Regulation of GRP1-catalyzed ADP Ribosylation Factor Guanine Nucleotide Exchange by Phosphatidylinositol 3,4,5-Trisphosphate*

Cellular levels of phosphatidylinositol 3,4,5-trisphosphate (PtdIns(3,4,5)P3) are rapidly elevated in response to activation of growth factor receptor tyrosine kinases. This polyphosphoinositide binds the pleckstrin homology (PH) domain of GRP1, a protein that also contains 200 residues with high sequence similarity to a segment of the yeast Sec7 protein that functions as an ADP ribosylation exchange factor (ARF) (Klarlund, J., Guilherme, A., Holik, J. J., Virbasius, J. V., Chawla, A., and Czech, M. P. (1997)Science 275, 1927–1930). Here we show that dioctanoyl PtdIns(3,4,5)P3 binds the PH domain of GRP1 with aK d = 0.5 μm, an affinity 2 orders of magnitude greater than dioctanoyl-PtdIns(4,5)P2. Further, the Sec7 domain of GRP1 is found to catalyze guanine nucleotide exchange of ARF1 and -5 but not ARF6. Importantly, PtdIns(3,4,5)P3, but not PtdIns(4,5)P2, markedly enhances the ARF exchange activity of GRP1 in a reaction mixture containing dimyristoylphosphatidylcholine micelles, 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid, and a low concentration of sodium cholate. PtdIns(3,4,5)P3-mediated ARF nucleotide exchange through GRP1 is selectively blocked by 100 μm inositol 1,3,4,5-tetrakisphosphate, which also binds the PH domain of GRP1. Taken together, these data are consistent with the hypothesis that selective recruitment of GRP1 to PtdIns(3,4,5)P3 in membranes activates ARF1 and -5, known regulators of intracellular membrane trafficking.

Recent work in this laboratory identified a potential effector protein, GRP1, containing a PH domain that binds PtdIns-(3,4,5)P 3 but not Ptd(3,4)P 2 or PtdIns(3)P (14). GRP1 also contains a 200-amino acid sequence with high similarity to a domain in the yeast Sec7 protein, which is known to be required for transport of polypeptides from the endoplasmatic reticulum to and through Golgi membranes (15). A similar Sec7 domain in the mammalian proteins cytohesin-1 and ARNO and in the yeast protein Gea1 were reported to catalyze guanine nucleotide exchange on the small GTP-binding protein ARF1 (16 -18). ARF proteins in the GTP-bound form promote membrane vesicle trafficking pathways by recruiting coat proteins to membranes, causing membrane budding (19 -21). A potential interaction between PI 3-kinase and ARF mutants had been suggested based on characteristic changes of endosome morphology elicited by wortmannin in various cultured cell lines (12). Taken together, these observations provide a framework for the hypothesis that GRP1 or its homologs may mediate cellular effects of 3Ј-polyphosphoinositides that require ARF proteins.

EXPERIMENTAL PROCEDURES
Recombinant GRP1 Protein-To map the active regions of GRP1, GST proteins fused to various domains of GRP1 were purified from lysates of bacteria expressing the relevant pGEX5X-3 constructs as described (14). For other experiments, GRP1 was cloned into pGEX-4T, and the recombinant protein was purified. The GST-GRP1 fusion protein was cleaved by incubation with 5 g/ml thrombin in 200 l of 20 mM Tris, pH 8.0, 2.5 mM CaCl 2 , 150 mM NaCl overnight at 4°C. Complete cleavage was verified by SDS-polyacrylamide gel electrophoresis. The proteins were transferred to assay buffer (50 mM HEPES, pH 7.5, 1 mM MgCl 2 , 100 mM KCl, 1 mM dithiothreitol) using Centricon 30 microconcentrators (Amicon) and stored at Ϫ20°C in 50% glycerol.
ARF Exchange Assay-Recombinant baculovirus encoding ARF1, -5, or -6 (ARF cDNAs kindly provided by Dr. R. Klausner) fused at the C terminus to a 9-amino acid sequence corresponding to the major antigenic determinant of influenza virus hemagglutinin was constructed. Sf9 cells were infected with the recombinant baculovirus, the cells were harvested 3 days later by centrifugation for 5 min at 3000 rpm, and the pellets were stored at Ϫ70°C. A cell pellet corresponding to 50 ml of culture medium was dissolved in 1 ml of assay buffer supplemented with 1% Triton X-100, 1 mM benzamidine, 5 g/ml leupeptin, 5 g/ml aprotinin, 1 mM phenylmethylsulfonyl fluoride, and 50 M GDP. After clarification by centrifugation at 20,000 rpm for 5 min, 50 l of a rabbit antiserum that had been produced by immunization with a peptide (YPYDVPDYA) conjugated to hemocyanin was added and incubated overnight on ice. The following day, 75 l of Protein A conjugated to Sepharose CL-4B (Sigma) was added and incubated on an end-over-end mixer for 1 h. The beads were collected by centrifugation and washed five times with 1 ml of assay buffer. An additional 500 l of Sepharose CL-4B was added as a carrier. Analysis of bound ARF from Sf9 cells labeled with [ 3 H]myristic acid by Western blotting and autoradiography revealed that at least 90% of the ARF molecules was modified by the lipid (22).
For the standard assay, 5 l of assay buffer containing GRP1, 6 mM dimyristoylphosphatidylcholine (Avanti Polar Lipids Inc.), 0.2% cholate, 1% bovine serum albumin, 10 M GTP␥S, and 0.75 Ci of [ 35 S]GTP␥S was added to 5-l beads containing immunoabsorbed ARF. Approximately 50% of the cholate molecules are incorporated into the micelles under these conditions, based on the rate of diffusion of [ 3 H]cholate through a dialysis membrane in the presence versus absence of phospholipid micelles. 2 After 40 min at room temperature, the beads were washed four times with 1 ml of assay buffer, and 35 S was quantitated by liquid scintillation counting. Synthetic dipalmitoyl 3Јphosphoinositides were from Matreya Inc., and bovine brain PtdIns(4)P and PtdIns(4,5)P 2 were from Calbiochem.
Lipid Binding and Competition Assays-Binding assays were performed as described (23). Briefly, GST fused to the PH region of GRP1 (amino acids 239 -399) was bound to glutathione immobilized on agarose beads (Sigma). Synthetic 3 H-labeled dioctanoyl PtdIns(3,4,5)P 3 (kindly provided by C.-S. Chen (24)) was added, and after 1 h the beads were separated from the supernatant by centrifugation. The amount of [ 3 H]PtdIns(3,4,5)P 3 bound to the PH domain was calculated by subtracting the amount of free 3 H present in supernatants from GST-GRP1 from the amount of free 3 H in supernatants from control GST. The data were fitted to the equation, , by least squares curve fit. For the competition assays the beads containing the GST fusion proteins were incubated with 2.5 M 3 H-labeled dioctanoyl PtdIns(3,4,5)P 3 in the presence of different concentrations of unlabeled lipids. After 1 h of incubation at room temperature, the beads were washed with buffer containing 0.5% Nonidet P-40 and counted in a scintillation counter. The percentage of [ 3 H]PtdIns(3,4,5)P 3 bound was calculated based on the amount of 3 H bound to the beads in the absence of competitor. The total 3 H bound to the GST control was less than 0.5% of the total 3 H bound to GRP1 PH domain. The data were fitted to the equation, % bound ϭ 100 Ϫ n ϫ L/(K I(app) ϩ L), where n is the percent specific binding, L is the concentration of unlabeled lipid added, and K I(app) is the apparent competitive dissociation constant. The ratios of the apparent dissociation constants accurately reflect the ratios of the true dissociation constants under the present experimental conditions (23).

RESULTS
To test whether GRP1 functions as an ARF guanine nucleotide exchange factor, hemagglutinin-tagged ARF1, -5, or -6 proteins produced in Sf9 cells were incubated with recombinant GRP1 protein in the presence of [ 35 S]GTP␥S for various times prior to determination of ARF-bound label. Fig. 1 shows a nearly linear rate of labeled guanine nucleotide binding to these ARF proteins during this time course in the presence or absence of GRP1. A 4 -10-fold stimulation of ARF1 binding to [ 35 S]GTP␥S was observed in response to GRP1, whereas no effect of GRP1 on ARF6 binding to nucleotide was detected. Up to 10 molecules of [ 35 S]GTP␥S could be bound to ARF1 per molecule of GRP1 under our experimental conditions, showing that GRP1 acts catalytically. Guanine nucleotide exchange on ARF5 was also significantly enhanced by GRP1, although to a lesser extent than that for ARF1. Fig. 1D indicates that GRP1 maximally activates labeled nucleotide binding to ARF1 and -5 when present at 2.5 M under the conditions of this assay.
Because the Sec7 homology domains of the related protein ARNO (16) have been shown to be sufficient for catalysis of guanine nucleotide exchange of ARF1, we analyzed the ARF1 exchange activity intrinsic to each of three segments of GRP1, including the Sec7 homology region (Fig. 2). GST fusion proteins of the N terminus and PH regions of GRP1 exhibited no effect on ARF1 binding to labeled nucleotide in our assay. In contrast, a GST fusion protein containing the Sec7 homology domain was as effective as the full-length GRP1 in catalyzing ARF1 exchange activity (Fig. 2).
The data depicted in Fig. 3 suggested that high specificity of PtdIns(3,4,5)P 3 -mediated activation of ARF exchange may characterize GRP1 function. We tested the specificity of GRP1stimulated ARF exchange by these phosphoinositides under conditions of various charge densities in detergent-phosphatidylcholine micelles (Fig. 4). Interestingly, when a relatively high concentration of negative charge is present in the micelles (0.1% cholate), PtdIns(3,4,5)P 3 , PtdIns(4,5)P 2 , and PtdIns(3,4)P 2 all stimulated ARF1 exchange activity in the presence of GRP1. In the absence of charge (0.1% CHAPS, no cholate), none of these phosphoinositides were able to enhance the ARF exchange activity of GRP1 (Fig. 4). In contrast, PtdIns(3,4,5)P 3 selectively stimulated [ 35 S]GTP␥S binding to ARF1 in the presence of GRP1 when a low level of negative charge (0.05% cholate) was present with the phosphatidylcholine in our assay. Under these conditions PtdIns(4,5)P 2 and PtdIns(3,4)P 2 had significantly less effect. These experiments establish an in vitro assay system that reveals selective regulation of GRP1-catalyzed ARF1 guanine nucleotide exchange activity by PtdIns(3,4,5)P 3 .
Previous results (16,25,26) and data in Fig. 4 suggest that guanine nucleotide exchange of ARF proteins require an interface of ARF, exchange factor, and negatively charged phospholipid, indicating that recruitment of GRP1 to the phospholipid membrane may be a key element of the regulation of ARF guanine nucleotide exchange. Thus we reasoned that Ins(1,3,4,5)P 4 , the polar head group of PtdIns(3,4,5)P 3 , should compete for binding and recruitment of GRP1 to PtdIns(3,4,5)P 3 in micelles, thus blocking GRP1-mediated ARF1 exchange. Fig.  5 shows results that are consistent with this prediction. Addition of 100 M Ins(1,3,4,5)P 4 to our assay virtually ablated the increased ARF1 guanine nucleotide exchange activity catalyzed by GRP1 in response to PtdIns(3,4,5)P 3 . In concert with data in Fig. 4 indicating high binding specificity of the GRP1 PH domain, other polar head groups tested, including Ins(1,3,4)P 3 , Ins(1,4,5)-P 3 , Ins(1,3,4,6)P 4 , and Ins(1,2,5,6)P 4 , failed to inhibit GRP1mediated ARF1 guanine nucleotide exchange (Fig. 5). DISCUSSION The results presented here demonstrate the specific binding and regulation of GRP1-catalyzed ARF1 guanine nucleotide exchange by PtdIns(3,4,5)P 3 , a product of receptor-regulated PI 3-kinase activity (Figs. 2, 4, and 5). These experiments (Fig. 2) also localize ARF exchange activity to an approximately 200residue segment of the GRP1 protein that exhibits high sequence similarity to a region of the Saccharomyces cerevisiae Sec7 protein (27). GRP1 was found to catalyze nucleotide ex- change on ARF1 and -5 but not ARF6 (Fig. 1). Interestingly, ARF5 was reported not to be a substrate for cytohesin-1 (17), but a detailed description of the specificities of the various isoforms of GRP1 is still lacking.
The identification of GRP1 based on its binding to PtdIns(3,4,5)P 3 in an expression cDNA library screen suggested a particularly high affinity for this polyphosphoinositide (14). Our present results confirm this expectation, revealing a K d ϭ 0.5 M for binding [ 3 H]dioctanoyl PtdIns(3,4,5)P 3 (Fig. 2). Importantly, the PH domain of GRP1 exhibits an apparent affinity for dioctanoyl PtdIns(4,5)P 2 that is about 2 orders of magnitude lower than that for dioctanoyl PtdIns(3,4,5)P 3 (Fig.  2). This extraordinary specificity for the 3Ј-polyphosphoinositide is a distinctive characteristic of the GRP1 PH domain among the many PH domains that have been characterized for phosphoinositide binding (28). Furthermore, this degree of specificity is consistent with that required of a protein regulated in intact cells by PtdIns(3,4,5)P 3 signaling, given that PtdIns(4,5)P 2 is much more abundant in cell membranes (6). Taken together, the data in Figs. 2-5 provide strong support for the hypothesis that GRP1 functions in intact cells in conjunction with membrane-localized PtdIns(3,4,5)P 3 generated by receptor-activated PI 3-kinase activity.
Particular attention in this study was focused on the relationship between the specificity of PtdIns(3,4,5)P 3 binding to GRP1 and the specificity of PtdIns(3,4,5)P 3 -mediated activation of GRP1 exchange activity for ARF1 (Figs. 2 and 4). Previous work has emphasized the requirements for ARF myristoylation and phospholipid to obtain optimal nucleotide exchange rates for ARF proteins, suggesting exchange occurs at the membrane surface (26). In the case of ARNO, PtdIns(4,5)P 2 was found to be sufficient for its recruitment to membranes as well as activation of ARF exchange (16), suggesting specificity of its PH domain for binding this phosphoinositide or a nonspecific effect of PtdIns(4,5)P 2 to recruit ARNO to membranes based on a charge effect, or both. All three polyphosphoinositides tested were effective in stimulating GRP1-mediated exchange activity when a high charge density was present on micelles in the assay (Fig. 4), indicating that GRP1 can be nonspecifically bound to membranes under these conditions. Selectivity of PtdIns(3,4,5)P 3 in stimulating ARF1 nucleotide exchange by GRP1 over that observed for PtdIns(4,5)P 2 or PtdIns(3,4)P 2 was revealed at lower charge density (Fig. 4). These data are consistent with the hypothesis that binding of GRP1 to membranes containing ARF may include two components: specific interaction with PtdIns(3,4,5)P 3 through its PH domain and interactions with acidic phospholipids through one or more clusters of its basic amino acid residues. At concentrations of GRP1 that yield maximal [ 35 S]GTP␥S loading of ARF1, no stimulation by PIP 3 was observed, consistent with the concept that activation of GRP1 exchange activity results from recruitment to membranes (data not shown).
The high activity observed for ARF1 guanine nucleotide exchange catalyzed by the GRP1 Sec7 domain suggests that ARF1 function may be closely related to the physiological role of GRP1 (Fig. 1). ARF1 has been implicated in vesicle transport related to Golgi membrane function as well as in secretory and exocytosis pathways (19,21). It is noteworthy that several targets of receptor-activated PI 3-kinase signaling involve intracellular membrane-trafficking systems, including mast cell secretion and insulin-sensitive GLUT4 glucose transporter translocation to the plasma membrane in muscle and adipocytes (9,11,29). Thus it is tempting to hypothesize a role for GRP1 in regulating such processes in response to localized synthesis of PtdIns(3,4,5)P 3 .