homology (PH) domains to cellular membranes, thereby initiating various cellular responses (1,2). Class I PI3K are further subdivided into classes IA and IB according to their mode of activation. Class IB PI3K␥ is chiefly activated by G protein-coupled receptors (GPCR) and therefore grouped separately from the class IA PI3K that are activated down-stream of receptor tyrosine kinases. Insight into the physiological role of PI3K␥ has been mostly derived from the characterization of p110␥ knockout mice, which show defects in chemoattractant-induced neutrophil migration and oxidative burst, thymocyte development (3)(4)(5)(6), macrophage and dendritic cell (DC) migration (7), and the GPCR-dependent autocrine amplification of FceRI-mediated mast cell degranulation (8). Moreover, characterization of p110␥ knockout mice revealed a role for PI3K␥ both in the regulation of myocardial contractility and in cardiac remodeling processes (9,10). Recently, characterization of mice with a knockin of a catalytically inactive mutant of p110␥ revealed that the impact on contractility is probably mediated by a scaffolding interaction with phosphodiesterase (PDE) 3B, whereas remodeling processes are governed by pathways depending on catalytic activity of p110␥ (11).
Besides the catalytic p110␥ subunit, PI3K␥ consists of a p101 regulatory subunit that binds both p110␥ and G␤␥ (12). Although lipid kinase activity of p110␥ can be stimulated by G␤␥ in the absence of the regulatory p101 subunit (13), p101 appears to be necessary for G␤␥-mediated activation of PI3K␥ in living cells. In a heterologous reconstitution system, p101 binds to G␤␥ subunits and thereby recruits p110␥ to the plasma membrane, whereas p110␥ was neither recruited to the plasma membrane by G␤␥ nor activated by GPCR stimulation in the absence of p101 (14). However, p110␥ is functional and physiologically important in tissues where neither expressed sequence tag data nor direct experimental evidence validate an expression of p101, rendering the role of p101 still controversial. Recently, we were able to map the determinants relevant for interaction with p110␥ and G␤␥ to distinct areas within the p101 primary structure (15). These findings enabled us to identify an mRNA sequence within the DDBJ/EMBL/GenBank TM data base that encodes a distantly related p101 homologue (15). Whereas the encoded putative protein showed little overall sequence similarity to p101, a higher degree of conservation was observed within the regions that correspond to the p101 functional domains. Meanwhile, an initial characterization of this gene product was published by Suire et al. (16), who showed that it indeed interacts with p110␥ and G␤␥.
Here we report a different cloning strategy and further functional characterization of this novel PI3K␥ regulatory subunit, which we designated as p87 PIKAP (p87 PI3K adapter protein) (17,18). Similarly to p101, p87 PIKAP interacts with p110␥ and G␤␥. p87 PIKAP and p101 bind to p110␥ in a mutually exclusive fashion with a similar orientation within the dimeric complex. p87 PIKAP was necessary and sufficient to reconstitute a PI3K␥ signaling pathway in transfected HEK293 cells, mediating G␤␥-dependent activation of p110␥ downstream of a G icoupled receptor. p87 PIKAP mRNA was detected in various tissues, albeit most prominently in the heart. By contrast, p101 is only weakly expressed in the heart, whereas B and T cells feature p101 as the only p110␥ regulatory subunit. In DCs, neutrophils, and macrophages, both p87 PIKAP and p101 are coexpressed, raising the possibility of isoform-specific signaling with respect to the PI3K␥ regulatory subunit. Moreover, we present evidence that p87 PIKAP interacts with PDE3B, pointing to an additional involvement of p87 PIKAP in the recently described noncatalytic scaffolding interaction of p110␥ with PDE3B.
Cell Culture and Transfection-HEK293 cells were grown at 37°C and 5% CO 2 in Dulbecco's modified Eagle's medium or minimal essential medium with Earle's salts, supplemented with 10% fetal calf serum, 2 mM glutamine, 100 g/ml streptomycin, and 100 units/ml penicillin. COS-7 cells were cultivated at 37°C and 7% CO 2 in Dulbecco's modified Eagle's medium containing 4.5 g/liter glucose supplemented with 10% fetal calf serum, 2 mM glutamine, 100 g/ml streptomycin, and 100 units/ml penicillin. Transfection of HEK293 and COS-7 cells was performed using the FuGENE 6 transfection reagent (Roche Applied Science) according to the manufacturer's recommendations. The amount of transfected plasmid cDNA was 2 g/35-mm dish or 4 g/60-mm dish for coimmunoprecipitation (co-IP) experiments. The total amount of transfected plasmid was always kept constant by the addition of empty expression vector (pcDNA3) where necessary. For fluorescence microscopy experiments, the cells were seeded on glass coverslips 24 -48 h prior to the experiments.
Immunoprecipitation and Immunoblot Analysis-Immunoprecipitation and immunoblot analyses were carried out as described previously (15) using anti-FLAG M2 (Sigma) and anti-GFP antibodies (BD Biosciences) and suitable secondary antibodies (Sigma). For the analysis of p87 PIKAP and p101 stability, whole cell lysates were prepared by lysing the cells directly into Laemmli sample buffer followed by sonification (5 s) to ensure complete lysis. The lysates were analyzed by SDS-PAGE and immunoblot with anti-FLAG M2 antibody. Akt phosphorylation was analyzed in whole cell lysates using anti-Akt and anti-phospho-Akt (Ser 473 ) antibodies (Cell Signaling Technology).
Fluorescence Imaging and Confocal Microscopy-FRET efficiencies were determined using the acceptor photobleaching method as described previously (15). FRET efficiencies E were calculated using the equation E ϭ 1 Ϫ (F DA /F D ), with F DA representing the CFP fluorescence measured before bleaching YFP and F D representing the CFP fluorescence in absence of YFP acceptor. F D was obtained by linear regression of the increase in CFP fluorescence with the decrease in YFP fluorescence and extrapolation to zero YFP fluorescence, i.e. complete YFP photobleach (22). The expression levels of YFP-and CFP-tagged proteins were assessed by calibration of fluorescence intensities using an intramolecularly fused CFP-YFP construct. A 1.5-3-fold excess of YFPover CFP-tagged proteins was maintained to ensure comparability and to avoid situations where availability of the FRET acceptor may limit FRET efficiencies. For FRET competition studies, the excess of YFPover CFP-tagged proteins was limited to 1.2-1.5-fold, allowing for an effective competition with untagged proteins. For the analysis of p87 PIKAP and p101 expression, fluorescence intensities of YFP-tagged regulatory subunits of PI3K␥ and of a cotransfected free CFP (to identify transfected cells) were quantified using a 40ϫ/1.3 F-Fluar objective and CFPand YFP-selective band pass filters used for FRET microscopy (15). Pixel fluorescence intensities were integrated over single cells and corrected for background. Confocal microscopy was performed essentially as described (15). All of the imaging experiments were performed at room temperature in 10 mM HEPES, pH 7.4, 128 mM NaCl, 6 mM KCl, 1 mM MgCl 2 , 1 mM CaCl 2 , 5.5 mM glucose, and 0.2% bovine serum albumin.
Northern Blot Analysis-Multiple tissue Northern blots were purchased from BD Biosciences. RNA probes were generated with the StripEZ-SP6 kit (Ambion) using either BglII-linearized p87 PIKAP -FLAG or the supplied ␤-actin control plasmid as a template and ␣-[ 32 P]UTP (10 mCi/ml; PerkinElmer Life Sciences) as the radioactive label. Hybridizations were performed overnight at 68°C in UltraHyb hybridization buffer (Ambion) according to the manufacturer's protocol. The signals were detected on an image plate (Fujifilm), which was read by a phosphorimaging device (Fujifilm BAS Reader 1500).
Multiplex and Competitive PCR-Bone marrow-derived macrophages and splenocytes from C57BL/6 mice were obtained and cultured as described (23). Magnetic cell sorting of leukocyte subtypes from splenocytes of C57BL/6 mice was performed according to the manufacturer's conditions (Miltenyi Biotec). Purity of sorted cell populations was assessed with appropriate antibodies by flow cytometry on a LSR II cytometer (BD Biosciences) using the FlowJo analysis software (Tree-Star Inc.) as described (24). Total RNA was prepared using the High pure RNA kit (Roche Applied Science), and poly(A ϩ ) RNA was prepared with the MACS mRNA kit (Miltenyi Biotec). Both were carried out according to the manufacturer's protocols.
For the quantification of template copy numbers by competitive PCR, internal standards were generated for p87 PIKAP and p101 fragments. These were designed to have each a 40-bp deletion upstream of the 3Ј-primer-binding site to yield a construct of distinguishable size but a composition similar to the respective fragment to be quantified. These competitor constructs were generated by 20 cycles of PCR using the primers p87 PIKAP forward (see above) and p87 PIKAP competitor reverse (5Ј-GTGGGGCTGTCAGTGTAAATGGCTGCCCCTGGACC-3Ј) for the p87 PIKAP internal standard and p101 forward (see above) and p101 competitor reverse (5Ј-GCAGAGCCCCACTGAATGTCGT-CTCTGCTGGCTGG-3Ј) for the p101 internal standard. The PCR products were purified (High pure PCR product purification kit; Roche Applied Science), and the yield and purity were assayed by agarose gel electrophoresis and quantified by UV spectroscopy. Serial dilutions of these purified PCR fragments in H 2 O were then used as internal standards in subsequent PCRs. For quantitative analysis of band intensities, TINA 2.09 software (Raytest) was used.
PDE Assay-PDE activities were determined by a modification (25) of the two-step radioactive method described by Thompson and Appleman (26). Briefly, the cells were lysed in lysis buffer containing 20 mM Tris, pH 7.5, 0.5% Igepal Nonidet P-40, 100 mM NaCl, 50 mM NaF, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 3.2 g/ml soybean trypsin inhibitor, 0.5 mM benzamidine, and 2 g/ml aprotinin. Protein concentrations of cleared lysates were determined using the BCA protein assay (Pierce). The cell lysates were diluted in 20 mM Tris-HCl, pH 7.4, 5 mM MgCl 2 , and cAMP as the substrate (0.075 Ci/reaction (2,8)-3 H-cAMP (PerkinElmer Life Sciences) supplemented with 1 M unlabeled cAMP) and incubated for 10 min at 30°C (100 l of reaction volume). Substrate turnover did not exceed 15% of the total cAMP amount, which is well in the linear portion of the reaction. The reactions were terminated by boiling (3 min at 95°C). To convert the AMP to adenosine, the samples were incubated with 25 g of snake venom from Crotalus atrox (Sigma) for 10 min at 30°C. The samples were rocked on ice for 20 min with 400 l of a 1:1:1 slurry of Dowex (1 ϫ 8, 200 -400; Sigma), H 2 O, and ethanol and then centrifuged for 3 min. (2,8)-3 H-adenosine in the supernatant was then quantified by liquid scintillation counting. To specifically assess PDE3 activity, PDE activity was measured in the presence of the PDE3-specific inhibitor cilostamide (10 M; Sigma) in the reaction mix and compared with untreated samples. In each transfection experiment, PDE3 activities were determined in duplicate.

RESULTS
Identification and Cloning of p87 PIKAP -As previously reported, we have identified an mRNA (GenBank TM accession number BC028998), whose encoded putative protein shows homology to p101 within the p110␥and G␤␥-interacting domains (15), although the overall sequence similarity is relatively low (about 24% amino acid similarity). Based on expressed sequence tag data and the origin of the identified sequence, we chose murine DCs as a suitable source for RT-PCR-based amplification and cloning of the coding sequence of the p101 homologue. We could amplify and subclone the expected 2.3-kb fragment from total RNA of CD11c ϩ DCs of C57BL/6 mice. Clones derived from three independent PCRs were sequenced and yielded the coding sequence deposited in the DDBJ/EMBL/GenBank TM data base (accession code AY753194). It corresponds to the predicted coding region in BC028998, which was also used in the study of Suire et al. (16). Multiple clones were obtained containing a 12-bp insertion at the boundary of exons 12 and 13 (DDBJ/EMBL/GenBank TM accession number DQ295832). This insertion, however, did not result in obvious functional differences to the protein encoded by the deposited sequence (data not shown). The p87 PIKAP coding sequence consists of 20 exons, and the gene is located on murine chromosome 11 immediately next to the p101 gene (see the Ensembl data base at www.ensembl.org for further information on gene structure). An alignment of the protein sequence has been published previously (15). Although Suire et al. (16) proposed p84 as a name for the novel regulatory subunit, we intend to stick to our previously introduced nomenclature that is also used in the DDBJ/EMBL/GenBank TM data base entries pertaining to this gene (Refs. 17 and 18; see also GenBank TM entry AY753194).
Subcellular Distribution and Stability of p87 PIKAP -The subcellular distribution of p87 PIKAP was assayed in living cells by confocal microscopy. Unlike p101 fusion proteins, YFP-tagged p87 PIKAP localized almost exclusively to the cytosol of HEK293 and COS-7 cells regardless of the position of the fluorescent tag (Fig. 1A, upper panels). Because p110␥ also localizes to the cytosol in these cell types, it is not surprising that coexpression of p110␥ did not change the localization of p87 PIKAP (Fig. 1A, lower panels). In contrast, YFP-p101 was predominantly localized within the nucleus of both COS-7 and HEK293 cells in the absence of p110␥, whereas coexpression of p110␥ led to a redistribution to the cytosol (Fig. 1A, right panels; see also Ref. 14). For p101, a conserved nuclear localization signal can be located at positions 499 -502 (residues numbered as in pig p101). The p87 PIKAP sequence, however, lacks such nuclear targeting signals, and the protein partitions to the cytosol.
Furthermore, as observed previously (14), the overall fluorescence intensity of YFP-tagged p101 was substantially lower if it was expressed without p110␥. If cotransfected with free CFP as a control for transfection efficiency, expression levels of p101-YFP were reduced to 36 Ϯ 3% (n ϭ 3, 75-120 cells each) upon expression without p110␥. In contrast, expression levels of p87 PIKAP -YFP remained largely unchanged (reduction to 75 Ϯ 2%). To further test a dependence on p110␥ expression, HEK293 cells were transfected with different ratios of FLAG-tagged PI3K␥ subunits, whose expression was assayed in whole cell lysates. A dependence on p110␥ expression was observed for p101-FLAG (Fig.  1B), which was comparable with previous results with CFP-p101 (14). We thus conclude that these p101 fusion proteins are stabilized by coexpression of p110␥. In contrast, the expression level of p87 PIKAP -FLAG was largely independent of the amount of p110␥-FLAG coexpressed (Fig. 1B), indicating a higher stability of the p87 PIKAP protein in the absence of p110␥.

p87 PIKAP , a Novel Regulatory Subunit of PI3K␥
CFP-p101 copurified with p110␥-FLAG under the same conditions. A CFP-YFP fusion protein employed as a control for unspecific binding to the fluorescent protein moiety was not detectable in immunoprecipitates containing p110␥-FLAG. To validate the interaction between p87 PIKAP and p110␥ in living cells and to obtain further spatial information about the p110␥-p87 PIKAP complex, FRET measurements were performed. A representative acceptor photobleaching FRET experiment on HEK293 cells expressing YFP-p110␥ and CFP-p87 PIKAP is shown in Fig. 3A. A FRET efficiency of about 17% was determined (22) as depicted in Fig. 3B. To verify the specificity of the FRET signals, FRET competition assays were performed. If FRET is due to a specific interaction between two proteins, coexpression of untagged protein is expected to displace the tagged protein from its interaction partner, thereby leading to a reduction in FRET efficiency. Indeed, such a reduction in FRET efficiencies was observed if CFP-p87 PIKAP and YFP-p110␥ were coexpressed with either untagged p87 PIKAP or p110␥ (Fig. 3, C and  D), showing that the FRET signals measured arise from a specific protein-protein interaction.
FRET efficiencies were further determined for all combinations of CFP-and YFP-tagged p110␥ and p87 PIKAP in living HEK293 cells (Fig. 3, E and F). All of the FRET efficiencies were significantly higher than those measured for CFP-and YFP-tagged p87 PIKAP coexpressed with free YFP and CFP, respectively. FRET efficiencies comparable with those of the negative controls were obtained if YFP-p110␤ or CFP-p110␤ were used instead of p110␥ fusion proteins, indicating that p87 PIKAP is a class IB PI3K adapter protein that does not interact with class IA catalytic subunits. Similar to the situation observed for p110␥ and p101 (14), higher FRET efficiencies were obtained if fluorescent proteins were fused to the same termini in both p110␥ and p87 PIKAP . Assuming unhindered rotation of the fluorochromes, this indicates that the relative orientation of the polypeptide chains within the complex of p87 PIKAP and p110␥ is similar to that of p101 and p110␥.
Binding of p87 PIKAP and p101 to p110␥ Is Mutually Exclusive-To test whether p101 and p87 PIKAP bind to p110␥ in a pairwise or in a mutually exclusive fashion, we further employed FRET competition assays. The interaction between p110␥-YFP and p101-CFP was monitored by determining FRET efficiencies under conditions of cotransfection with either a control vector or plasmids encoding p87 PIKAP , p101, or p85␣. We observed that p101-CFP was displaced from its interaction with p110␥ upon coexpression of either p87 PIKAP or p101, whereas coexpression of p85␣ had no effect on the FRET efficiency (Fig. 4A). Similar results were obtained if the interaction between p87 PIKAP -CFP and p110␥-YFP was assayed under the same conditions (Fig. 4B). Thus, p87 PIKAP and p101 bind to p110␥ in a mutually exclusive fashion, indicating overlapping binding surfaces of p87 PIKAP and p101 on p110␥. Also, a rough measure of the relative binding affinities can be deduced from the competition assays. Because both p87 PIKAP and p101 competed with p101-CFP and p87 PIKAP -CFP with comparable efficiency (i.e. bring about comparable reductions in FRET efficiencies), both proteins should have a comparable affinity for the common interaction partner p110␥.
To more directly assess their relative affinity for p110␥, various amounts of CFP-tagged p87 PIKAP and p101 were cotransfected with p110␥-FLAG. p110␥-FLAG was immunoprecipitated, and the recovery of fluorescently tagged p87 PIKAP and p101 was analyzed by comparing the signals obtained in immunoprecipitates with those obtained in the lysates. If comparable amounts of CFP-p87 PIKAP and CFP-p101 were expressed, an excess of CFP-p87 PIKAP was detected in the immunoprecipitate (Fig. 4C, lane 3). Under conditions of a slight excess of CFP-p101, CFP-p87 PIKAP was still enriched to a greater extent within the immunoprecipitate (Fig. 4C, lanes 4 and 5). Still, CFP-p101 was able to  To control for unspecific binding, p110␥-FLAG, p101, and p87 PIKAP were each expressed alone. Additionally, p110␥-FLAG was coexpressed with a CFP-YFP fusion protein to control for unspecific binding of the fluorescent proteins. The data shown are from a representative experiment of three. IB, immunoblot.
Interaction between p87 PIKAP and G␤␥-According to the crucial role of p101 in the activation of p110␥ by G␤␥, we examined the ability of p87 PIKAP to bind to G␤␥ in living cells. Suire et al. (16) reported that p87 PIKAP has an ϳ4-fold lower affinity for G␤␥ as compared with p101 in in vitro lipid kinase assays using purified recombinant protein. Probably in line with these findings, we failed to observe a predominant membrane staining in HEK293 or COS-7 cells coexpressing YFP-p87 PIKAP and an excess of G␤ 1␥2 , although a membrane accumulation can be observed for YFP-p101 under the same conditions (data not  as well as either empty pcDNA3 vector or vectors encoding p87 PIKAP , p101, or p85␣ at the amounts indicated. The means and S.E. of four independent measurements with at least six cells each are shown. C, HEK293 cells were transfected with plasmids encoding p110␥-FLAG and CFP-tagged p101 and p87 PIKAP at the relative amounts indicated below. p110␥-FLAG was immunoprecipitated with an anti-FLAG antibody, and copurified CFP-p101 and CFP-p87 PIKAP was detected with an anti-GFP antibody (IP). The cell lysates were analyzed with anti-FLAG and anti-GFP antibodies to measure expression (load). The lanes are numbered from left to right. The data shown are from a representative experiment of three. IB, immunoblot.

p87 PIKAP , a Novel Regulatory Subunit of PI3K␥
shown; for YFP-p101 see Ref. 14). To test whether the interaction between p87 PIKAP and G␤␥ is still sufficient to drive p110␥ activation, we assayed G␤␥ interaction based on PI3K␥ activity. The YFP-fused PtdIns 3,4,5-P 3 -binding PH domain of Grp1 (YFP-Grp1-PH) acts as a translocating biosensor for class I PI3K activity in living cells (27). In HEK293 cells transfected with plasmids encoding G␤ 1␥2 , YFP-Grp1-PH was almost exclusively located within the cytosol (Fig. 5A, left panels). Coexpression of wild-type p110␥ led to a membrane localization of a minor fraction of YFP-Grp1-PH. An almost quantitative membrane localization pattern was only observed in cells that were additionally cotransfected with either wild-type p101 or p87 PIKAP , indicating a strong and sustained activation of PI3K␥ in this context (Fig. 5A, middle panels). Thus, like p101, p87 PIKAP functions as an adapter to drive activation of p110␥ by G␤␥. The degree of activation appeared to be slightly higher with p101 than with p87 PIKAP (Fig. 5), corresponding to the observation that p101 probably has a higher affinity for G␤␥. G␤␥ is necessary for a p101-or p87 PIKAP -mediated activation, because coexpression of the G␤␥-scavenging C terminus of ␤-adrenergic receptor kinase (␤ARK-CT-CFP) reduced the degree of YFP-Grp1-PH membrane association and because omission of G␤␥ completely abolished the translocation signal in the presence of p87 PIKAP and p110␥ (Fig. 5A, right panels; only shown for p87 PIKAP ). These results could be confirmed by analyzing the phosphorylation state of Akt, which is a primary downstream effector of PI3K signaling. G␤␥-mediated activation of p110␥ and subsequent phosphorylation of Akt on Ser 473 was only observed if either adapter protein p101 or p87 PIKAP was present (Fig. 5B).
p87 PIKAP Mediates Activation of PI3K␥ Downstream of GPCR Stimulation-To assess the role of p87 PIKAP in the activation of PI3K␥ downstream of chemokine receptor stimulation, YFP-Grp1-PH translocation upon treatment with fMLP was monitored in HEK293 cells expressing a reconstituted PI3K␥ signaling cascade consisting of the fMLP receptor, wild-type PI3K␥ subunits, and YFP-Grp1-PH. In agreement with the findings shown in Fig. 5, expression of either p101 or p87 PIKAP was required for fMLP-induced PI3K␥ activation (Fig. 6A). Expression of p101 resulted in a slightly more pronounced translocation of YFP-Grp1-PH, which was reminiscent of the results obtained for static overexpression of G␤␥ (Fig. 5). The fMLP-induced translocation of YFP-Grp1-PH was disrupted in the absence of p110␥ or upon expression of a kinase-deficient p110␥ (CFP-p110␥(K833R); Fig. 6B). Additionally, the fMLP-induced PI3K␥ activation mediated by p87 PIKAP was significantly reduced upon coexpression of ␤ARK-CT-CFP (Fig. 6B). Thus, we conclude that the observed translocation is due to catalytic activity of p110␥ and depends on the release of G␤␥ complexes and is mediated by either adapter p101 or p87 PIKAP . Essentially the same results were obtained using Akt phosphorylation as an independent read-out system (Fig. 6, C and D).
Expression Pattern of p87 PIKAP and p101-To explore in which physiological context p87 PIKAP may be important for the activation of PI3K␥, we examined the expression of p87 PIKAP mRNA using Northern blot analysis and semi-quantitative multiplex RT-PCR. The Northern blots containing 2 g of poly(A ϩ ) RNA isolated from tissues of 8 -10-weekold mice showed that transcripts of the expected size of 3.2 kb are most prominent in heart, but weaker signals in the other lanes indicate that p87 PIKAP is also broadly expressed in a variety of tissues including brain, spleen, lung, liver, kidney, prostate, thyroid, and salivary glands (Fig. 7). Expression in thymus was barely detectable, probably because of the adult age of the mice. In testis, a shorter transcript variant was detected, which is too short to encode a full-length protein and has therefore so far not been characterized further. Additional information can be obtained from expressed sequence tag data bases, which corroborate the results of Northern blots and also extend the expression pattern by bone marrow (see the UniGene entry Mm.234573).
We then went on to assay the relative expression of the PI3K␥ subunits p110␥, p101, and p87 PIKAP in heart and various leukocyte species, i.e. tissues that are known to harbor physiologically important PI3K␥ signaling cascades. In multiplex PCRs on reverse-transcribed RNA, fragments of GAPDH, p110␥, p101, and p87 PIKAP cDNA were simultaneously amplified and then analyzed by agarose gel electrophoresis. As expected, all cell types assayed showed expression of p110␥, albeit to a varying extent (Fig. 8). In agreement with the Northern blot data, p87 PIKAP was highly expressed in heart, where p101 was only marginally present. By contrast, in thymus and spleen, expression of p101 was more

p87 PIKAP , a Novel Regulatory Subunit of PI3K␥
abundant than that of p87 PIKAP , which was also barely detectable in thymus on the Northern blot (Fig. 7). However, examination of leukocyte subspecies revealed that p87 PIKAP and p101 are differentially expressed in leukocyte subpopulations. Although B and T cells feature p101 as the only p110␥ regulatory subunit, p87 PIKAP is clearly expressed, along with p101, in macrophages, neutrophils, and DCs.
To quantify the relative expression of p101 and p87 PIKAP , competitive PCR was performed. The amount of amplified cDNA fragments was compared with the amount of product obtained from an internal standard template of known copy number that has identical sequence expect for an ϳ40-bp deletion with respect to the native cDNA fragments. By varying the copy number of internal standard, conditions can be found where amplifi-cation for the cDNA fragment and the internal standard are equally efficient. In such reactions, the number of cDNA fragments equals the number of internal standard molecules initially introduced as template. Such sets of PCRs were generated for heart, neutrophils, and CD11b ϩ DC, which all contain different ratios of p101 and p87 PIKAP . Based on these assays, p87 PIKAP mRNA is expressed in heart at an about 5-fold higher level than p101 (about 21,500 and 4,400 copies of p87 PIKAP and p101 mRNA, respectively, per 100 ng of total RNA; Fig. 8B). In neutrophils, less p87 PIKAP than p101 mRNA was detectable (37,900 and 107,000 copies, respectively, per ng of poly(A ϩ ) RNA). An even higher excess of p101 mRNA was detected in CD11b ϩ DC (about 30,800 and 120,000 copies of p87 PIKAP and p101, respectively, per 100 ng of total RNA).  . Northern blot analysis of p87 PIKAP expression. Northern blots of multiple murine tissues were hybridized with a p87 PIKAP probe covering bases 1624 -2259 of the coding sequence deposited in AY753194 (spanning exons [15][16][17][18][19][20]. Equal loading was controlled by hybridizing the same blots with a ␤-actin control probe. p87 PIKAP , a Novel Regulatory Subunit of PI3K␥ APRIL 14, 2006 • VOLUME 281 • NUMBER 15 p87 PIKAP Interacts with PDE3B-In heart, an interaction between p110␥ and PDE3B has been shown to occur and to lead to an activation of PDE3B (11). However, this interaction seems to be mediated by additional unknown proteins, because recombinant p110␥ does not activate PDE3B in PDE3B-containing immunoprecipitates derived from hearts of p110␥ knockout mice (11). Because p87 PIKAP was found to be strongly expressed in heart (Figs. 7 and 8), we explored the possibility that p87 PIKAP may connect PDE3B and p110␥. To this end, HEK293 cells were transfected with combinations of PDE3B-FLAG and CFP-tagged PI3K subunits. Only CFP-p87 PIKAP was efficiently copurified in PDE3B-FLAG immunoprecipitates (Fig. 9, left panels). Weak signals were obtained for CFP-p101 and CFP-p110␥, whereas no signal was detectable for CFP-p85␣. If cells were cotransfected with CFP-p87 PIKAP , p110␥-CFP, and PDE3B-FLAG and subjected to IP with anti-FLAG antibodies, copurification of both CFP-tagged proteins was observed, but CFP-p87 PIKAP was more efficiently copurified than p110␥-CFP (Fig.  9, left panels). Moreover, CFP-p87 PIKAP was less efficiently copurified in the presence of p110␥, indicating that heterodimerization of p87 PIKAP and p110␥ may result in a lower affinity to PDE3B. Similar results were obtained with a different protocol for lysis and co-IP that employs a RIPA buffer (data not shown). To control for cell lysis artifacts, cells transfected with TRPV1-FLAG instead of PDE3B-FLAG were assayed, resulting in very faint to undetectable signals of CFP-tagged proteins in the immunoprecipitates (Fig. 9, right panels). To further exclude lysis artifacts, the cells were separately transfected with PDE3B-FLAG, p110␥, p101, or p87 PIKAP . The lysates were then combined and subjected to the same IP procedure. Under these conditions, copurification of PI3K␥ subunits was not observable (data not shown). In the reciprocal setting, CFP-PDE3B could also be copurified by immunoprecipitation of coexpressed p87 PIKAP -FLAG (Fig. 9, middle panels). Thus, we conclude that p87 PIKAP interacts with PDE3B.
To assess whether interaction of p87 PIKAP with PDE3B modulates PDE3B activity, PDE assays were performed on lysates of HEK293 cells transfected with PDE3B-FLAG and different combinations of PI3K␥ subunits. PDE3 activity was 8 Ϯ 2 pmol/min/mg protein in mock-transfected HEK293 cells. PDE3 activity remained unchanged if either p110␥ or p87 PIKAP -FLAG were coexpressed (Fig. 10). In cells transfected with PDE3B-FLAG, activities of 890 Ϯ 210 pmol/min/mg protein were determined. The activity of recombinant PDE3B remained unchanged if an excess of p110␥, p87 PIKAP -FLAG, or both p110␥ and p87 PIKAP -FLAG was cotransfected. A reduction in PDE3B activity was observed upon coexpression with p101-FLAG. To verify equal expression of recombinant PDE3B-FLAG, immunoblotting experiments were performed. Whereas coexpression of p87 PIKAP or p110␥ did not affect the expression of PDE3B-FLAG, coexpression of p101 markedly reduced the expression of PDE3B, thereby explaining the reduced PDE3B activity in these samples (see blot in Fig. 10). Furthermore, no change in PDE3B activity was observed if cell lysates of PDE3B-FLAG-expressing cells were assayed for PDE activity in the presence of 100 nM recombinant hexahistidine-tagged p87 PIKAP affinity-purified from baculovirus-infected Sf21 cells (Fig. 10, right panel). We therefore conclude that the interaction of p87 PIKAP or of a p87 PIKAP /p110␥ dimer with PDE3B is not sufficient for modulating PDE3B activity.

DISCUSSION
Here we report the cloning and characterization of a novel p101 homologue. We previously suggested its existence and possible functional relevance based on its amino acid similarity to the N-and C-terminal functional domains of p101 (15). Based on a similar notion, Suire et al. (16) also identified and initially characterized this novel PI3K␥ regulatory subunit and termed it p84. Because we previously presented data on and submitted sequence data pertaining to this regulatory subunit, we used our previously introduced name p87 PIKAP within this publication (Ref. 17; see also GenBank TM entry AY753194).
We could confirm that p87 PIKAP interacts with p110␥ and extend knowledge about this interaction using co-IP and FRET assays. Both p101 and p87 PIKAP bind to the same surface or to at least overlapping binding surfaces on p110␥ because their binding is mutually exclusive. The results of the FRET measurements on N-or C-terminally tagged p87 PIKAP and p110␥ constructs indicate that the complex between p87 PIKAP and p110␥ probably resembles that of p101 and p110␥ in that both N termini and both C termini are closer to each other than to the opposite termini (14). Based on competitive FRET and co-IP assays, the affinity of both p101 and p87 PIKAP for p110␥ is in a similar range and perhaps slightly higher in the case of p87 PIKAP . In contrast to p101, p87 PIKAP is lacking nuclear localization signals and remains mostly within the cytosol also in the absence of p110␥. Furthermore, in the absence of p110␥, p87 PIKAP was found to be more stable than p101. Neutrophil lysates from p110␥ knockout mice, however, showed strong reductions in protein levels for both p101 and p87 PIKAP (16). The difference between these findings probably results from additional or cell type-specific regulatory effects on the mRNA or protein level within the native context. Moreover, p87 PIKAP may have a somewhat extended half-life compared with p101, which is visible 48 h after transfection, Equal amounts of reverse-transcribed RNA were used as template in each PCR, whereas the amount of internal standard introduced into the reaction was varied from 10 3 to 10 6 copies/reaction in steps of 0.5 orders of magnitude as indicated below. The last reaction of each PCR set for p87 PIKAP and p101 contains 10 6 copies of internal standard but no RT. In reactions yielding equal amounts of the slightly smaller internal standard and of the original cDNA fragment, equal amounts of template molecules had been present in the reaction set-up. The ethidium bromide-stained agarose gels shown are representative of two sets of PCRs each.

p87 PIKAP , a Novel Regulatory Subunit of PI3K␥
but negligible compared with an additional long term stabilization by p110␥. We further observed that expression of monomeric p101 results in reduced expression of cotransfected cDNAs (for example of PDE3B-FLAG in Figs. 9 and 10, or of free CFP (data not shown)). A nuclear function of monomeric p101 in regulating the expression of other proteins may be supported by yeast two-hybrid data showing the interaction of p101 with transcriptional regulators (Alliance for Cellular Signaling; www. signaling-gateway.org/).
Although overexpression of G␤␥ did not result in a marked plasma membrane localization of p87 PIKAP , its interaction with G␤␥ could nevertheless be revealed via G␤␥ stimulation of p110␥ activity in the presence of p87 PIKAP . These findings may extend the in vitro data of Suire et al. (16) to a context of living cells; in vitro lipid kinase assays showed 4-fold stronger stimulation by G␤␥ for the p101/p110␥ than for the p87 PIKAP /p110␥ heterodimer. In accordance with this finding, we could demonstrate that, in living cells, receptor-mediated activation of p87 PIKAP /p110␥ heterodimer results in a less pronounced translocation of YFP-Grp1-PH, which may be explained by the lower affinity of p87 PIKAP for G␤␥.
The relative level of expression of p87 PIKAP and p101 in subtypes of leukocytes differs between cell types. The presence of p87 PIKAP in macrophages, neutrophils, and DCs may render their PI3K␥ signaling different from that in B and T cells, from which p87 PIKAP is virtually absent. Both regulatory subunits may differ in specificity for subtypes of the upstream activator G␤␥. A systematic screen for activation by various G␤␥ dimers performed on the p101/p110␥ heterodimer revealed that the G ␤1␥11 dimer, although highly expressed in tissues containing p110␥, is ineffective in stimulating the p101/p110␥ heterodimer (28). A similar screen on the p87 PIKAP /p110␥ heterodimer may be helpful to settle this question of different G␤␥ specificities. Moreover, the slower FIGURE 9. p87 PIKAP interacts with PDE3B. HEK293 cells were transfected with plasmids encoding the indicated proteins. The cells were lysed, and FLAG-tagged proteins were precipitated with an anti-FLAG antibody. The recovery of FLAG-tagged protein was tested by probing with an anti-FLAG antibody, and copurification of CFP-tagged protein was analyzed with an anti-GFP antibody (IP). Aliquots of cell lysates used for IP were probed with anti-FLAG or anti-GFP antibodies to assay for expression of FLAG-and CFP-tagged proteins (load). Left panels, IP of PDE3B-FLAG; middle panels, IP of p87 PIKAP -FLAG; right panels, IP of TRPV1-FLAG employed as a negative control. Prominent bands of copurified protein can be observed for p87 PIKAP in the PDE3B-FLAG immunoprecipitate as well as for CFP-PDE3B in the p87 PIKAP -FLAG immunoprecipitate. The experiments shown are representative of three each. IB, immunoblot.
FIGURE 10. Effect of PI3K␥ subunits on PDE3B activity. Left panel, HEK293 cells were transfected with plasmids encoding FLAG-tagged versions of the indicated proteins (0.2 g of PDE3B-encoding plasmid and a total of 1.8 g of plasmid cDNA encoding the indicated PI3K␥ subunits). The cells were lysed, and PDE activity was determined in the presence and absence of 10 M cilostamide to assess PDE3 activity. To maintain comparability between assays from different transfection experiments, PDE3B activities (in pmol/min/mg protein) were normalized to the activity in lysates of cells transfected with only PDE3B-FLAG. The means and S.E. of three independent transfection experiments are given. The amount of recombinant PDE3B-FLAG in cell lysates was analyzed by immunoblotting (IB) with anti-FLAG antibody. A blot from a representative experiment is shown. Right panel, lysates of HEK293 cells transfected with a plasmid encoding PDE3B-FLAG were incubated with 100 nM purified recombinant p87 PIKAP in either native (nat.) state or denatured (boiled for 5 min, denat.) state as a control. The means and S.E. of two independent experiments are given. accumulation of PtdIns 3,4,5-P 3 observed in cells expressing the p87 PIKAP /p110␥ heterodimer instead of the p101/p110␥ heterodimer (see above) may result in different kinetics of PI3K␥ signaling events in cells expressing both p87 PIKAP and p101.
Based on data showing an indirect interaction of p110␥ with PDE3B that regulates PDE3B activity in heart, a PDE3B-regulating multi-protein complex, which is disrupted upon genetic ablation of p110␥, has been postulated (11). According to our Northern blot and RT-PCR data, p87 PIKAP is highly expressed in the heart. Because the expression of p87 PIKAP and p101 is strongly reduced in p110␥ knockout mice (16), the absence of p87 PIKAP in hearts of p110␥ knockout mice may explain why recombinant p110␥ is unable to reconstitute regulation of PDE3B immunoprecipitated from hearts of p110␥ knockout mice. Therefore, we asked whether p87 PIKAP is a component of the p110␥-containing complex regulating PDE3B activity in the heart. We could show that p87 PIKAP and also p101 interact with PDE3B.
To test whether the regulation of PDE3B observed in heart can be reconstituted by p87 PIKAP in vitro, we assayed PDE3B activity in the presence of p87 PIKAP , p110␥, or the p87 PIKAP /p110␥ heterodimer. However, for both monomeric p87 PIKAP and the p87 PIKAP /p110␥ heterodimer, the effects on PDE3B activity were not observed in vitro. Because p87 PIKAP interacts with PDE3B, it may be an essential part of a PDE3B-regulating protein complex in the heart, although the lack of effects on PDE3B activity indicates that additional proteins are required. Gene knockdown and knockout studies on regulatory subunits of PI3K␥ will be necessary to obtain further knowledge on the role of p87 PIKAP in cardiac PI3K␥ signaling and to reveal the relative contributions of p87 PIKAP and p101 to PI3K␥ signaling.