90
Biochemical Society Transactions (2003) Volume 31, part 1
Integrin–protein kinase C relationships J. Ivaska*, S. Kermorgant*, R. Whelan*, M. Parsons†, T. Ng†‡ and P.J. Parker*1 *Cancer Research U.K., London Research Institute, 44 Lincoln’s Inn Fields, London WC2A 3PX, U.K., †Richard Dimbleby/Cancer Research U.K., Department of Cancer Research, GKT School of Medicine, St. Thomas’ Hospital, London SE1 7EH, U.K., and ‡Randall Centre, King’s College London, Guy’s Medical School Campus, London SE1 1UL, U.K.
Abstract The integrins have an ability to interact with extracellular matrix proteins to confer adhesive and motile properties on cells. The means by which these activities operate and the manner in which they are integrated with cell functions is of particular relevance to many biological processes. In the present paper, the developing understanding of the bi-directional relationship between the protein kinase C family of signal transducers and integrins is discussed.
Introduction Cell–matrix interactions are essential elements in maintaining a variety of cellular properties including polarity, migration, survival and growth-factor responsiveness. These interactions are supported by a family of heterodimeric (α/β), cell surface, matrix receptors – the integrins [1]. Combinations of distinct α- and β-integrins determine binding specificity for the different matrix proteins (fibronectin, laminin, vitronectin, collagen, gelatin), the basis of which has become more clearly defined through structural analysis [2]. The ability to bind matrix and the consequent assembly of cellular signalling complexes is determined not only by cell-surface expression, but also through ‘inside-out’ signalling that can confer high affinity/avidity on α/β-integrins [3,4]. Notable among agents that can effect this inside-out signalling are phorbol esters and, although there are now multiple classes of phorbol-ester-binding proteins established, it is thought that the protein kinase C (PKC) family has an important role in this integrin control [3]. Reciprocally, there is accumulating evidence that integrins have a key role in the function of PKC family members [5]. The inter-relationship of these two families of proteins is reviewed in the present paper. The particular aspects considered below relate to the permissive control of PKCs and their influence on the traffic of, and migratory responses to, integrins.
Integrin control of PKC Although members of the PKC family have long been considered to be under the direct allosteric control of lipid second messengers, a recent study has emphasized the importance of a series of conserved phosphorylations that are important for optimum catalytic function [5] Key words: adherence, migration, phosphorylation. Abbreviations used: ERM, ezrin-radixin-moesin; nPKC, novel PKC; PAR, proteinase-activated receptor precursor; PDK1, phosphoinositide-dependent kinase-1; PI3K, phosphoinositide 3-kinase; PI-PLC; phosphoinositide-specific phospholipase C; PKC, protein kinase C; PTEN, phosphatase and tensin homologue deleted on chromosome 10. 1 To whom correspondence should be addressed (e-mail
[email protected]).
C 2003
Biochemical Society
(Figure 1). One key reason underlying the generally slow appreciation of the importance of these phosphorylations is that the PKC proteins, as isolated from cells or tissues, are usually highly phosphorylated and hence have a respectable specific activity. This clearly contrasts with more conventional mitogen-activated protein kinases such as protein kinase B (PKB; also known as Akt) and extracellular-signal-related kinase (ERK)1/2, where acute stimulation is needed to trigger robust phosphorylation and activity. In part, this high degree of phosphorylation for the novel PKCs (nPKCs) δ and ε is a function of cell– cell and/or cell–matrix contacts, since, in some, but not all, cell types, phosphorylation is readily maintained while cells remain adherent to substrate. The importance of integrins in this pattern of behaviour is clearly shown by replating experiments that demonstrate matrix-dependent rephosphorylation of these nPKCs [6]. In fact, recent evidence indicates that release from, and replating on, matrix can have a profound effect upon the PKCε phosphorylation state, with the accumulation of dephosphorylated species [7]. The precise underlying mechanisms in this pattern of behaviour remain unresolved, although the acute nature of this response indicates that this is a protein-phosphatasedependent process, as opposed to the accumulation of de novo unphosphosphorylated PKCε. Interestingly this response to cell–matrix disengagement is not a universal one for all PKC isoforms. For example, PKCα does not appear to become acutely dephosphorylated at its three priming sites on disengagement of integrins (P. Whitehead and P.J. Parker, unpublished work). Interestingly, sustained maintenance of cells in suspension has shown that both PKCε and PKCδ accumulate as dephosphorylated forms [6]. One well-established element in the control of PKC phosphorylation is the phosphoinositide 3-kinase (PI3K)/phosphoinositide-dependent kinase-1 (PDK1) pathway [8]. PDK1 has been shown to form complexes with many PKC isoforms and will phosphorylate the conserved activation loop TFCG motif (where underlining indicates the site of phosphorylation) [9]. The involvement of PI3K is indicated, by the effects of inhibition and the established properties of PDK1, as
Signalling the Future
Figure 1
Priming phosphorylation sites on PKC family members Representative PKC family proteins are shown with sequences highlighting regions covering the activation loops (targets for PDK1) and the C-termini. The latter are putative autophosphorylation sites (for PKCδ there is very good evidence that this is the case [31]) and a further site (an acidic residue in some members) for which the kinase is not defined.
a PtdIns(3,4,5)P3 -binding protein [10]. There is, however, some question concerning the in vivo requirement of PtdIns(3,4,5)P3 in the PDK1 phosphorylation of some PKC isoforms [11]. Despite this, however, the prolonged suspension culture accumulation of dephosphorylated PKCε and PKCδ is influenced by the status of PTEN (phosphatase and tensin homologue deleted on chromosome 10) [6]. This tumour suppressor gene encodes a PtdIns(3,4,5)P3 3-phosphatase [12] and its loss of function is associated with elevated PtdIns(3,4,5)P3 and signal output through the PI3K pathway. Whether this particular adherence-dependent PTEN–PKC relationship underlies the association of PTEN loss with advanced, metastatic tumours is of significant interest and may relate to the roles that PKC isoforms have in cell migration (see below). The coupling of integrins to cellular signals is effected through the recruitment of various signalling and structural/scaffold proteins. Among these, the protein tyrosine kinases FAK (focal adhesion kinase), PYK2 (proline-rich tyrosine kinase 2) and c-Src seem to have key roles in signal propagation, whereas the cytoskeletal proteins such as α-actinin, talin and paxillin contribute changes in actin organization [13,14]. With respect to the control of PKC, the key transducers are phosphoinositide-specific phospholipase C (PI-PLC) and PI3K. These transducers have been reported to become associated with certain activated α/β integrins. In
the case of PI-PLCγ 1 , this may be through direct interaction with the β 1 -integrin tail [15], although in the context of the activated α 1 β 1 -integrin complex, PI-PLC was not identified [16]. In relation to these PKC pathways, evidence for integrindependent spreading involving interaction with Rac, Ras, PKCε and PI3K has also been presented [17]. The specificity and details of integrin complex formation with PI-PLC, PI3K etc. and hence their signalling capacity via diacylglycerol and PtdIns(3,4,5)P3 to PDK and PKC requires further analysis.
PKC control of integrins Adhesive properties of integrins that control cell-spreading properties, and the dynamic properties of integrin–matrix engagement that control cell movement, are under the influence of a number of different intracellular regulators. These inside-out signals include the actions of small G-proteins (Ras, Rac and cdc42), PI3K and PKC [18]. A significant element of the literature on the influence of PKC on integrins derives from the use of phorbol esters. These PKC activators have long been known to affect integrin affinity/avidity. Although it is now well established that PKCs are not the only phorbol ester receptors [19], studies demonstrate that PKCs do indeed regulate integrin properties. C 2003
Biochemical Society
91
92
Biochemical Society Transactions (2003) Volume 31, part 1
A general theme arising out of the recent studies on integrins and PKC is the demonstration of co-localization. This is evident in spreading responses, migration, traffic and, more particularly, in direct or indirect complex formations. Such observations of co-localization are important in assessing the actions of signal transducers and it is certainly the case for PKCs where a plethora of binding proteins are considered to be important in determining PKC function through localization [20]. The direct association of PKCα with β 1 integrin [21] and of classical or conventional PKCs (cPKCs) with CD81 (an integrin-associated tetraspanin protein) [22] provides clear evidence for a role ‘by association’. For CD81, the region of interaction has been partly refined to exclude the small internal loop, the transmembrane regions 3 and 4 and the large extracellular loop. In case of β 1 -integrin interaction with PKCα, the binding occurs through the association of the β 1 -integrin tail with the V3 variable region between the regulatory and catalytic domains [23]. This latter association was originally shown to correlate with β 1 -integrin dependent migration [21]. More recent studies exploiting the mapping of this binding site through the use of a competitive binding peptide and deletion mutants have provided direct evidence for this role of PKC [23]. Catalytically, there is no evidence for a role of PKCα in the direct phosphorylation of β 1 integrin in vivo. In contrast, several serine/threonine sites in the β 2 -integrin cytoplasmic domain have been shown to be phosphorylated by PKC in vitro [24]; one of these PKC sites, Thr-758, could lead to the recruitment of 14-3-3 proteins to the activated integrin following phosphorylation in situ. However, the assembly of the β 1 -integrin–PKC complex is correlated with recruitment of the ERM (ezrin-radixinmoesin) family protein ezrin and its phosphorylation on the Thr-567 site [25]. This phosphorylation is associated with stabilizing the open conformation of ezrin with the capacity for interactions through its tail domain (with actin) and its N-terminal FERM domain (with various membrane localized partners) [26]. The phosphorylation of ezrin is essential for supporting integrin dependent migration and in vitro studies, alongside pharmacological ones, indicating that this is PKCmediated phosphorylation [25]. The physical and functional relationships between PKC, β1 integrin and the ERM proteins have important implications in biological settings other than integrin-mediated migration. For instance, a constitutive PKCα–β1-integrin complex has also been found in multiple myeloma cells and has a significant role in the development of a vascular endothelial growth factor (VEGF)-responsive migratory phenotype [27]. In addition, CD44-mediated motility upon phorbol ester stimulation is dependent on β1- and not αV-integrin ([28]; M. Parsons and T. Ng, unpublished results). It is possible that the dissociation of ezrin from other surface receptors such as CD44, L-selectin or intracellular adhesion molecule-1 (ICAM-1) may be a prerequisite for ezrin association with PKCα and its subsequent phosphorylation, which is essential for supporting integrin-dependent migration. The role of PKC described above in mediating this characteristic integrin-dependent event is perhaps more reflective C 2003
Biochemical Society
of integrin control of PKC than vice versa. However, there is an additional element, more poorly understood, that suggests there is, in fact, a reciprocal relationship. The association of PKCα with β 1 -integrin in cells, as determined by FRET (fluorescence resonance energy transfer), is characterized by a greater interaction between the ‘activated’ conformer (12G10-positive) than with the ligand-unoccupied conformer (Mab13-positive) [21]. Since the interaction is promoted by PKC activation, it is reasonable to consider that the binding of PKCα to β 1 -integrin shifts the conformer equilibrium towards the ‘activated’ conformer. This is also associated with increased cell-surface expression of the integrin. The overall working hypothesis would be that the open conformer–PKCα interaction promotes the accumulation of the ‘activated’ conformer of β 1 -integrin and its delivery to the cell surface, thereby regulating matrix interactions. This correlates with the many observations of phorbol-ester-induced up-regulation of integrins and integrin–matrix association. A further aspect of this relationship relates to the cellular traffic of integrins. For β 1 -integrin, it appears that the catalytic activity of PKCα contributes to the rate of internalization from the cell surface, a process critical to the dynamic migratory process. Intracellular traffic associated specifically with directional movement has been shown to be associated with PKCε function. In this instance, inhibition of PKC activity leads to the reversible accumulation of β 1 integrin that is internalized from the cell surface within an intracellular vesicular compartment alongside PKCε [29]. Loss of PKCε from these isolated β 1 -integrin-positive vesicles requires PKC activity alongside as yet unidentified cytosolic component(s). How this process confers directionality to β 1 -integrin traffic (i.e. trailing edge to leading edge movement) awaits further analysis, as indeed does understanding where and when PKCs functionally interact with other integrins.
Perspectives The bi-directional relationship between integrins and PKC, although on a firm foundation, requires a deeper level of understanding. The individual elements that contribute to functional complexes that exert the mutual controls outlined above, as well as the proximal outputs, are only partially documented. Furthermore, within a broader perspective, the polarity of events associated with localized matrix disengagement, localized signals and leading edge redistribution of integrins, will demand sophisticated cell biological analysis to achieve a molecular working model on integrin–PKC function. In this context, it is of interest to consider the relationship of the ‘cell matrix–integrin– PKC’ paradigm with the emerging one involving cell–cell contact: ‘cdc42–PAR6–PAR3–PKC’ [30]. In the latter case, the evidence indicates a role for PKC in tight-junction formation, but not in its maintenance. Is the integrin–PKC relationship only concerned with delivery and assembly of functional integrins or does PKC contribute to the reporting of integrin engagement status?
Signalling the Future
References 1 Howe, A., Aplin, A.E., Alahari, S.K. and Juliano, R.L. (1998) Curr. Opin. Cell Biol. 10, 220–231 2 Humphries, M.J. (2002) Arthritis Res. 4, S69–S78 3 Kolanus, W. and Seed, B. (1997) Curr. Opin. Cell Biol. 9, 725–731 4 Hughes, P.E. and Pfaff, M. (1998) Trends Cell Biol. 8, 359–364 5 Parekh, D.B., Ziegler, W. and Parker, P.J. (2000) EMBO J. 19, 496–503 6 Parekh, D.B., Katso, R.M., Leslie, N.R., Downes, C.P., Procyk, K.J., Waterfield, M.D. and Parker, P.J. (2000) Biochem. J. 2, 425–433 7 England, K., Watson, J., Beale, G., Warner, M., Cross, J. and Rumsby, M. (2001) J. Biol. Chem. 276, 10437–10442 8 Vanhaesebroeck, B. and Alessi, D.R. (2000) Biochem. J. 346, 561–576 9 Le Good, J.A., Ziegler, W.H., Parekh, D.B., Alessi, D.R., Cohen, P. and Parker, P.J. (1998) Science 281, 2042–2045 10 Stephens, L., Anderson, K., Stokoe, D., Erdjument-Bromage, H., Painter, G.F., Holmes, A.B., Gaffney, P.R., Reese, C.B., McCormick, F., Tempst, P. et al. (1998) Science 279, 710–714 11 Sonnenburg, E.D., Gao, T. and Newton, A.C. (2001) J. Biol. Chem. 276, 45289–45297 12 Maehama, T. and Dixon, J.E. (1998) J. Biol. Chem. 273, 13375–13378 13 Liu, S., Calderwood, D.A. and Ginsberg, M.H. (2000) J. Cell Sci. 113, 3563–3571 14 Frame, M.C., Fincham, V.J., Carragher, N.O. and Wyke, J.A. (2002) Nat. Rev. Mol. Cell Biol. 3, 233–245 15 Vossmeyer, D., Hofmann, W., Loster, K., Reutter, W. and Danker, K. (2002) J. Biol. Chem. 277, 4636–4643 16 Zhang, X.A., Bontrager, A.L. and Hemler, M.E. (2001) J. Biol. Chem. 276, 25005–25013 17 Berrier, A.L., Mastrangelo, A.M., Downward, J., Ginsberg, M. and LaFlamme, S.E. (2000) J. Cell. Biol. 151, 1549–1560
18 Schoenwaelder, S.M. and Burridge, K. (1999) Curr. Opin. Cell Biol. 11, 274–286 19 Kazanietz, M.G. (2000) Mol. Carcinog. 28, 5–11 20 Jaken, S. and Parker, P.J. (2000) Bioessays 22, 245–254 21 Ng, T., Shima, D., Squire, A., Bastiaens, P.I., Gschmeissner, S., Humphries, M.J. and Parker, P.J. (1999) EMBO J. 18, 3909–3923 22 Zhang, Q., Sakai, T., Nowlen, J., Hayashi, I., Fassler, R. and Mosher, D.F. (1999) J. Biol. Chem. 274, 368–375 23 Parsons, M., Keppler, M.D., Kline, A., Messent, A., Humphries, M.J., Gilchrist, R., Hart, I.R., Quittau-Prevostel, C., Hughes, W.E., Parker, P.J. et al. (2002) Mol. Cell Biol. 22, 5897–5911 24 Fagerholm, S., Morrice, N., Gahmberg, C.G. and Cohen, P. (2002) J. Biol. Chem. 277, 1728–1738 25 Ng, T., Parsons, M., Hughes, W.E., Monypenny, J., Zicha, D., Gautreau, A., Arpin, M., Gschmeissner, S., Verveer, P.J., Bastiaens, P.I. et al. (2001) EMBO J. 20, 2723–2741 26 Bretscher, A. (1999) Curr. Opin. Cell Biol. 11, 109–116 27 Podar, K., Tai, Y.T., Lin, B.K., Narsimhan, R.P., Sattler, M., Kijima, T., Salgia, R., Gupta, D., Chauhan, D. and Anderson, K.C. (2002) J. Biol. Chem. 277, 7875–7881 28 Legg, J.W., Lewis, C.A., Parsons, M., Ng, T. and Isacke, C.M. (2002) Nat. Cell Biol. 4, 399–407 29 Ivaska, J., Whelan, R.D., Watson, R. and Parker, P.J. (2002) EMBO J. 21, 3608–3619 30 Ohno, S. (2001) Curr. Opin. Cell Biol. 13, 641–648 31 Ziegler, W.H., Parekh, D.B., Le Good, J.A., Whelan, R.D.H., Kelly, J.J., Frech, M.M., Hemmings, B.A. and Parker, P.J. (1999) Curr. Biol. 9, 522–529
Received 6 September 2002
C 2003
Biochemical Society
93
