Report

Signaling by Phosphoinositide-3,4,5-Trisphosphate Through Proteins Containing Pleckstrin and Sec7 Homology Domains

See allHide authors and affiliations

Science  28 Mar 1997:
Vol. 275, Issue 5308, pp. 1927-1930
DOI: 10.1126/science.275.5308.1927

Abstract

Signal transmission by many cell surface receptors results in the activation of phosphoinositide (PI) 3-kinases that phosphorylate the 3′ position of polyphosphoinositides. From a screen for mouse proteins that bind phosphoinositides, the protein GRP1 was identified. GRP1 binds phosphatidylinositol-3,4,5-trisphosphate [PtdIns(3,4,5)P3] through a pleckstrin homology (PH) domain and displays a region of high sequence similarity to the yeast Sec7 protein. The PH domain of the closely related protein cytohesin-1, which, through its Sec7 homology domain, regulates integrin β2 and catalyzes guanine nucleotide exchange of the small guanine nucleotide-binding protein ARF1, was also found to specifically bind PtdIns(3,4,5)P3. GRP1 and cytohesin-1 appear to connect receptor-activated PI 3-kinase signaling pathways with proteins that mediate biological responses such as cell adhesion and membrane trafficking.

Cellular 3-phosphoinositides are generated through the actions of a family of PI 3-kinases and appear to have regulatory roles in multiple cell functions. In yeast the Vps34 gene product, a PI 3-kinase with substrate specificity restricted to phosphatidylinositol (PtdIns), is required for correct sorting of carboxypeptidase Y to the vacuole (1). In mammalian cells three classes of PI 3-kinases have been identified in addition to a Vps34 homolog (2). These include p110 isoforms regulated by p85 subunits containing SRC homology 2 (SH2) domains (3), a p110γ PI 3-kinase regulated by heterotrimeric guanine nucleotide-binding proteins (4), and a PI 3-kinase containing a homology domain (C2 domain) thought to bind membrane lipids (5). The p85-p110 and p110γ type PI 3-kinases that are specifically activated by receptor signaling systems exhibit broad substrate specificities, and their activation leads to rapid phosphorylation of the inositol D-3 positions on PtdIns, PtdIns-4-phosphate [PtdIns(4)P], and PtdIns-4,5-bisphosphate [PtdIns(4,5)P2] (6). Signaling by these 3-polyphosphoinositides appears to regulate such diverse cellular responses as membrane ruffling (7), chemotaxis (8), secretory responses (9), insulin-mediated membrane translocation of glucose transporters (10), membrane trafficking of growth factor receptors (11), and regulated cell adhesion (12).

Several protein targets of 3-phosphoinositides have been identified. These include certain protein kinase C isoforms (13) and the pleckstrin homology (PH) domain-containing protein kinases c-Akt and Btk (14). A series of sequentially activated protein kinases stimulated in response to 3-phosphoinositides causes activation of the p70 ribosomal protein S6 kinase (15). Additionally, 3-phosphoinositides appear to bind certain SH2 domains, including those within the p85 regulatory subunit of p110 PI 3-kinase (16). Thus, the phosphoinositides may serve as membrane localization elements that recruit target proteins to specific cellular organelles (17).

To identify general receptors for phosphoinositides (GRPs), we developed an expression library screening procedure using bovine brain phospholipids labeled with [γ-32P]adenosine triphosphate (ATP) in the presence of the glutathione-S-transferase (GST) fusion protein GST-p110 PI 3-kinase (18). High-performance liquid chromatography (HPLC) analysis of the reaction products confirmed the presence of 32P-labeled PtdIns(3)P, PtdIns(3,4)P2, and PtdIns(3,4,5)P3 in this probe mixture. We screened mouse 3T3-F442A adipocyte and brain cDNA expression libraries with the labeled brain phosphoinositides (Fig. 1). A single clone from each cDNA library screen reproducibly bound the 32P label from the probe mixture upon subcloning and plaque purification (Fig. 1A). The proteins encoded by both brain and adipocyte cDNA clones bound [32P]PtdIns(3,4,5)P3 but not [32P]PtdIns(3,4)P2 or [32P]PtdIns(3)P under the experimental conditions of the library screening procedure (Fig. 1B).

Fig. 1.

Expression cloning of GRP1 cDNA. (A) Autoradiographs of nitrocellulose filters at different stages of purification of the cDNA clone identified in a mouse 3T3-F442A adipocyte cDNA expression library. Filters were incubated with mixed brain phosphoinositides labeled at the 3′ position with p110 PI 3-kinase and [γ-32P]ATP, and then washed. The primary screen was performed in 15-cm dishes and subsequent screens in 10-cm dishes. (B) Binding specificity of the isolated cDNA clone. About 1000 pfu of the cDNA library (Con.) or the isolated cDNA clone (GRP1) were spotted on plates containing a layer of Escherichia coli and incubated with nitrocellulose filters as described. The filters were incubated with 0.5 × 106 cpm of either [32P]PtdIns(3)P, [32P]PtdIns(3,4)P2, or [32P]PtdIns(3,4,5)P3 (PtdIns is abbreviated in labels as PI in all relevant figures), processed as described (18), and subjected to autoradiography and densitometry. The values are the means of four experiments, and the error bars represent the standard deviations. Similar results were obtained with a mouse brain cDNA expression library.

Both the brain and adipocyte cDNA clones encoded amino acid sequences of the same protein, GRP1. Standard hybridization techniques were used to obtain additional cDNA clones of this species that encode four more amino acids including a putative NH2-terminal methionine, and a full-length sequence was deduced (Fig. 2). Database searches showed GRP1 to be highly similar to a protein encoded by the human cDNA B2-1, originally cloned from cytolytic natural killer T lymphocytes (19). Both GRP1 and the B2-1 protein contain PH domains (20) and Sec7 homology regions (21, 22) (Fig. 2A). The B2-1-encoded protein has been called cytohesin-1 on the basis of its binding to the integrin β2 cytoplasmic domain through this Sec7 homology region (23), and the Sec7 domain of an isoform denoted ARNO (for ARF nucleotide binding site opener) has nucleotide exchange activity for the small guanine nucleotide-binding protein adenosine diphosphate-ribosylation factor 1 (ARF1) (23). A partial sequence of another probable isoform (Cts18) has been reported (22). The divergent amino acid sequences between GRP1, ARNO, cytohesin-1, and Cts18 appear not to result from species variation because a human expressed sequence tag (EST) found in the Institute for Genomic Research database shows a predicted amino acid sequence identical to that of mouse GRP1 between residues 34 and 120 (Fig. 2B). The sequence of the human EST differs from those of ARNO, cytohesin-1, and Cts18. Because there are additional EST sequences in the database, it is likely that more human isoforms of these proteins exist.

Fig. 2.

Structure of GRP1. (A) Overall structure of GRP1 and cytohesin-1 (Ch-1). (B) Comparison of the deduced sequences of GRP1, B2-1/cytohesin-1, and EST 01394. The region corresponding to the Sec7 domain is boxed with a solid line, and the region corresponding to the PH domain is boxed with a dashed line. Sequence similarity between GRP1 and cytohesin-1 is 88%, and that between the Sec7 and PH domains of these two proteins is 93% and 94%, respectively. Single-letter abbreviations for the amino acid residues are as follows: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; and Y, Tyr.

To determine whether PtdIns(3,4,5)P3 binding could be localized to specific sequences within the GRP1 structure, we expressed GST fusion proteins containing either residues 239 to 399 (PH domain), residues 52 to 260 (Sec7 domain), residues 5 to 71, or residues 5 to 399 of GRP1 in bacteria and purified them on glutathione-conjugated beads (Fig. 3A). The nearly full-length GRP1 or its PH domain bound [32P]PtdIns(3,4,5)P3 (24), whereas the fusion proteins containing the Sec7 homology region or residues 5 to 71 did not (Fig. 3A). The GRP1 PH domain associated with [32P]PtdIns(3,4,5)P3 specifically and failed to bind [32P]PtdIns(3,4)P2 or [32P] PtdIns(3)P (Fig. 3B). PH domains from PLC- δ (25) and various other proteins (26) bind PtdIns(4,5)P2, but do not preferentially associate with a 3-phosphoinositide. In our [32P]polyphosphoinositide binding assay, binding of [32P]PtdIns(3,4,5)P3 to GST fusion proteins containing the PH domains of the Son of Sevenless (SOS) protein and insulin receptor substrate (IRS)-1 was negligible (Fig. 4B).

Fig. 3.

Localization of PtdIns(3,4,5)P3 binding to the PH domain of GRP1. (A) Identification of the binding domain for PtdIns(3,4,5)P3 in GRP1. Proteins (150 pmol) were spotted on nitrocellulose filters and incubated with [32P]PtdIns(3,4,5)P3 and the amount of bound lipid was determined as described (24). GST-N contains residues 5 to 71 of GRP1. The values are the means of quadruplicate determinations, and the error bars are the standard deviations. (B) Binding specificity of the PH domain of GRP1. Nitrocellulose filters with GST-PH were incubated with 25,000 cpm of each of the phospholipids, and the amount of bound lipid determined as described (24). The values are the means of quadruplicate determinations and the error bars are the standard deviations.

Fig. 4.

Effect on binding of [32P]- PtdIns(3,4,5)P3 to GST-PH by inositol polyphosphates. (A) [32P]PtdIns- (3,4,5) P3 (25,000 cpm) was incubated with various concentrations of Ins(1,3,4,5)P4, and bound radioactivity was determined and expressed as the percent of that bound in the absence of competitor. The values are means of quadruplicate measurements, and the error bars are standard deviations. (B) Specificity of [32P]PtdIns(3,4,5)P3 binding to the PH domains of GRP1 and cytohesin-1 (Ch-1). The binding to GST fusion proteins containing amino acids 462 to 569 of murine SOS or amino acid 13 to 115 of murine IRS-1 was assayed with 150 pmol of protein bound to the filters. Competition of binding to 7.5 pmol of GST-PH fusion proteins was measured in the presence of 100 μM of the competitors. The values were calculated as in (A) and are means of quadruplicate measurements and the error bars represent standard deviations. (C) Comparison of the amino acid sequences of the PH domains of murine GRP1, B2-1/cytohesin-1, SOS-1, and IRS-1. The basic motifs in the NH2-terminal parts of the GRP1 and B2-1/cytohesin-1 molecules are outlined with heavy boxes and with asterisks. Light boxes delineate amino acid identities among the PH domains.

The specificity of the GRP1 PH domain for binding [32P]PtdIns(3,4,5)P3 was also examined with respect to competition by the inositol phosphate head groups of the phosphoinositides. Unlabeled inositol-1,3,4,5-tetraphosphate [Ins(1,3,4,5)P4] (100 μM) completely inhibited [32P]PtdIns(3,4,5)P3 binding to GST-PH(GRP1), whereas Ins(1,4,5)P3, Ins(1,3,4,6)P4, and Ins(1,2,5,6)P4 had little or no effect at the same concentration (Fig. 4B) even though the latter two carry the same charge as the inhibitor Ins(1,3,4,5)P4. The concentration of Ins(1,3,4,5)P4 that inhibited binding of [32P]PtdIns(3,4,5)P3 half maximally was about 3 to 8 μM (Fig. 4A). Taken together, the data in Figs. 3 and 4 indicate extraordinary specificity of the GRP1 PH domain for PtdIns(3,4,5)P3. A GST fusion protein of the cytohesin-1 PH domain also bound [32P]PtdIns(3,4,5)P3 (Fig. 4B), but not [32P]- PtdIns3P or [32P]PtdIns(3,4)P2 (27). Analysis of the NH2-terminal amino acid sequences of the GRP1 and cytohesin-1 PH domains in comparison with those of the IRS-1 and SOS PH domains (Fig. 4C) shows an additional lysine at position 273 and a Lys282-Arg283-Arg284 motif in the NH2-terminal region, a region important for 4,5-polyphosphoinositide binding in other PH domains (26).

The selectivity of the GRP1 and cytohesin-1 PH domains for binding PtdIns(3,4,5)- P3 indicates that the PH domain may function in the recruitment of these proteins to sites of PtdIns(3,4,5)P3 synthesis in response to the action of receptor-regulated p110-type PI 3-kinases. Such recruitment to specific cell membrane regions would define their sites of action mediated presumably through their Sec7 domains. One function of the cytohesin-1 Sec7 domain is enhancement of cellular adhesion through direct association with the cytoplasmic region of integrin β2 (22). Transfection of Jurkat cells with the B2-1/cytohesin-1 cDNA or cDNA encoding this Sec7 homology domain alone enhanced their adhesion to ICAM-1, a ligand of integrins containing the β2 polypeptide. Our data thus suggest a molecular basis for regulation of integrin through PI 3-kinase (Fig. 5).

Fig. 5.

Model of the mechanism by which GRP1 family proteins link receptor-activated PI 3-kinase signaling pathways to ARF1 and integrin β2 responses. Receptor tyrosine kinases (RTKs) recruit p85-p110-type PI 3-kinases to tyrosine phosphate sites, promoting the generation of 3,4,5-phosphoinositide, which binds the PH domain of GRP1 or cytohesin-1. Membrane-bound cytohesin-1 or GRP1 may interact through their Sec7 homology regions either with ARF1 to cause guanine nucleotide exchange, or with integrin β2 to modulate cell adhesion.

Another reported function of the Sec7 homology domain within cytohesin-1 and the similar protein ARNO is the catalysis of guanine nucleotide exchange on ARF1 (23). Regulation of ARF proteins by a PI 3-kinase-mediated pathway has been previously suggested on the basis of morphological data (28). Our demonstration that the cytohesin-1 PH domain is a target for PtdIns(3,4,5)P3 (Fig. 4B) suggests a model in which receptor-activated PI 3-kinase generates PtdIns(3,4,5)- P3 to localize cytohesin-1, which in turn can regulate the guanine nucleotide exchange of ARF1 (Fig. 5). The PH domains of GRP1, cytohesin-1, and ARNO exhibit very high sequence similarity. Thus, the PH domain of ARNO may also bind PtdIns(3,4,5)P3. This family of proteins appears to mediate the regulation of protein sorting and membrane trafficking by PtdIns(3,4,5)P3.

REFERENCES AND NOTES

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
View Abstract

Navigate This Article