Review
Structure and function of phosphoinositide 3-kinases

https://doi.org/10.1016/S0005-2760(98)00139-8Get rights and content

Abstract

Phosphoinositide kinases (PI3Ks) play an important role in mitogenic signaling and cell survival, cytoskeletal remodeling, metabolic control and vesicular trafficking. Here we summarize the structure–function relationships delineating the activation process of class I PI3Ks involving various domains of adapter subunits, Ras, and interacting proteins. The resulting product, PtdIns(3,4,5)P3, targets Akt/protein kinase B (PKB), Bruton’s tyrosine kinase (Btk), phosphoinositide-dependent kinases (PDK), integrin-linked kinase (ILK), atypical protein kinases C (PKC), phospholipase Cγ and more. Surface receptor-activated PI3Ks function in mammals, insects, nematodes and slime mold, but not yeast. While many members of the class II family have been identified and characterized biochemically, it is presently unknown how these C2-domain containing PI3Ks are activated, and which PI substrate they phosphorylate in vivo. PtdIns 3-P is produced by Vps34p/class III PI3Ks and operates via the PtdIns 3-P-binding proteins early endosomal antigen (EEA1), yeast Vac1p, Vps27p, Pip1p in lysosomal protein targeting. Besides the production of D3 phosphorylated lipids, PI3Ks have an intrinsic protein kinase activity. For trimeric GTP-binding protein-activated PI3Kγ, protein kinase activity seems to be sufficient to trigger mitogen-activated protein kinase (MAPK). Recent disruption of PI3K genes in slime mold, Caenorhabditis elegans, Drosophila melanogaster and mice further underlines the importance of PI3K signaling systems and elucidates the role of PI3K signaling in multicellular organisms.

Introduction

Phosphoinositides were recognized early as precursors for second messengers in cell surface receptor-coupled signal transduction pathways. By the action of various phospholipases C (PLC), PtdIns(4,5)P2 is hydrolyzed to Ins(1,4,5)P3, releasing calcium from internal stores [1] and diacylglycerol activating protein kinases C (reviewed in [2], [3], [4]). In the mid-1980s, a phosphoinositide kinase activity was found to associate with polyoma middle T antigen, pp60v-src and pp68v-ros [5], [6], [7], and with anti-phosphotyrosine immunoprecipitates from PDGF-stimulated fibroblasts [8]. This activity was found to phosphorylate the D3-hydroxy-group of the inositol headgroup of phosphoinositides and thus generated a previously unknown class of lipids ([9], [10], reviewed in [11]).

The observation that increased levels of 3-phosphorylated phosphoinositides were found in transformed and/or mitogen-stimulated cells implied an involvement of PI3K in oncogenic and mitogenic signal transduction [12]. This was supported by the finding that mutants of polyoma middle T and pp60v-src, which are unable to associate with PI3K, also failed to transform cells [5], [13]. Moreover, transforming mutants of pp60c-src resulted in an association of pp60c-src with PI3K activity (see [14] and [11], [15] for a contemporary review).

The cloning of two isoforms of the regulatory subunit of PI3K, p85α and p85β [16], [17], [18], revealed that the molecules each contained two Src homology 2 (SH2) domains that mediate the interaction with tyrosine phosphorylated pp60v-src and pp68v-ros or autophosphorylated PDGF receptor. Finally, the identification of p85-associated p110α and p110β subunits [19], [20] confirmed that the catalytic activity was intrinsic to the larger subunit of the prototypic p85/p110 PI3K heterodimer. The Saccharomyces cerevisiae VPS34 gene product Vps34p, essential for vacuolar protein sorting and vacuole segregation [21], was subsequently shown to possess PI3K activity and was thus the first PI3K with an assigned function [22].

Numerous PI3Ks have been identified by biochemical approaches and PCR-based up-to-date cloning strategies. PI3Ks have been isolated from many eukaryotes, including mammals, yeast, flies, slime mold, plants and algae. The integration of sequence data, binding of adaptor proteins and PI substrate specificity permitted a segregation of PI3Ks into three classes [23], [24], [25], [26]. In this review, we summarize the structural data available for PI3Ks and match it with the mechanisms exploited by PI3Ks to relay signals from cell surface receptors.

Section snippets

The PI3K family

PI3Ks catalyze the transfer of the γ-phosphate group of ATP to the D3 position of phosphoinositides (for the inositol head-group nomenclature see [27]). Based on their selective in vitro substrate specificity, PI3Ks can be grouped into three classes: class I PI3Ks identified up to now are heterodimers of approximately 200 kDa, composed of a 110–120 kDa catalytic subunit and a 50–100 kDa adaptor subunit and are capable to phosphorylate in vitro, PtdIns, PtdIns 4-P and PtdIns(4,5)P2. The

Lipid products as mediators of PI3K downstream signaling

As discussed above, ligand receptor interactions trigger a rapid rise of cellular PtdIns(3,4,5)P3, and with some delay, PtdIns(3,4)P2, the latter generated by the action of phosphoinositide 5-phosphatases (for a review on PI phosphatases see Ref. [194]). PtdIns(3,4,5)P3 is synthesized directly from PtdIns(4,5)P2 [28], [29], but it has been recognized that the initially proposed route via PtdIns→PtdIns 4-P→PtdIns(4,5)P2→PtdIns(3,4,5)P3 is trespassed by all possible sequences of phosphorylations

Is PI3K only a lipid kinase?

When is a lipid kinase not a lipid kinase? When it is a protein kinase is the title of a review of Hunter [31] and seems to apply handily to answer the above question. Taken literally, class I and III kinases would then all target protein substrates, as their intrinsic protein Ser kinase activity has been demonstrated conclusively [84], [85], [86], [87], [171].

However, while a large body of evidence illustrates the actions of phosphorylated lipid products emerging from PI3Ks, there is no clear

PI3K knock-out and transgenic models

The function of class I PI3Ks described in the previous paragraphs were mainly inferred from studies of cells in culture. To explore the importance of tissue specific expression and activation of PI3K isoforms, studies at the level of the entire organisms are under way.

Concluding remarks

During the past few years we have witnessed a steady interest in PI3K signaling, and an impressing amount of biochemical, structural and physiological data has been acquired. Yet, a lot of open questions remain to be addressed.

The 3D structures of the sub-domains of the class I regulatory p85 subunit, except for the iSH2 domain, have been solved. However, we still do not understand completely how structural changes in p85, notably the putative coiled-coil structure of the iSH2 region, promote

Acknowledgements

We thank M. Zvelebil and M.D. Waterfield for the p85 model depicted in Fig. 2; L.C. Cantley, S. Dedhar, M.J. Fry, E. Hirsch, T. Kadowaki, S. Koyasu, S. Leevers, R.L. Nussbaum, Y. Terauchi, and R. Wetzker, for the release of information before publication; and S. Leevers for critical comments.

References (311)

  • N. Divecha et al.

    Cell

    (1995)
  • M.D. Bootman et al.

    Cell

    (1995)
  • D.R. Kaplan et al.

    Cell

    (1987)
  • L.C. Cantley et al.

    Cell

    (1991)
  • K.R. Auger et al.

    Cell

    (1989)
  • J.A. Escobedo et al.

    Cell

    (1991)
  • E.Y. Skolnik et al.

    Cell

    (1991)
  • M. Otsu et al.

    Cell

    (1991)
  • I.D. Hiles et al.

    Cell

    (1992)
  • J. Domin et al.

    FEBS Lett.

    (1997)
  • B. Vanhaesebroeck et al.

    Trends Biochem. Sci.

    (1997)
  • T. Hunter

    Cell

    (1995)
  • J. Kunz et al.

    Cell

    (1993)
  • D.M. Sabatini et al.

    Cell

    (1994)
  • M.F. Lavin et al.

    Trends Biochem. Sci.

    (1995)
  • K.O. Hartley et al.

    Cell

    (1995)
  • G.W. Booker et al.

    Cell

    (1993)
  • S. Koyama et al.

    Cell

    (1993)
  • J. Liang et al.

    J. Mol. Biol.

    (1996)
  • G.B. Cohen et al.

    Cell

    (1995)
  • H. Yu et al.

    Cell

    (1994)
  • R. Kapeller et al.

    J. Biol. Chem.

    (1994)
  • S.P. Soltoff et al.

    J. Biol. Chem.

    (1996)
  • J. Wang et al.

    J. Biol. Chem.

    (1995)
  • P. Mak et al.

    FEBS Lett.

    (1996)
  • G. Siegal et al.

    J. Mol. Biol.

    (1998)
  • R.B. Russell et al.

    FEBS Lett.

    (1992)
  • T.C. He et al.

    Blood

    (1993)
  • H. Shigematsu et al.

    J. Biol. Chem.

    (1997)
  • Y. Zheng et al.

    J. Biol. Chem.

    (1994)
  • K.F. Tolias et al.

    J. Biol. Chem.

    (1995)
  • S. Ahmed et al.

    J. Biol. Chem.

    (1994)
  • B.R. Dey et al.

    Gene

    (1998)
  • K. Inukai et al.

    J. Biol. Chem.

    (1996)
  • D.A. Fruman et al.

    Genomics

    (1996)
  • K. Inukai et al.

    J. Biol. Chem.

    (1997)
  • K. Baltensperger et al.

    J. Biol. Chem.

    (1994)
  • P. End et al.

    J. Biol. Chem.

    (1993)
  • H. Hayashi et al.

    J. Biol. Chem.

    (1993)
  • M. von Willebrand et al.

    J. Biol. Chem.

    (1998)
  • L.E. Rameh et al.

    Cell

    (1995)
  • I. Gout et al.

    Cell

    (1993)
  • T. Rordorf-Nikolic et al.

    J. Biol. Chem.

    (1995)
  • F.X. Yu et al.

    J. Biol. Chem.

    (1992)
  • M.J. Berridge et al.

    Nature

    (1984)
  • Y. Nishizuka

    Science

    (1992)
  • M. Whitman et al.

    Nature

    (1985)
  • Y. Sugimoto et al.

    Proc. Natl. Acad. Sci. U.S.A.

    (1984)
  • I.G. Macara et al.

    Proc. Natl. Acad. Sci. U.S.A.

    (1984)
  • A.E. Traynor-Kaplan et al.

    Nature

    (1988)
  • Cited by (588)

    • Biology and pathophysiology of central nervous system metastases

      2022, Neurological Complications of Systemic Cancer and Antineoplastic Therapy
    View all citing articles on Scopus
    View full text