Coordinate Regulation of Catecholamine Uptake by rab3 and Phosphoinositide 3-Kinase*

Previously we observed that rab3 GTPases modulate both the secretion of catecholamines from PC12 neuroendocrine cells and the steady-state accumulation of exogenous norepinephrine (NE) into these cells (Weber, E., Jilling, T., and Kirk, K. L. (1996) J. Biol. Chem. 271, 6963–6971). Here we addressed the mechanisms by which these monomeric GTPases stimulate NE uptake by PC12 cells including their effects on uptake kinetics, their sites of action (secretory granule membrane versus plasma membrane), and the involvement of rab3-inter-acting proteins in this process. We observed that rab3B stimulated the rate and maximal accumulation of radiolabeled NE into large dense core vesicles within intact PC12 cells. rab3A and rab3B also increased NE uptake into large dense core vesicles in digitonin-permeabi-lized PC12 cells, which indicates that these GTPases stimulate catecholamine uptake at the level of the secretory granule membrane. In an attempt to identify rab3B targets that may mediate this effect on NE uptake, we found that rab3B interacts directly with phosphoinositide 3-kinase (PI3K) in a GTP-dependent fashion and that PI3K activity was elevated in PC12 cells overexpressing rab3B. Furthermore, two structurally distinct inhibitors of PI3K (wortmannin and LY294002) inhibited NE uptake in intact as well as digitonin-permeabi-lized PC12 difluoride membrane, rinsed in Tris-buffered saline, and then blocked in 5% dry milk plus 0.1% Tween 20 in Tris-buffered saline (10 m M Tris-HCl, 150 m M NaCl, pH 7.5). Blots were incubated with rab3B polyclonal antibody (16) or monoclonal antibodies against p85 (cid:1) regulatory subunit of PI3K (Transduction Laboratories) or phosphotyrosine (PY20; Transduction Laboratories). Secondary antibodies were horseradish peroxidase-con-jugated anti-mouse or anti-rabbit secondary antibodies as appropriate incubated at room temperature fo r 1 h in 1 M glucose, 0.5% Tween 20, 10% glycerol, and 5% dry milk in Tris-buffered saline. The blots were developed by Renaissance Chemiluminescence (PerkinElmer Life Sciences). Measurement of NE Transport in Intact Cells— Transport assays were performed on 2-day-old PC12 cultures seeded on 12- or 24-well poly- L -lysine-coated culture dishes. 343 n M [ 3 H]NE with varying con- centrations of unlabeled NE (0–3.43 (cid:3) M ) were added to 1.0–1.5 (cid:3) 10 6 cells in RPMI 1640 medium (Sigma) supplemented with 10% heat- inactivated horse serum, 5% fetal bovine serum, 1% penicillin/strepto-mycin, and 400 units/ml hygromycin. After incubation for varying time periods at 37 °C (typically 30 min), uptake was terminated by removal of the [ 3 H]NE-containing media. Cells were rinsed twice in fresh RPMI media and then chased 30–45 min in RPMI media at 37 °C. At the end of the chase, cell-associated radioactivity was determined by lysing the cells in 0.25 N NaOH and measuring the amount of [ 3 H]NE in the extracts by liquid scintillation counting. For PI3K inhibitor studies, transport was performed as before except that cells were pretreated with various concentrations of LY294002 A , antibody Anti-rab3B incubated with whole cell lysates from two mock and two rab3B clones the immune complexes were subjected to an in vitro PI3K assay. The of phosphoinositide 3-phosphate formed in each was quantified by phosphorimage GST-rab3B

Transport and storage of the classical monoamines by neuroendocrine cells occur through the activities of two pharmacologically and functionally distinct transporters. Neurotransmitter uptake across the plasma membrane is facilitated by specific plasma membrane transporters that remove neurotransmitter from the extracellular milieu (1,2). Vesicular monoamine transporters (VMATs) 1 localized on the membranes of secretory vesicles concentrate monoamine neurotransmitters into large dense core vesicles (LDCVs), chromaffin granules, and secretory vesicles. VMAT is an electrogenic antiporter that uses a H ϩ electrochemical gradient maintained by an ATP-dependent vacuolar H ϩ pump located in the secretory vesicle membrane to drive neurotransmitter uptake. A number of drugs, including reserpine and tetrabenazine, inhibit neurotransmitter uptake by VMAT (3,4). In both neurons and neuroendocrine cells, catecholamine transport across the vesicular membrane serves to rapidly and efficiently package newly synthesized and recycled catecholamines into secretory vesicles to control the supply of neurotransmitter available for exocytosis. Catecholamine uptake by neurons also helps terminate synaptic transmission by the removal of neurotransmitter from the synaptic cleft.
Both steps in catecholamine uptake (plasma membrane and vesicular membrane) are subject to regulation by second messengers that are in turn activated by receptor binding or membrane depolarization (5,6). Neurotransmitter uptake is also regulated by heterotrimeric and monomeric GTPases. The heterotrimeric G-protein, G␣ o2 , inhibits catecholamine uptake by a neuroendocrine cell line (PC12) that accumulates norepinephrine into secretory granules by mechanisms similar to those utilized by chromaffin cells and noradrenergic neurons (7). In addition, monomeric GTPases of the rab3 subfamily can stimulate NE uptake by PC12 cells (8). rab3 GTPases are expressed in a variety of secretory cells including PC12 cells, where they localize to catecholamine-containing secretory granules (LDCVs). rab3 GTPases have been implicated in the regulation of exocytosis (8,11,12). Interestingly, however, we observed in a previous study that stably transfected PC12 cell lines that overproduce rab3A or rab3B also accumulated substantially greater amounts of exogenous NE than mock transfected or untransfected cells (8). This effect was greater for rab3B than rab3A and was seen for multiple rab3-overproducing clones (i.e. was not the result of clonal variation). This raises the interesting possibility that rab3 GTPases regulate not only the efficiency of secretion but also the amount of vesicle cargo available for secretion. The mechanism for the rab3-induced increased in NE uptake was not addressed in the earlier study and is the subject of the present investigation.
Here we addressed three outstanding questions concerning the regulation of NE uptake by rab3 GTPases in PC12 cells: (i) how does rab3B affect the kinetics (time course and concentration dependence) of NE uptake by intact PC12 cells, (ii) do rab3B and rab3A regulate NE uptake at the plasma membrane or the secretory granule membrane, and (iii) do rab3 GTPases interact with other molecules to regulate this process? We report that rab3B stimulates the rate and maximum accumulation of NE by PC12 cells, and that one of the sites of action for rab3B and rab3A is the secretory granule uptake step. Furthermore, we identify phosphoinositide 3-kinase (PI3K) as a putative rab3B effector protein and demonstrate that PI3K, like rab3B, positively regulates vesicular catecholamine uptake by PC12 cells. The involvement of PI3K in the regulation of NE uptake could be generally significant, because PI3K participates in a variety of cell signaling pathways that are activated by multiple growth factors and hormones (13,14).

MATERIALS AND METHODS
Cell Culture-PC12 cells stably transfected with rab3A, rab3B, the GTP binding mutant rab3BN135I, and vector alone (mock) were generated as described (8). In our previous analysis of these clones, we observed that, like the endogenous rabA in PC12 cells, recombinant rab3A and recombinant rab3B localized to LDCVs in these cells (8).
PC12 cells were cultured in HEPES-modified RPMI with sodium bicarbonate and supplemented with 10% heat-inactivated horse serum, 5% fetal bovine serum, 1% penicillin/streptomycin (Sigma), and 400 units/ml hygromycin B (Calbiochem, La Jolla, CA) at 37°C in a humidified incubator with 95% CO 2 and 5% O 2 . All experiments were performed using PC12 cells cultured in the absence of nerve growth factor.
Productin of Glutathione S-Transferase (GST) Fusion Proteins, in Vitro Binding Assays, and Immunoblotting-GST fusion proteins containing full-length human rab3B, rab5, and the rat p85␣ regulatory subunit of PI3K were expressed in Escherichia coli using the pGEX-2T prokaryotic expression vector (Amersham Biosciences, Inc.). Transformed E. coli were induced with 100 M isopropyl-1-thio-␤-D-galactopyranoside at 37°C for 3 h. GST fusion proteins were affinity purified from bacterial cell lysates with glutathione-agarose, as described (15). The GST moiety was removed from GST-p85 PI3K by incubation with thrombin (1% thrombin w/w GST fusion protein) at 25°C for 1 h. Thrombin was neutralized by incubation with p-aminobenzamidine (Sigma) for 1 h at 4°C. 1 g of immobilized GST, GST-rab3B, or GST-rab5 prebound to 0.5 mM GDP or GTP␥S was mixed with 1 g of p85 (GST moiety removed) in the following buffer (20 mM Tris-HCl, pH 8.0, 100 mM NaCl, 10 mM MgCl 2 , 10% glycerol, and 1 mM dithiothreitol) supplemented with leupeptin, aprotinin, and pepstatin A (10 g/ml) and 2 mM phenylmethylsulfonyl fluoride for 1 h at 4°C. After 1 h, the bound proteins were washed extensively with the same buffer, and then eluted in 5ϫ SDS sample buffer. The samples were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), transferred to polyvinylidene difluoride membrane (Bio-Rad), and blotted with a monoclonal antibody to p85␣ (Transduction Laboratories, Lexington, KY).
Measurement of NE Transport in Intact Cells-Transport assays were performed on 2-day-old PC12 cultures seeded on 12-or 24-well poly-L-lysine-coated culture dishes. 343 nM [ 3 H]NE with varying concentrations of unlabeled NE (0 -3.43 M) were added to 1.0 -1.5 ϫ 10 6 cells in RPMI 1640 medium (Sigma) supplemented with 10% heatinactivated horse serum, 5% fetal bovine serum, 1% penicillin/streptomycin, and 400 units/ml hygromycin. After incubation for varying time periods at 37°C (typically 30 min), uptake was terminated by removal of the [ 3 H]NE-containing media. Cells were rinsed twice in fresh RPMI media and then chased 30 -45 min in RPMI media at 37°C. At the end of the chase, cell-associated radioactivity was determined by lysing the cells in 0.25 N NaOH and measuring the amount of [ 3 H]NE in the extracts by liquid scintillation counting. For PI3K inhibitor studies, transport was performed as before except that cells were pretreated with various concentrations of LY294002 (Biomol) or wortmannin (Calbiochem) for 6 h prior to transport assay. Transport assays were performed in duplicate. The protein concentration of each lysate was determined using a Micro BCA protein assay reagent kit (Pierce), and the uptake values were normalized to the amount of total protein (mg) in the lysate. K m and V max values were determined from plots of normalized transport activity versus total NE concentration (see "Results").
Measurement of Vesicular NE Uptake in Digitonin-permeabilized Cells-NE uptake by digitonin-permeabilized cells was performed as described with minor modifications (18,23). PC12 cells were seeded on 12-well poly-L-lysine-coated dishes, cultured for 2 days, and then permeabilized with 100 M digitonin in complete media for 15 min at room temperature. The cells were then rinsed in HTMS buffer containing 20 mM HEPES, 20 mM Tris-HCl, 6  The cells were then homogenized in 10 mM HEPES, 0.32 M sucrose buffer (1 ϫ 40 strokes with a Potter homogenizer) and centrifuged at 14,000 rpm to pellet unbroken cells. The clarified homogenate was layered on a 0.6 -1.8 M sucrose gradient with a 2.2 M sucrose pad (10-ml volume). The gradient was centrifuged at 110,000 ϫ g for 6 h at 4°C. Fractions (0.5 ml) were collected from the bottom of the tube. An equal aliquot of each fraction was analyzed by liquid scintillation counting for 3

[H]NE and by Western blotting for markers of LDCVs (secretogranin) and light vesicles (synaptophysin).
Measurement of PI3K Activity in Immunoprecipitates-PC12 cells plated on 10-cm poly-L-lysine-coated dishes were cultured until confluent. Cells were lysed in lysis buffer containing 20 mM Tris, pH 7.5, 1% Nonidet P-40, 137 mM NaCl, 135 mM KCl, 5 mM MgCl 2 , 5 mM CaCl 2 , 4 mM Na 3 VO 4 , and 10% glycerol supplemented with leupeptin, aprotinin, and pepstatin A (all at 10 g/ml) and 2 mM phenylmethylsulfonyl fluoride for 15 min at 4°C. Lysates were clarified by centrifugation at 14,000 ϫ g. 10 g of anti-phosphotyrosine (PY20) (Transduction Laboratories), anti-p85 PI3K (Transduction Laboratories), or normal mouse IgG (Santa Cruz Biotechnology Inc., Santa Cruz, CA) were mixed for 2 h at 4°C with lysates (1 ml) that were prepared from mock and rab3Bexpressing PC12 cells and matched for equal protein concentrations. Protein A/G plus agarose (Santa Cruz Biotechnology Inc.) was added to the lysates followed by mixing for an additional 2 h. To determine whether rab3B interacts with PI3K, 25 g of anti-rab3B polyclonal IgG (8,16) bound to cyanogen bromide-Sepharose was used for the immunoprecipitation step.

RESULTS
Previously we observed that recombinant rab3A and rab3B localized to LDCVs in PC12 cells and markedly stimulated the steady-state accumulation of radiolabeled NE by these cells (8). We observed no discernible differences in either the total number of LDCVs per cell or the number of LDCVs within 110 nm of the plasma membrane (8), suggesting that other factors were responsible for the observed effects of rab3 on NE uptake. In the present study we addressed the mechanism by which rab3 GTPases regulate catecholamine uptake by PC12 neuroendocrine cells. We focused on rab3B because this GTPase had quantitatively greater effects on NE uptake, although rab3A appears to exhibit qualitatively similar effects (Ref. 8; see below).
rab3B Stimulates the Rate and Maximal Accumulation of NE by Intact PC12 Cells- Fig. 1 shows the effects of rab3B expres-rab3 and PI3K Stimulate Catecholamine Uptake sion on the kinetics of [ 3 H]NE uptake by intact PC12 cells. The uptake of NE was time-dependent and reached a plateau within 2 h for either the rab3B transfected cells or mock transfected cells (Fig. 1A). The initial rate of uptake and the steadystate accumulation of radiolabeled NE were consistently much higher in rab3B-expressing cells than in mock cells. The higher steady-state accumulation observed here was consistent with what we had found previously for multiple rab3B-expressing clones (8). To characterize further the effects of rab3B on uptake kinetics, we measured uptake over a range of NE concentrations at a single time point where NE uptake is a linear function of time (i.e. 30 min; Fig. 1B). In each case the uptake data could be well fit by a single Michaelis-Menten function (see Fig. 1 legend), suggesting that one of the two steps in NE uptake by intact PC12 cells (plasma membrane or vesicle mem-brane) is limiting under these conditions. The primary effect of rab3B was to increase the maximal accumulation (V max ) of NE uptake by intact PC12 cells.
rab3 Stimulates NE Uptake at the Secretory Granule-Extracellular NE is concentrated into secretory granules (LDCVs) via a two-step process in which NE is transported across the plasma membrane into the cytosol and then across the vesicular membrane into the LDCV (19,20). It is conceivable that the greater amount of NE uptake exhibited by rab3B-expressing cells reflects an increase in cytosolic accumulation with no corresponding increase in uptake into LDCVs. Thus, to determine whether rab3B increased the accumulation of exogenous NE into secretory granules as opposed to simply increasing the cytoplasmic levels of NE, intact cells preloaded with [ 3 H]NE were fractionated by sucrose gradient ultracentrifugation, and the amount of [ 3 H]NE in each fraction was determined by liquid scintillation counting. As shown in Fig. 2, the peak of [ 3 H]NE was localized in fractions 8 -12 (corresponding to 1.4 -1.6 M sucrose), and co-fractionated with secretogranin, a soluble protein marker of LDCVs (21). Fractions were also immunoblotted with antibodies against synaptophysin, a marker of endosomes and small secretory vesicles in PC12 cells (22). Secretogranin and synaptophysin were enriched in distinct fractions, indicating that LDCVs and small secretory vesicles were fractionated into separate pools. The NE signal at the top of the gradient probably represents NE released from LDCVs that were lysed during centrifugation because the soluble LDCV cargo protein, secretogranin, was also present in these fractions. The rab3B-expressing cells exhibited a greater amount of [ 3 H]NE in the LDCV fractions than in mock cells, which indicates that the greater amount of NE accumulated by these cells represents at least in part more NE within secretory granules.
Although our fractionation data indicate that rab3B expression leads to an increase in the amount of NE within LDCVs, these data do not establish that the vesicle uptake step is directly stimulated by this GTPase. rab3B could secondarily increase NE accumulation into LDCVs by stimulating the plasma membrane uptake step and thereby increasing the cytosolic pool of NE available for transport into LDCVs. To rab3 and PI3K Stimulate Catecholamine Uptake identify the site of regulation of NE transport by rab3B and rab3A (e.g. plasma membrane or LDCV membrane), we measured NE uptake in digitonin-permeabilized cells. In this assay NE diffuses into the cytosol via bulk flow through pores in the plasma membrane created by the detergent, digitonin, and is then actively transported into secretory vesicles (17,18). We were encouraged to use this assay because our earlier membrane fractionation results indicated that rab3A and rab3B remain tightly associated with secretory granules following cell homogenization (8) and, therefore, would be expected to remain on these granules in semi-intact cells. To validate this assay we first showed that vesicular uptake in permeabilized cells was stimulated by ATP and that only the ATP-dependent component of uptake was sensitive to reserpine, an inhibitor of the vesicle membrane monoamine transporter (Fig. 3A). To assess the efficiency of permeabilization, we monitored NE uptake in the presence and absence of nomifensine, a plasma membrane NE transporter inhibitor (Fig. 3B). NE uptake was inhibited by nomifensine in intact cells but not in permeabilized cells. Conversely, the VMAT inhibitor, reserpine, inhibited uptake by both intact and permeabilized cells (Fig. 3C). Collectively, these results indicate that NE uptake by digitonin-permeabilized cells represents NE uptake at the vesicle membrane independent of a significant contribution from the plasma membrane transport process.
Using this assay we observed that reserpine-sensitive NE uptake by digitonin-permeabilized cells was stimulated ϳ3-fold in rab3B PC12 cells compared with mock PC12 cells (Fig. 3D). We also measured NE uptake in permeabilized cells that were overexpressing rab3A or that were expressing a rab3B mutant with accelerated on and off rates for GTP binding (rab3BN135I ( Refs. 8 and 26)). NE uptake was increased 1.5-fold in permeabilized PC12 cells overproducing rab3A compared with mock cells (Fig. 3D). Thus, both rab3 isoforms positively regulate NE uptake in permeabilized PC12 cells as well as in intact cells (8). Interestingly, rab3BN135I had no effect on the amount of [ 3 H]NE accumulated in permeabilized cells, despite the fact that this mutant does stimulate NE uptake by intact cells (8).
The lack of effect of rab3BN135I expression on vesicular uptake in semi-intact cells may reflect the fact that this mutant binds much less tightly to membranes as compared with wild type rab3B or rab3A (8) and could have leaked out of the cells during permeabilization. However, we cannot rule out the possibility that rab3 may regulate uptake at the plasma membrane as well as at the granule surface, and that its modulation of granule uptake is more sensitive to mutations that influence the kinetics of nucleotide binding. Either way the data obtained for semi-intact PC12 cells show that rab3B and, to a lesser extent, rab3A increase NE uptake at the level of the granule membrane.
rab3 Interacts Physically and Functionally with PI3K-To further characterize the mechanism by which rab3B regulates NE transport, we sought to identify putative rab3B effectors that might function in concert with rab3B to stimulate NE uptake. PI3K attracted our attention because of its role in vesicular trafficking and signal transduction; processes in which rab proteins and monomeric GTPases in general are thought to play important roles. In addition, we reported previously that rab3B expression in PC12 cells affects the profile of tyrosine phosphorylated proteins that co-immunoprecipitate with the 85-kDa regulatory subunit of PI3K (24). To determine whether there is an interaction between rab3B and PI3K, we immunoprecipitated rab3B from lysates of rab3B-expressing PC12 cells and tested for co-immunoprecipitated PI3K activity using an in vitro lipid kinase assay. Mock-transfected cells lacking detectable rab3B (8) served as negative controls. A rab3B-specific antibody (8, 16) precipitated 4-fold more PI3K activity from rab3B-expressing PC12 cells as compared with the mock transfected cells (Fig. 4A). To test whether native rab3B can also interact with PI3K in cells that normally express this GTPase, we also immunoprecipitated rab3B from HT29-CL19A colonic epithelial cells that express more endogenous rab3B than PC12 cells (8,16) and observed 10 -15 times more PI3K activity in these immunoprecipitates as compared with nonimmune controls (performed two times; results not shown). Unfortunately, our attempts to show an interaction between rab3A and PI3K using this strategy were unsuccessful because we could not immunoprecipitate rab3A from PC12 cells with any of three available antibodies.
To determine whether the association of rab3B with PI3K is mediated directly through one of the PI3K subunits or indirectly through an adaptor molecule, we performed a direct pairwise binding assay with GST-rab3B and recombinant p85  6). # indicates a statistically significant difference (p Ͻ 0.001). C, comparison of effects of reserpine on NE uptake by digitonin-permeabilized and intact rab3B-expressing cells. Data are mean Ϯ S.D. (n ϭ 6). # indicates a statistically significant difference (p Ͻ 0.001). D, rab3B and rab3A stimulate NE uptake at the granule surface. Values represent reserpine-sensitive NE uptake by digitonin-permeabilized cells calculated by subtracting NE uptake measured in the presence of reserpine from that measured in the absence of reserpine. The data points are means Ϯ S.E. (n ϭ 6). ** indicates a statistically significant difference (p Ͻ 0.01 relative to mock). rab3 and PI3K Stimulate Catecholamine Uptake regulatory subunit of PI3K (p85␣). The results of a typical experiment shown in Fig. 4B indicate that p85␣ binds to GST-rab3B in the presence of 500 M GTP␥S, a nonhydrolyzable analog of GTP. No p85 binding was detected with control GST protein either in the presence of GDP or GTP␥S or with GST-rab3B in the presence of 500 M GDP. The specificity of this interaction was assessed further by incubating GST-rab5 (an endosome-associated rab) (9, 10) with p85␣ PI3K under the same conditions. In this case we observed no binding in the presence of GTP and weak binding in the presence of GDP. (The binding of PI3K regulatory subunit to rab5 in the presence of GDP possibly relates to the fact that these two proteins coordinately regulate endosome fusion (Ref. 25).) In summary, the results of Fig. 4 show that rab3B can physically interact with the p85 regulatory subunit of PI3K and that this interaction is potentiated by GTP.
To evaluate the functional relevance of this interaction, we determined whether rab3B could modulate the lipid kinase activity of PI3K. Fig. 5A compares the PI3K activities in p85 and phosphotyrosine immunoprecipitates from lysates of mock transfected cells and rab3B-expressing cells. We observed a 2.5-and a 4.5-fold increase in PI3K activity in p85 and phosphotyrosine (PY20) immunoprecipitates, respectively, for rab3B-expressing PC12 cells compared with the mock transfected cells. PI3K activities assayed in other mock and rab3B clones were similar to those shown in Fig. 5A excluding the possibility of clonal variation (data not shown). No differences in p85 protein amount or tyrosine phosphorylation of p85 were detected in the lysates of the various clones (Fig. 5B). Thus, the simplest explanations of these results are that PI3K is associated to a greater extent with tyrosine phosphoproteins in rab3B-expressing cells and/or that it has a higher activity per mole in the rab3B cells. We also assayed PI3K activity in PY20 immunoprecipitates from a rab3A and a rab3BN135I clone, and observed that PI3K activity was also stimulated in these cells relative to the mock controls (results not shown). These results indicate that rab3 directly or indirectly stimulates PI3K activity in PC12 cells.
PI3K Regulates NE Uptake in Intact and Digitonin-permeabilized PC12 Cells-Because PI3K physically and functionally interacts with rab3B, we determined whether PI3K, like rab3B, can regulate NE uptake in PC12 cells. We used two structurally distinct inhibitors of PI3K activity, LY294002 and wortmannin, to test for an involvement of PI3K in the regulation of NE uptake. Each inhibitor produced a dose-dependent decrease in NE uptake in intact mock and rab3B-expressing cells (Fig. 6). The rab3B-expressing cells and the mocks exhibited similar decreases in NE uptake in response to either PI3K inhibitor. LY294002 at 50 M also reduced NE uptake by the rab3A and rab3BN135I clones (results not shown). These findings indicate that PI3K activity, like rab3, positively regulates NE uptake in intact PC12 cells, and that PI3K modulates NE transport even in the absence of rab3A or rab3B overexpression (i.e. in the mocks).
We also examined the effects of LY294002 on NE uptake over a range of NE concentrations to determine how PI3K affects uptake kinetics (Fig. 7). Compared with vehicle-treated controls, LY294002-treated rab3B-expressing cells showed a 2-fold FIG. 4. rab3B interacts physically with PI3K. A, rab3B antibody precipitates PI3K activity from rab3B-expressing PC12 cells. Anti-rab3B antibody was incubated with whole cell lysates from two mock clones and two rab3B clones and the immune complexes were subjected to an in vitro PI3K assay. The amount of phosphoinositide 3-phosphate formed in each sample was quantified by phosphorimage analysis. Results are expressed as the average of two separate experiments for each clone (Ϯ S.E.). To confirm that rab3B was precipitated under these conditions each immunoprecipitate was probed for the presence of rab3B by Western blotting (inset). # indicates a statistically significant difference (p Ͻ 0.001 relative to mock clones). B, rab3B binds directly to the PI3K regulatory subunit in a nucleotide-dependent manner. Immobilized recombinant GST, GST-rab3B, or GST-rab5 (1 g each) was incubated with recombinant p85␣ PI3K (1 g) in the presence of 0.5 mM GDP or GTP␥S. Bound proteins were resolved by to SDS-PAGE and analyzed for the presence of p85 PI3K by immunoblotting with a monoclonal antibody that recognizes the p85␣ isoform. This experiment was repeated twice with GST-rab5 and three times with GST-rab3B with similar results.

FIG. 5. PI3K activity is enhanced in rab3B PC12 cells.
A, comparison of the levels of PI3K activities in p85 and phosphotyrosine (PY20) immunoprecipitates between a mock and rab3B clone. The mean Ϯ S.D. of duplicate samples is shown. * indicates a statistically significant difference (p Ͻ 0.05 relative to mock). Similar results were obtained in eight separate experiments. PI3K activity was also stimulated in p85 and PY20 immunoprecipitates from rab3A and rab3BN135I-expressing PC12 cells relative to mock controls (results not shown). B, comparison of the amount and tyrosine phosphorylation of p85 between the mock and rab3B clones used in the experiments above. This experiment was repeated three times with similar results. rab3 and PI3K Stimulate Catecholamine Uptake diminution in apparent V max as well as a significant decrease in the apparent affinity for NE (Fig. 7 legend). Thus, the effects of the PI3K inhibitor on uptake kinetics were somewhat more complicated than that of rab3B expression, which significantly changed only the apparent V max (Fig. 1B). Given this more complicated effect of the PI3K inhibitor on uptake kinetics, we also counted the total numbers of LDCVs per cell and within 110 nm of the plasma membrane to determine whether this inhibitor had any effect on the number or distribution of LD-CVs within PC12 cells. Table I shows that treatment with LY294002 for 6 h had no apparent effect on LDCV number or distribution in PC12 cells.
Finally, we determined whether PI3K, like rab3B, influences NE uptake at the vesicle membrane by testing the effects of PI3K inhibitors on NE uptake in digitonin-permeabilized cells. Fig. 8 shows that NE uptake in digitonin-permeabilized cells was inhibited in the presence of LY294002 or wortmannin during the 30 min uptake period. The results of the PI3K inhibitor experiments indicate that PI3K activity is a significant factor in the control of monoamine uptake at the granule surface.

DISCUSSION
In an earlier study we observed that rab3 isoforms stimulate the steady-state accumulation of exogenous NE by PC12 cells (8). The mechanism underlying the rab3-induced increase in NE uptake was not addressed in the previous study and was the impetus for the present investigation. Neurotransmitter uptake can be regulated at different levels; thus, a focus of this FIG. 6. Dose-dependent decrease in NE uptake by PI3K inhibitors. A, effect of varying LY294002 concentration on NE uptake by intact mock and rab3B-expressing PC12 cells. LY294002 also reduced NE uptake by rab3A-and rab3BN135I-overexpressing PC12 cells (data not shown). B, effect of varying wortmannin concentration on NE uptake by intact PC12 cells. * and ** and # indicate statistically significant differences relative to vehicle control (p Ͻ 0.05, p Ͻ 0.01, and p Ͻ 0.001, respectively). Experiments were performed in triplicate and are representative of two independent experiments.   rab3 and PI3K Stimulate Catecholamine Uptake study was to identify the site of regulation by rab3 (i.e. plasma membrane or secretory granule). Because rab GTPases generally interact with accessory proteins to perform their function, an additional goal was to identify rab3 binding partners and to determine whether such binding partners also participate in the regulation of NE uptake.
The results of our kinetic analysis revealed that rab3B expression primarily increased the apparent V max of NE uptake by intact PC12 cells with no significant effect on the K m . The kinetic data could be well fit by single Michaelis-Menten functions suggesting that one of the two uptake steps (plasma membrane or LDCV membrane) was limiting. Because rab3B and rab3A also increased NE accumulation in digitonin-permeabilized cells, the simplest explanation of our kinetic data is that monoamine uptake into secretory granules is the ratelimiting step. NE uptake by permeabilized cells was ATP-dependent, reserpine-sensitive and nomifensine-insensitive, as expected for VMAT-mediated uptake at the LDCV membrane. Thus, our data strongly indicate that these rab3 GTPases modulate NE uptake at least in part at the level of the secretory granule membrane.
The increase in apparent V max induced by rab3B expression could be explained in one of several ways: (i) increased numbers of secretory granules, (ii) increased numbers of NE transporters (VMAT) per granule, and (iii) increased proton motive driving force (voltage plus pH gradient) for NE uptake into secretory granules mediated by VMAT (7). Our morphologic results indicate that neither rab3B (8) nor PI3K (Table I) appear to affect the numbers or distributions of secretory granules in PC12 cells. At present we cannot distinguish between the latter two possibilities, at least with respect to the long term effects of rab3B expression on the uptake process (see below for discussion of the effects of the PI3K inhibitors on NE uptake).
An additional goal of this study was to identify putative rab3B effectors and to determine whether they participate in the regulation of NE uptake. Like other GTPases, rab proteins perform their functions through interactions with effector molecules. Using two different biochemical assays, we have identified PI3K as a rab3B-binding protein. Other monomeric GTPases, including Ras, Rac, and cdc42, have been shown to interact with PI3K (27)(28)(29). Ras interacts with the p110 catalytic subunit, and Rac and cdc42 can modulate PI3K activity by interacting with the p85 regulatory subunit. A possibility that is consistent with the data presented here would be that PI3K functions as an effector for rab3B. In accordance with criteria that classify proteins as rab effectors, the current data demonstrate that: (a) PI3K (in particular, the p85␣ regulatory subunit) binds preferentially to the active, GTP-bound conformation of rab3B; (b) rab3B modulates the lipid kinase activity of PI3K; and (c) PI3K activity regulates the same cellular process regulated by rab3B, namely NE uptake into LDCVs.
Several observations made in this and a previous study (24) are consistent with the proposal that rab3B controls PI3K activity by facilitating its interaction with tyrosine phosphoproteins and/or through direct binding to the p85 regulatory subunit. There is evidence that the basal function of p85 is to inhibit the activity of the p110 catalytic subunit and that binding of p85 to tyrosine phosphoproteins overcomes this inhibition (30). We observed increased PI3K activity in phosphotyrosine immunoprecipitates from rab3B-and rab3A-expressing cells. This elevation in activity was not because of an increase in tyrosine phosphorylation of the p85 subunit, implying that rab3B potentiates the recruitment of PI3K to tyrosine phosphoproteins. Consistent with this observation, we previously reported that rab3B expression affected the profile of tyrosine-phosphorylated proteins that co-immunoprecipitate with p85 (24). A role for rab3 in facilitating the interaction of p85 with tyrosine phosphoproteins would result in activation of PI3K activity. The elevated lipid kinase activity measured in p85 immunoprecipitates is also consistent with the proposal that PI3K is more active in cells expressing rab3B.
Given that rab3B both interacts with PI3K and stimulates NE uptake, we investigated whether PI3K could also participate in the regulation of NE uptake. Our results obtained with the PI3K inhibitors, LY294002 and wortmannin, implicate PI3K activity in the regulation of NE uptake. These inhibitors have no effect on the efficiency of NE secretion induced by raising cytosolic Ca 2ϩ (Ref. 24, and results not shown); thus, the involvement of PI3K in NE transport appears to be limited to NE uptake. Presumably rab3B utilizes other effectors to regulate secretion (e.g. see Refs. [31][32][33]. Inhibiting PI3K had somewhat complicated effects on the kinetics of NE uptake involving changes in both the apparent K m and V max . Perhaps this indicates that PI3K regulates NE uptake in intact cells at multiple levels (i.e. plasma membrane and LDCV membrane). NE uptake in mock-transfected cells was also inhibited by the PI3K inhibitors indicating that a basal level of PI3K activity, possibly involving endogenous rab3A, positively regulates NE uptake.
We found that PI3K activity regulates NE uptake at least in part at the level of the granule membrane. Both PI3K inhibitors caused a substantial reduction in NE uptake by digitoninpermeabilized cells over a relatively short period of time (30 min of incubation). The rapid effect of both PI3K inhibitors on NE uptake would seem to rule out an effect on the numbers of secretory granules (Table I) or the numbers of VMAT molecules. As discussed above, it is conceivable that PI3K and rab3B regulate vesicular NE uptake by affecting one or both of the driving forces for uptake across the granule membrane (i.e. the transmembrane pH gradient and the transmembrane voltage). G␣o 2 inhibits NE uptake by PC12 cells by reducing the proton electrochemical gradient across the LDCV membrane (7). Future experiments will be required to determine whether rab3B and/or PI3K affect NE uptake across the LDCV membrane by this or other mechanisms.
In summary, we have shown that rab3 stimulates NE uptake at the level of the granule membrane; thus, this GTPase can regulate both limbs of NE transport (granule uptake (this study) and calcium-induced release (Ref. 8)). We identified PI3K as a rab3B binding partner, showed that PI3K activity is enhanced in cells expressing rab3, and observed that PI3K inhibitors reduced NE uptake in intact and permeabilized PC12 cells. Given that rab3B can interact with PI3K, it is reasonable to conclude that rab3B stimulates NE uptake, in part, through its interaction with PI3K. It is also possible that rab3B can affect uptake by mechanisms not involving PI3K, as suggested by the fact that NE uptake was still elevated in rab3B-expressing cells relative to mock cells, even in the presence of the highest concentrations of PI3K inhibitors that were tested (Fig. 7). As PI3K can participate in multiple signaling pathways, the results of this and future studies could provide novel insights into the regulation of NE uptake in neuroendocrine cells and perhaps neurons in various physiological settings.