ACTAUNIVERSITATISUPSALIENSISUPPSALA2006

Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 136

Structural and Functional Aspectsof β1 Integrin Signalling

STINA NILSSON

ISSN 1651-6206ISBN 91-554-6529-3urn:nbn:se:uu:diva-6747

To my family!

List of Papers

This thesis is based on the following papers, referred to in the text by their roman numerals:

I Nilsson S, Kaniowska D, Brakebusch C, Fässler R, Johansson S. (2006) Threonine 788 in integrin subunit 1 regulates integrin activation. Exp Cell Res. April;312(6):844-853

II Velling T, Nilsson S, Stefansson A, Johansson S. (2004) 1-Integrins induce phosphorylation of Akt on serine 473 independ-ently of focal adhesion kinase and Src family kinases. EMBO Rep. Sep;5(9):901-5

III Nilsson S, Wärmegård B, Velling T, Johansson S. (2006) Activa-tion of Erk downstream of 1 integrins independently of FAK and ShcA. Manuscript

Reprints were made with permission from the publishers.

Front cover: Light microscopy image of 1A-expressing GD25 cells in culture.

Contents

Background...................................................................................................11Integrins....................................................................................................11

Integrin family of receptors .................................................................11Integrin ligands ....................................................................................14Integrin structure..................................................................................16

The extracellular domain ................................................................16The transmembrane domain............................................................17The cytoplasmic domain .................................................................18

Integrin activation – inside-out and back in again...................................18Inside-out signalling ............................................................................18

Integrin extracellular conformation.................................................18Integrin transmembrane and intracellular conformation.................20Talin dependent integrin activation.................................................22Regulation of integrin activation.....................................................23

Outside-in signalling............................................................................27Clustering of integrin receptors.......................................................27Modulation of integrin cytoplasmic interactions ............................27

Focal adhesion sites – protein ensembles with inositol guest stars .........29Molecular components of focal adhesion sites ....................................29

Cytoskeletal proteins.......................................................................30Signalling proteins and their regulators ..........................................32Signalling lipids and their regulators ..............................................38

Adhesion dependent and integrin regulated cell functions.......................40Spreading and organisation of the actin cytoskeleton .........................41Migration & integrin cell surface recycling.........................................41Survival................................................................................................42

Crosstalk with other cell surface receptors...............................................43

Present investigation .....................................................................................44Aims of thesis...........................................................................................44Results and Discussion.............................................................................44

Integrin activation – depending on common features of subunits? .44Paper I – Extracellular activity of 1 integrins is regulated by the intracellular residue threonine 788..................................................45

1 mediated signalling – relying on activation of the FAK/Src complex?..............................................................................................46

Paper II – FAK and PI3K activation as two parallel key events.....47Paper III – Adhesion via 1 integrins controls activation of Erk at several checkpoints, with and without FAK involvement ..............48

Future Perspectives .......................................................................................50Concluding remarks .................................................................................51

Acknowledgement ........................................................................................52

References.....................................................................................................54

Abbreviations

Cas Crk-associated substrate Csk C-terminal Src kinase ECM Extracellular matrix ER Endoplasmic reticulum Erk1/2 Extracellular signal-regulated kinase 1 and 2 FAK Focal Adhesion Kinase FERM Band 4.1, Ezrin, Radixin, Moesin GAP GTPase-activating protein GEF Guanine nucleotide exchange factor GFR Growth factor receptor Grb2/7 Growth-factor-receptor-bound protein 2 and 7 ICAP-1 Integrin cytoplasmic domain-associated protein 1 ILK Integrin-linked kinase JNK c-Jun NH2-terminal kinase MAPK Mitogen-activated protein kinase PAK p21-activated kinase PDK1 3-phosphoinositide-dependent protein kinase 1 PH Plextrin Homology PI3K Phosphatidylinositol-3-kinase PIP2 PI(4,5)P2 or phosphatidylinositol-4,5-bisphosphate PIP3 PI(3,4,5)P2 or phosphatidylinositol-3,4,5-trisphosphate PIPK-I Phosphatidylinositol phosphate kinase type I PKB/Akt Protein kinase B / Akt PKC Protein kinase C PP2A/B Protein phosphatase 2A and 2B PTB Protein tyrosine binding PTEN Phosphatase and Tensin homologue deleted on chromosome 10 PTP-PEST Protein tyrosine phosphatase-proline, glutamate, serine, and threonine se-

quencesPYK2 Proline-rich tyrosine kinase 2 RGD Arginine-Glycine-Aspartic acid RPTP Receptor protein tyrosine phosphatase S/T Serine/threonine SAPK Stress-activated protein tyrosine kinase SH2 Src homology 2 SH3 Src homology 3 Shc Src-homology containing protein SHIP1/2 SH2-containing inositol 5-phosphatase 1 and 2 SHP1/2 SH2-domain-containing protein tyrosine phosphatase 1 and 2 Sos Son of sevenless TM Transmembrane

11

Background

Cell adhesion molecules make cells stick together into multi-cellular tissues and organisms. They also provide means for the cell to communicate with its surroundings. Cell adhesion plays a critical role both in normal processes such as embryonic development and healing of wounds as well as in patho-logical processes like tumour cell invasion and metastasis, thrombosis, and inflammation [1, 2].

Most of the cell surface receptors that mediate adhesion can be grouped into four protein families: cadherins, immunoglobulin superfamily members, selectins, and integrins (reviewed in [3]). Integrins primarily mediate interac-tions of cells with components of the extracellular matrix (ECM) such as fibronectin, vitronectin, laminins, and collagens. Attachment of cells to the matrix can affect the cells in a number of ways; the response of the cell de-pends on the type of cell involved, their state at the moment and the specific make-up of the matrix. Together with signals arising from receptors for soluble growth factors integrins enforce a cellular decision on what biologi-cal action to take, be it attachment, movement, death or differentiation. The wide range of actions in the cell, under the influence of integrins, obviously requires a large set of players and a fine-tuned control.

Integrins Integrin family of receptors During the 1980’s this family of transmembrane (TM) glycoproteins were identified and named “integrins” in recognition of their importance to struc-tural integrity of cells and tissues as well as to denote their role as an integral membrane complex associating the exterior protein network with the cells interior cytoskeletal network across the plasma membrane [2, 4].

Integrin receptors act as heterodimeric non-covalent complexes formed by one - and one -subunit. Today, 18 - and 8 -subunits have been de-scribed in mammals and together they form 24 known receptors (Figure 1).Roughly, integrins fall into three subfamilies dependent on their subunit composition and ligand specificity. (i) The 1 subunit can combine with 12 different subunits and thereby form the largest subfamily. The 1 contain-ing integrins are widely expressed and mediate cell adhesion mainly to ECM proteins. (ii) The 2 and 7 integrins are expressed exclusively on blood

12

cells and mainly mediate cell-cell interactions with e.g. ICAM-s and E-cadherin. (iii) The last subfamily all contain the V subunit and recognize an arginine-glycine-aspartic acid (RGD)-motif found in a broad range of ligands. Blood cells, endothelial cells, epithelial cells and osteoclasts express

V integrins. The highly specialised IIb 3 in platelets and 6 4 in kerati-nocytes, with their important role in blood coagulation and hemidesmosome adhesion, respectively, fall outside the three mentioned subfamilies.

Figure 1. Outlined are the mammalian integrin subunits and their associa-tions. Here they are grouped according to their ligand specificity and, in the case of

2 and 7, their restricted expression on leukocytes. Asterisks denote the presence of alternatively spliced cytoplasmic domains. (Adapted from [5].)

Alternative splicing of the mRNA encoding some of the integrin subunits further expands the diversity of integrins (for review see [6, 7]). Variant sequences have been described for the extracellular domains of subunit 6( 6X1 and 6X1+X2; [8]), 7 ( 7X1 and 7X2; [9]), and finally two variants of

IIb [9]. The variability may influence the ligand binding capacity as well as supplying a mechanism for regulating the affinity of the receptor as in the case of 7 1, or determine the preferences for heterodimerisation as for

6X1+X2 do 4 over 1.Variant cytoplasmic domains may, by providing alternative docking sites

for cytoskeletal and signalling molecules, activate different signalling path-ways. For seven of the integrin subunits ( 3, 6, 7, 1, 3, 4, and 5)

13

splice variants in the cytoplasmic domain are identified (Figure 2). The three alternatively spliced subunits have a close evolutionary relationship and two variants, A and B, have been described for each of them. In addition, for

7 a third variant, 7C, is found [10].

Figure 2. Linear alignment of the amino acid sequences of the cytoplasmic do-mains of the alternatively spliced integrin subunits (except 4). The conserved membrane-proximal residues are marked in red. Note that the cyto-1 motif (under-lined in 1) is conserved in all splice variants of 1 and 3, but cyto-2 and -3 motifs (in blue) only are conserved in some variants (see text for details). (* The sequence of 7 is of rat origin, and 5 is of mouse origin; adapted from [6].)

There are five alternatively spliced variants of the integrin 1 subunit ( 1A,B, C-1, C-2, and D). They are identical in amino acid composition until the AKWDT777 segment of the cytoplasmic domain, whereupon each isoform continues with a unique sequence of differing length (Figure 2). Even though five splice variants for 1 have been described, most likely only two of them ( 1A and 1D) are proteins with physiological functions. The splice variant 1A, mostly referred to as 1, is abundantly expressed in all tissues. Expression of the muscle specific 1D variant is strictly regulated during development and displays stronger adhesion to ligand and association with cytoskeleton than 1A [11]. 1B and 1C variants are minor and with re-stricted distribution, generally co-expressed with 1A. The origin, biosyn-thesis and function of 1-splice variants are thoroughly reviewed in [12]. The similarity between 1 and 3 subunits is particularly strong. Like the 1splice variants, the three 3 isoforms ( 3A, B, and C) share a common membrane proximal amino acid sequence of 20 residues followed by entirely

14

different sequences [13, 14]. Neither of these unique sequences resembles any of those in the 1-variants.

For the keratinocyte specific 4-integrin several different inserted se-quences are accountable for the five variations of the cytoplasmic domain ( 4A-E) [15, 16]. The tissue distribution of the 4 cytoplasmic variants is poorly characterised, and hardly any functional studies have been performed. Like most integrin subunits, the variants of 4 are expressed together with the most abundant form, 4A. Two alternative splicing variants of 5 are found in mouse ( 5A and B), but the splice site is not present in human, and no other splice variants have been detected [17]. The 5A and 5B are co-expressed in most mouse tissue; although 5B mRNA is present at much lower levels than 5A mRNA no functional differences have been described [17].

Integrin ligands The number of integrins and the remarkable breadth of their cellular distribu-tion support the statement that the phenotype of virtually every cell is uniquely influenced by its display of integrins. Each cell usually express several types of integrin receptors on its surface, hence making it possible to bind multiple kinds of ligands and interpret parallel signals from the ex-tracellular space. In addition, most integrin receptors display affinity for a diverse set of ligands.

Table 1 summarises the integrin specificities for their major extracellular ligands and shows how the specificity preferences overlap between the dif-ferent receptors when expressed on the same cell. The list includes a large number of ECM proteins (bone matrix proteins, collagens, fibronectin, fi-brinogen, laminins, thrombospondins, vitronectin, and von Willebrand fac-tor) reflecting the primary function of integrins in cell adhesion to extracellu-lar matrices. Many “counter-receptors” are ligands (VCAM, ICAM, and E-cadherin), reflecting the role of integrins in mediating cell-cell interactions. Included are also some molecules involved in the immune system or dis-played by microorganisms to gain entry into cells (e.g. iC3b and invasin).

15

Table 1. Integrins and their major extracellular ligands Integrin Matrix molecule Other ligands

1 1 Col I, IV, VI; Ln 2 1 Col I, II, III, IV, VII, XI; Ln 3 1 Ln 2/4, 5, 10/11, TP Inv 4 1 Fn, CS-GAG VCAM-1, Inv, Im 5 1 Fn, Fg, dCol disintegrins, Im, Inv 6 1 Ln Inv, Sperm Fertilin 7 1 Ln 1, 2/4 8 1 Fn, Vn, Tn 9 1 Col I, Ln, Tn, OP VCAM-1 10 1 Col II

Subfamily (i)

11 1 Col I D 2 ICAM-3 L 2 ICAM-1, 2, 3, 4, 5 M 2 Fg ICAM-1, iC3b, FX X 2 Fg iC3b 4 7 Fn MAdCAM-1, VCAM-1,

disintegrins

Subfamily (ii)

E 7 E-cadherin

V 1 Fn, Vn TGF LAP V 3 Vn, Fg, Fn, bSp, Tn, TP, OP,

MAGP-2, fibrillins, Del1, dCol vWF, disintegrins, L1-CAM

V 5 Vn, bSp, Fn V 6 Fn, Tn TGF LAP

Subfamily (iii)

V 8 Col I, Fn, Ln IIb 3 Fg, Fn, Vn, TP, dCol, Dec vWF, Pl, disintegrins, L1-CAM 6 4 Ln

Abbreviations: bSp – bone sialoprotein; Dec – decorsin; Del1 – developmental endothelial locus-1; (d)Col – (denatured) collagen; CS-GAG – chondroitin sulphate glycosaminoglycan; Fg – fibrinogen; Fn – fibronectin; FX – Factor X; iC3b – inactivated fragment of complement factor C3; ICAM – intracellular adhesion molecule; Im – intimin; Inv – invasin; Ln – laminin; L1-CAM – neural cell adhesion molecule L1; MAGP – microfibril – associated glycoprotein; MAdCAM – mucosal addressin cell adhesion molecule; OP – osteopontin; Pl – plasminogen; TGF LAP – transforming growth factor latency-associated peptide; Tn – tenascin-C, TP – trombospondin; VCAM – vascular cell adhesion molecule; vWF – von Willebrand factor (Adapted from Armulik A. Thesis Uppsala University, 2000.)

16

Integrin structure Integrins are type-I transmembrane glycoproteins. The receptor is formed through dimerisation of an - and a -subunit in the endoplasmic reticulum (ER), and integrin dimers are transported from the ER via the Golgi network to the cell surface [18]. Normally, a single integrin subunit cannot leave the ER, but what determines that a given subunit should associate with a cer-tain subunit is not known. Each receptor subunit consists of a large ex-tracellular domain that participates in ligand binding, with an N-terminal globular headpiece and a stalk region leading into the single TM spanning domain. Followed by a short cytoplasmic domain interacting with cytoskele-tal and signalling proteins. For a schematic illustration of the overall struc-ture see Figure 3.

Figure 3. Schematic view of the integrin heterodimer. Indicated are the ligand binding regions: the -propeller and A-domain in the -subunit, and the A-domain in the -subunit. Cation binding sites are depicted with spheres. A whole receptor is 28 nm long as determined for 5 1 [19]. Adapted from [20, 21].

The extracellular domain The extracellular domains are 1000 and 750 amino acids long for the and subunits, respectively. In addition, during biosynthesis each subunit is modified by N-linked glycosylation in the endoplasmic reticulum, these N-linked sugars are further processed within the Golgi compartment and vari-ant glycosylation is suggested to regulate integrin function [22-24]. For ex-ample, both the 3 1 and 5 1 receptors contain as many as 26 potential

17

glycosylation sites in the extracellular part, and differential glycosylation have been associated with phenotypic variations [23].

The integrin subunits of 3, 5, 6, 7, 8, and V are posttranslation-ally cleaved, close to the TM border, into a light and a heavy chain, which are covalently linked by a disulfide bridge [1].

The extracellular domain of the subunits contains seven homologues repeats (I-VII), which have been proposed to fold cooperatively into a single domain known as the propeller [25]. In addition, nine subunits ( 1, 2,

10, 11 L, M, X, D, and E) contain an extra, independently fold-ing, domain of about 180 amino acids, termed the I domain (inserted) or Adomain (homologues to the collagen-binding A domains of von Willebrand factor) domain. The A-domains are inserted between the homologues re-peat domains II and III in the -propeller [26].

The propeller, with or without an A domain inserted, together with a conserved A-like domain, A, present in all the subunits appear to mediate the ability of integrins to bind ligand. Thus, both subunits contribute to the formation of the ligand-binding pocket, but the subunit is likely to play an important role in determining the binding selectivity since often het-erodimers sharing a common subunit but with distinct chains, bind dif-ferent ligands. The ligand binding requires the presence of divalent cations, in general, physiological concentrations of Mn2+ and Mg2+ positively con-tribute to the binding affinity whilst Ca2+ inhibits ligand binding [27].

The transmembrane domain The single TM domain is 25 amino acids long [28] and form an -helix together with the membrane proximal region of the cytoplasmic domain [29]. Alignment of the TM domains of integrin and subunits reveal re-peatedly insertion of small amino acids, like glycine and alanine, at three positions and could speculatively be involved in the interaction of the het-erodimer or other TM proteins.

The borders of the TM segment has for some time been unclear. This is due to the fact that all subunits contain 23 hydrophobic residues, followed by one single positively charged amino acid (lysine or arginine) and then an-other four (in the subunits) or 5 (in the subunits) hydrophobic residues before the stretch of non-TM amino acids starts. The C-terminal border has been suggested to start both at the point of the single charged amino acid and after the last 4-5 hydrophobic residues. By using a glycosylation mapping technique, both N- and C-terminal borders have been determined for a hand-ful of subunits ( 2, 5, 10, 1, and 8) and it was shown that the charged amino acid is buried within the membrane in contrast to what previously, most often, was assumed from the primary sequences [28]. However, it is possible that the TM segment with this charged amino acid is shifted out of the membrane when the heterodimer tails associate (see section about In-tegrin transmembrane and intracellular conformation).

18

The cytoplasmic domain Integrin receptors do not possess any known enzymatic activity in the cyto-plasmic domain, as compared to e.g. growth factor receptor; instead they transfer signals across the plasma membrane by conformational changes and by interacting with cytoplasmic adaptors and kinases.

The cytoplasmic domains of integrins are 20 and 50 amino acids long for and subunits, respectively ( 4 is unique with a >1000-amino acid cytoplasmic tail that binds intermediate filaments). There is little homology between the cytoplasmic domains of the subunits, except for the conserved seven amino acid sequence KXGFFKR (in which X represents a non-conserved residue) present at the TM-cytoplasm border (Figure 2). How-ever, each subunit is highly conserved among different species, which indicates that the cytoplasmic domains are important for integrin functions [1, 30].

On the contrary, there is considerable homology between the cytoplasmic domains of the 1, 2, 3, 5, 6, and 7 subunits, which suggests that they share functional properties [6, 31, 32] (Figure 2). In addition to conserved amino acids at the TM border, four patches within the cytoplasmic domain are recognised for their central influence on integrin function, they are termed cyto-1, cyto-2, cyto-3, [33] and S/T [34-36] and indicated in Figure 2and Figure 5. The cyto-1 region is highly charged and important for localisa-tion to focal adhesions [33], and the cyto-2 and -3 are two well conserved NPXY and NXXY motifs, respectively, with tyrosine residues modified by phosphorylation and influencing outside-in signalling ([37] among many). Between the two NXXY motifs a stretch of serine and threonine residues reside, that strongly influence the ligand binding ability of the 1, 2, and 3containing integrins [38]. These functionally important patches are discussed in further detail below.

Integrin activation – inside-out and back in again In stationary cells most integrin receptors are in an active form with a con-formation suitable for ligand binding (i.e. with high affinity for ligand) and subsequent cellular events, while in circulating blood cells the normal status is an inactive conformation (low affinity) of the integrins.

Inside-out signalling Integrin extracellular conformation A prerequisite for initial ligand binding and all subsequent events is the acti-vation of integrins by cytoplasmic signals that confer a large conformational

19

change to the extracellular domain, so called “inside-out signalling”. Inside-out signalling is accomplished when signals received from other cell surface receptors, such as G protein-coupled receptors, are transmitted from the in-tegrin cytoplasmic face to the extracellular domain, thereby increasing the affinity of the receptors for ligands (integrin activation) [5]. In the last few years, many steps have been taken towards understanding of the changes in conformational (activation) states of the integrin receptors. In 2001, high resolution data of the crystal structure of integrin V 3 extracellular seg-ment was presented [39], and following studies by electron microscopy and NMR propose a bent extracellular conformation of the inactive integrin with a switchblade-like opening motion extending the integrin to an up-right ac-tive conformation [40, 41] (Figure 4).

Figure 4. An illustration of the clasp-release model of integrin activation. Seetext for details.

For a very long time integrin receptors were believed to not possess any enzymatic activity. But in year 2000, O’Neill and co-workers reported that the integrin 3 subunit possesses an endogenous thiol isomerase activity in the cysteine-rich repeats of the extracellular domain [42]. Redox agents, like nitric oxide and glutathione, modulate both the thiol isomerase activity of purified IIb 3 and its active conformation in intact platelets, suggesting a molecular mechanism for integrin conformational regulation [43]. Thus,

20

rearrangement of intramolecular disulfide bridges could possibly regulate conformational restraints in the extracellular domain.

Integrin transmembrane and intracellular conformation The integrin TM-spanning and cytoplasmic domains have also been in focus of integrin activation investigations. Li and co-workers [44-46] put forward interesting results, from mutagenesis and the use of TOXCAT, on homodimeric and homotrimeric associations of the TM domains of IIb and

3 with a possible roll in activation of the IIb 3 receptor. They suggest that TM domain-mediated homo-oligomerisation provides a plausible structural basis for integrin clustering, which is important for outside-in signalling, but could promote integrin activation as well.

Characterisation of the 3 TM-cytoplasmic domains by NMR, for a monomeric peptide in phospholipid micelles [47], reveals that the helix structure of the TM amino acids extends into the conserved membrane-proximal region of the cytoplasmic tail (Figure 5). The length of the deter-mined helix extends well past a typical membrane interface motif (trypto-phan-lysine) and is interrupted by a hinge region after I721. In contrast, NMR studies of a complex of 3 and IIb cytoplasmic regions (including residues 3(K716-T762) and IIb(K989-E1008) [48]) reveal a non-interrupted helical structure until 3-K738, and a primary interface between the IIb and 3 tails of amino acids K716-D723 in 3 and K989-N996 in

IIb.Determination of the TM borders by glycosylation mapping (single and

subunits in microsomal membranes) suggest that the TM segment is longer than necessary for a straight helix to cross the membrane perpendicularly [28, 49], in addition to positioning the conserved lysine within the mem-brane. Several models of how this long TM helix could reside within the membrane have been suggested, including one where the helix has a tilted position to accommodate its full length. Such a tilted position would shield a part of the conserved region from potential binding partners in the cyto-plasm. Given the hydrophobic nature of the conserved region, transition from an upright positioning in the membrane with the conserved region pointing in to the cytoplasm, to a tilted or bent position with part of the the conserved region located in the lipid bilayer, would not be energetically costly [47] (Figure 4). Taken together, the results from the single subunit studies mentioned may describe the TM conformations of separated integrin subunits, while the other study may possibly represent the conformation of an inactive integrin with subunit interactions favouring a longer part of the integrin tails in contact with the cytoplasm.

21

Figure 5. Linear alignment of the amino acid sequences of the transmembrane and cytoplasmic domains of selected integrin subunits. The TM borders are de-picted by a vertical line [28, 49], Highlighted are motifs that are important for in-tegrin functions. The conserved membrane-proximal residues of and chains are denoted in red, the NPXY/F motifs in blue and the S/T-rich motifs in green.

The importance of the membrane proximal regions of integrin and sub-unit cytoplasmic domains for inside-out signalling has been under intensive investigation for more than a decade (reviewed in [50, 51]). Deletion of the conserved KXGFFKR motif in the subunit or the KLLXXXHDR in the subunit leads to constitutive activation of the receptor [52]. The compiled studies have led to an activation model where the and cytoplasmic tails associate to form a clasp that maintains the receptor at a low-affinity state, whereas cellular activators such as talin induce the dissociation of the clasp during inside-out activation [48, 52-55] (illustrated in Figure 4). This un-clasping hypothesis is supported by NMR data for IIb 3 [48] as well as

5 1 and V 3 [56], biochemical data [57, 58], and a fluorescence energy transfer experiment in vivo [59].

The side-chains of two well conserved amino acids, arginine and aspartic acid, in the - and -subunit respectively, at the +2 position downstream of the TM C-terminal borders (as determined by [28]) have been indicated to form an electrostatic bridge between the two subunits. Such an interaction could restrain the receptor in an inactive conformation by clasping and chains together [55]. Interestingly, Vinogradova and co-workers showed that the membrane-proximal helices of integrin IIb 3 become partially embed-ded into the membrane in the unclasped state, using membrane-mimetic micelles, thereby inducing substantial conformational change on each tail [56]. This supports the activation model where the membrane-proximal heli-ces may move upward into the membrane upon separation.

22

Talin dependent integrin activation Talin-dependent activation of 5 1 and V 3 was recently demonstrated by the use of small interference RNA technology [60] in concert with previous work. The interaction between talin and the integrin cytoplasmic domain is thought to be a final common step in activation of several integrins (re-viewed in [51]). Briefly, talin is a major cytoskeletal actin binding protein, found as an anti-parallel homodimer of two large subunits. Each subunit consists of a globular N-terminal head region and a C-terminal elongated rod. The talin head binds to the cytoplasmic domains of 3, and other sub-units, via a protein tyrosine binding (PTB) subdomain within the conserved FERM (Band 4.1, Ezrin, Radixin and Moesin) domain of talin [61, 62]. The major integrin binding site lies within the talin head, but when the talin homodimer is formed this site comes in close proximity of a low affinity-binding site found in the rod domain [63]. The close vicinity of the two in-tegrin-binding sites could possibly work together to modulate the talin affin-ity for integrin cytoplasmic tails in ways that, so far, have not been investi-gated. Alternatively, they might function independently and carry out sepa-rate functions.

Several distinct regions in the cytoplasmic tail have been mapped for ta-lin head binding [48, 64, 65], but the binding site(s) within the cytoplasmic domain mediating interaction with the talin rod domain has yet not been defined. Crystal structure, NMR data, and mutagenesis studies reveal that the talin head-binding regions include the N-terminal N744PxY/F-motif and the conserved tryptophan residue of the chains [66]. Residue N744 in 3 play a key structural role [66] and Y747 is necessary for the talin-integrin interac-tion [61, 62]. These results are supported by alanine mutations of each site, which disrupts talin binding [61, 64], while a more conservative substitution, Y783F in 1, is functional [64]. Phenylalanine is also found in this position in 2, which is known to bind strongly to talin, consistent with the hydro-phobic nature of the binding pocket in the talin head. Also the very mem-brane-proximal parts of the tail are shown to contribute to the interaction with the talin head [48, 56, 65, 67].

The talin head was shown to activate integrins containing the subunits 1, 2, 3, 5, and 7 [61, 62]. These results are consistent with studies show-

ing that the interaction between the IIb and 3 tails was disrupted when talin head was present and bound the 3 tail at a region completely overlap-ping the membrane-proximal region interacting with the IIb tail [48]. The favoured binding of talin is explained by a much higher affinity of 3 for talin than for the IIb tail. Thus, initially, talin interaction with the 3 in-tegrin cytoplasmic tail might occur at the NPXY motif, with subsequent engagement of the membrane-proximal region leading to the displacement of the cytoplasmic tai, tail separation, and integrin activation (Figure 4).

23

In addition to talin, other proteins have been suggested to activate in-tegrins or to inhibit activation by competitive binding, e.g. 3-endonexin [68], filamin [69], and other PTB-domain containing proteins. Interestingly, the binding site for filamin (an actin-crosslinking protein) within the integrin cytoplasmic domain was shown to partially overlap with that of talin. These proteins compete for binding integrin tails, allowing integrin-filamin interac-tions to impact talin-dependent integrin activation [69].

Regulation of integrin activation By PIPK-I One important pathway in regulating integrin ligand binding and connection to the cytoskeleton involves the generation of phosphatidylinositol-4,5-bisphosphate (PI(4,5)P2 or PIP2) by the phosphatidylinositol phosphate kinase type I (PIPK-I). PIP2 is a membrane-associated phosphoinositide, markedly restricted to the cell surface plasma membrane rather than intracel-lular membrane compartments. It transiently associates with talin and pro-motes the talin-integrin association by recruiting talin to the membrane and enhancing talin’s affinity for the integrin 1-subunit [70]. PIPK-I is recruited to nascent focal adhesions and activated upon binding to the talin’s head FERM domain [71, 72]. Thus, a positive circuit may promote adhesion in which talin activates PIPK-I and PIP2 activates talin to bind integrins. How-ever the picture is not clear, since PIPK-I might also be involved in trans-dominant inhibition of 1 integrin-activation via competitive sequestering of talin [73] by binding to the same region of the talin FERM domain as the integrin -tails [74]. On the other hand, it is plausible that integrins and PIPK-I bind to a talin dimer simultaneously in an anti-parallel fashion. The PIPK-I -talin interaction is, in turn, regulated by two other important focal adhesion components, namely Src (phosphorylates PIPK-I on Y644) and focal adhesion kinase (FAK) [75] that will be discussed later. However, the significance of this and other kinds of competition events in vivo remains to be elucidated.

By Calpain Calpain cleavage of talin, separating the head and rod domains, sets out an-other possible regulatory pathway of talin-mediated integrin activation [76]. Calpain is a cytosolic cysteine protease that is activated by a rise in intracel-lular Ca2+ [77]. Isolated talin head domains bind integrins with higher affin-ity than intact talin. Both the PIP2 and calpain regulatory pathways are pro-posed to unmask the high-affinity binding site for integrins by releasing a masking interaction, either intra- or intermolecular, of the talin C-terminal rod with the FERM domain [51]. The non-phosphorylated N-terminal NPXY-motif in integrin -subunits was proposed to be the first target of talin binding in a two-step pathway of integrin activation [66], in which the talin head domain first recognises high-affinity binding sites in the mem-

24

brane-distal region of the resting integrin, with the subsequent binding to a second, lower affinity, membrane-proximal site which would unclasp the association of the tails.

By ICAP-1 Additionally, the integrin cytoplasmic domain-associated protein-1 (ICAP-1) is suggested to regulate 1-integrin inside-out signalling. It binds in a spe-cific manner to the 13 C-terminal amino acids of 1 cytoplasmic domain and the interaction requires the N792 and Y795 residues in the membrane-distal NPKY-motif as well as V787 [78, 79]. The ICAP-1 protein contains only one domain, a PTB-domain ( 30 kDa), and is expressed in at least eight human tissues and seven tested cell lines (noteworthy is the much more abundant expression in blood cells compared to fibroblasts) [79]. Recombi-nant expression of ICAP-1 in fibroblasts disrupts focal adhesions although, surprisingly, ICAP-1 cannot be observed in focal adhesions but co-localises with 1 integrin in ruffling edges [80]. The ability of ICAP-1 to displace vinculin, -actinin, and talin from focal adhesions is caused by its direct interaction with the 1 cytoplasmic tail. Since talin recruitment is a prerequi-site for focal adhesion assembly, ICAP-1 is suggested as an upstream regula-tor.

The diffuse cytoplasmic distribution of ICAP-1 [80] suggests that the in-teraction with the 1 cytoplasmic domain is regulated and multiple potential serine- and threonine-phosphorylation sites in ICAP-1 could serve in regula-tion. Indeed, ICAP-1 is heavily phosphorylated upon cell attachment to fi-bronectin [79]. However, mutational studies suggests that phosphorylation of T38 in ICAP-1 possibly is promoting the interaction with the 1 cytoplasmic tail [81], thus preventing talin binding and subsequent focal adhesion assem-bly. Blocking of the calcium/calmodulin-dependent protein kinase II (CaM-KII) activity preserves the high affinity state of the 5 1 integrin in vivo,and CaMKII together with the counteracting phosphatase calcineurin (also known as PP2B) was reported to control 5 1 mediated inside-out signal-ling through ICAP-1 [81, 82]. Taken together, ICAP-1 seems to be regulated by phosphorylation in a complex manner.

By phosphorylation of integrin The phosphorylation of integrin cytoplasmic tails has been proposed as means of regulating integrin activity, since a reversible and tight regulation can be achieved through kinases and phosphatases. Phosphorylation of subunits, as well as of some subunits, has indeed been demonstrated [83-85], but the phosphorylated state of the integrin often seems to be transient and has been quite difficult to “capture”. Strong stimuli and multipotent in-hibitors of phosphatases are widely used in the attempt to study phosphoryla-tion events, thus the physiological effects must be interpreted with caution.

25

Phosphorylation of integrins could possibly regulate the interactions with activating or inactivating cytoplasmic proteins. Among the many potential phosphorylation sites of the cytoplasmic domains the two conserved tyrosi-nes in the NPXY-motifs of the -chains (Figure 5) serve as potential sub-strates for Src-family kinases [86]. Phosphorylation of the tyrosine residue in the N-terminal NPXY-motif negatively regulates talin binding and may serve as a “molecular switch” to facilitate binding of alternative partners. The crystallisation of the talin head-domain in complex with integrin 3-peptide provided an explanation for the abolished binding of talin. This is because the talin PTB-domain lacks the arginine residues, common to sev-eral PTB-pockets, that otherwise coordinate the phosphate moiety of NPXpY ligands [87], and the NPXY-talin binding is depending on hydro-phobic interactions with the acidic and hydrophobic pocket of the talin PTB-domain. This is consistent with efficient binding of talin to NPXF-motif in

2 and a functional tyrosine to phenylalanine mutant of 1 [64]. Likewise, the interaction with ICAP-1 is thought to be abolished by phosphorylation of 1-Y795, since structural modelling of the ICAP-1 PTB-domain pocket also lacks an arginine residue and the space for a phosphate moiety is al-ready occupied by other residues [78]. However, it is unclear how the cell can control the outcome of such displacing phosphorylation events, since displacing talin would prevent activation and connection to the cytoskeleton, whilst displacing ICAP-1 would remove a (possible) negative regulator of talin. In the case of IIb 3, the C-terminal NITY759 sequence is essential for inside-out signalling but phosphorylation of Y759 is unlikely to be involved in this process [88].

In between the NPXY/F motifs there is a well-conserved serine-threonine cluster. Phosphorylation of these residues is postulated to regulate integrin affinity for ligands. It is most clear and thoroughly investigated for 2 in-tegrins. Stimulation of T-cells, either by phorbol ester or T-cells receptor (CD3)-ligation, activates the leukocyte 2 and 7 integrins. In response to these stimuli, both integrins are phosphorylated on threonine residues in the threonine triplet [89] (Figure 5). Many of the protein kinase C (PKC) iso-forms are activated during leukocyte stimulus and are, indeed, capable of phosphorylating 2 in vivo [90, 91]. Threonine-phosphorylated 2 integrins distribute preferentially to the actin cytoskeleton in vivo [92], possibly medi-ated by binding of 14-3-3 proteins [91]. Okadaic acid, used to inhibit PP1 (protein phosphatase 1) and PP2A (protein phosphatase 2A) inhibits T-cell adhesion [93].

Threonine phosphorylation of the 3 chain of IIb 3 integrins in platelets were shown as early as 1990 with 32P -labelling [85], and the exposure and closure of ligand binding sites correlates closely with phosphorylated and dephosphorylated residues, respectively [37]. Mutational analysis of threonine residues in 3 and 1 show the importance of this motif for ex-tracellular domain conformation, cell adhesion and spreading [36, 94, 95].

26

For 1 integrins there is so far no direct evidence for threonine phosphoryla-tion of the conserved cluster in connection with integrin activation. How-ever, replacement of both T788/T789 in 1 with alanine residues revealed the loss of the 1 extracellular conformation-dependent 9EG7-epitope and defective fibroblast attachment to fibronectin [36]. In the present investiga-tion (Paper I), T788 was identified as the site of major impact on the double-alanine substitution and introduction of the phospho-mimicking T788D mu-tation was investigated and found to both display the 9EG7-epitope as well as to mediate strong cell attachment. Other studies using mutational analysis of the 1-threonine cluster (in the context of interleukin 2 receptor chimeras acting as dominant negative constructs) demonstrate that the conserved threonine motif regulates cytosolic protein interactions with the 1 cyto-plasmic domain [94]. By use of phospho-site specific monoclonal antibodies,

1A was reported to be phosphorylated on T788/789 after okadaic acid (which is a quite broad specific serine-threonine phosphatase inhibitor) treatment of myoblasts [96]. The protein phosphatase PP2A was suggested to be responsible for the dephosphorylation activity and by comparison with okadaic acid-treated myotubes (mainly expressing the 1D isoform) the formation of focal adhesions was reported to be reduced when 1A was phosphorylated. In another study, also using phospho-site specific antibod-ies, the 1-subunit was reported to be phosphorylated on the two threonine-residues during mitosis, with reduced dependence of 1 integrins during the G2/M phase and weakened connection to actin [97]. However the relation-ship between phosphorylated threonines and active conformation of the 1integrin was not addressed in either study.

The cytoplasmic domains hold more potential phosphorylation sites outside the NPXY and S/T-cluster. The phosphorylation-state of S785 in the integrin 1-subunit modulates the localisation of 1 integrins and is regu-lated, at least in part, by the action of an okadaic acid-sensitive phosphatase [98]. A likely candidate phosphatase is PP2A since it associates with 1 at focal adhesion site and also co-immunoprecipitates with the 1-subunit from parietal endoderm [98]. When 1 is phosphorylated at S785 in these cells (as the only phosphorylated site) it is not found at fibronectin-induced focal adhesion sites [99]. However, adhesion to laminins is promoted by a phos-pho-mimicking mutation of S785 to aspartic acid, indicating a regulatory role for this serine residue in integrin-mediated functions [98, 99] and in-tegrin-linked kinase (ILK) was put forward as the candidate kinase for this phosphorylation site. This serine is conserved in 2, as well as 7, and 2-S756 is the main phorbol-ester induced serine phosphorylation site. How-ever, it remains an open question whether phosphorylation of S756 has a function in the regulation of the 2 integrin. Additional serine residues in the

-cytoplasmic domains, outside the serine-threonine cluster, that become phosphorylated in response to stimuli are so far not implicated in the inside-out activation of integrins.

27

The reversible nature of the switch from inactive to active integrin is well established but the mechanism(s) have not been investigated. The cytoplas-mic phosphorylation-dephosphorylation turnover as well as the extracellular thiol isomerase activity are possible mechanisms of repeated cycling be-tween active and inactive states of integrins.

Outside-in signalling Upon ligand binding to integrins new conformational changes are induced in the extracellular domain. These rearrangements are somehow transmitted to the cytoplasmic domain, which allows, “outside-in signalling” to take place. Suggestively, new binding sites for cytoplasmic proteins are created within a single cytoplasmic tail or between adjacent ones newly positioned next to each other. Several possible mechanisms could be envisioned, but the con-formational changes outside – in are far from understood today.

Clustering of integrin receptors Ligand binding of integrins induces clustering (aggregation) of several bound integrin receptors [100] (Figure 4). Clustering can be mediated via ligands with multiple integrin binding sites, as for fibronectin and collagen polymers, but possibly also by an intrinsic property of the integrin (i.e. the suggested homo-oligomerisation properties discussed earlier). Clustering is thought to increase the integrin avidity for ligand, similar to a Velcro closing where many weak interactions hold tightly when positioned close to each other. For the extracellular connections to hold tightly, the integrins must be firmly anchored to the cytoskeleton. This is achieved by recruitment of cyto-plasmic proteins into highly organised complexes known as focal adhesion sites (discussed in more detail at page 29).

Modulation of integrin cytoplasmic interactions Because there are numerous proteins interacting with the integrin cytoplas-mic tails, there must be a spatio-temporal regulation of these interactions by different mechanisms. Regulation of when and what partners to bind the integrin cytoplasmic tail is, like in the activation step, sometimes modulated by phosphorylation of tyrosine, serine and threonine residues. For example, thrombin stimulation of platelets induces tyrosine phosphorylation on Y747 and Y759 in the conserved NXXY-motifs of the 3-cytoplasmic tail, which is important for key platelet function like formation of stable aggregates and retraction of clots [101]. The cytoskeletal protein myosin and the signalling proteins Src homology domain containing protein (Shc) and growth-factor-receptor-bound protein 2 (Grb2) bind to phosphorylated 3-peptide in vitro[102, 103]. In addition, Shc co-immunoprecipitates with tyrosine-phosphorylated 3 integrins from platelets and has thus been implicated in the activation of Erk (extracellular signal-regulated kinase) signalling path-

28

way. Unlike the binding of myosin, which requires phosphorylation of both 3 NXXY-motifs [103], binding of Shc only requires phosphorylation of

Y759 [102]. Nevertheless, most proteins that bind to the cytoplasmic do-mains of IIb 3 do not require 3-tyrosine phosphorylation, e.g. 3-endonexin, talin, and integrin-associated protein (IAP).

For the other 3 integrin, V 3, phosphorylation of Y747 in the 3-chainhas been reported to be necessary for Arp3-organisation into adhesion sites [104]. The Arp2-Arp3 complex mediates initiation and growth of actin fila-ments, and phosphorylation-dependent Arp3 localisation is a possible mechanism to enable coordination of adhesion and cytoskeletal reorganisa-tion. Furthermore, in vitro actin polymerisation was induced by V 3 com-plexes purified from phorbol ester-stimulated cells, but strongly reduced by a Y747/759F mutation, phosphatase treatment, and inhibition of the tyrosine kinase Syk [105].

The 2-subunit contains phenylalanines instead of tyrosines in the NPXY-like motifs of 3 and 1, and hence is not modulated by phosphory-lation at these motifs. Tyrosine to phenylalanine mutations of the 1-subunit lead to impaired cell spreading and migration [106, 107]. In addition, these mutations result in defective FAK-mediated outside-in signalling via tensin and paxillin [107]. The 1-tail NPXY-motifs are tyrosine phosphorylated in cells that are transformed by v-Src, accompanied with lost adhesion to fi-bronectin and loss of coupling of focal contacts to actin-containing cy-toskeleton [86]. Single Y/F mutations suggest that overactive phosphoryla-tion of cytoplasmic residues of 1, particularly Y783, accounts in part for the phenotype of v-Src-transformed cells. In more general terms, phosphory-lations of the NPXY-motifs are known to abolish the interaction with several PTB-containing proteins, e.g. talin.

Phosphorylation of 3-T753 has been implicated in outside-in signalling via IIb 3 in platelets [108] and is predicted to inhibit the binding of Shc [109]. Additional studies revealed that both phospholipid-dependent kinase 1 (PDK1) and protein kinase B (PKB/Akt) can phosphorylate 3-T753 in vitro[109], but whether these are true kinases for 3-T753 in cells is not known. The available data indicate that threonine and tyrosine phosphorylation has both negative and positive regulatory roles in integrin function, possibly reflecting the diversity and dynamics of integrin-mediated events.

Outside-in signalling regulates two kinds of events, which are not entirely separable, namely actin cytoskeletal reorganisation and cytoplasmic signal-ling cascades. As the integrin receptors lack enzymatic activity they must contract other molecules to do the job, and the resulting onset of recruitment of cytoplasmic adaptor proteins, kinases and phosphatases and subsequent cellular events have been widely studied. The assembly of focal adhesions and major adhesion dependent events will be discussed next.

29

Focal adhesion sites– protein ensembles with inositol guest stars At the cytoplasmic face of the plasma membrane integrins bind to cytoskele-tal proteins of the actin contractile system. This can be clearly visualised in stationary cells in culture where integrins are concentrated in small patches at the ventral cell surface, in contact with the adhesive substratum. The patches are known as “focal adhesions”. A corresponding phenomena is presumably occurring in vivo and to mimic this more truly, adhesion de-pendent events are begun to be investigated by use of three dimensional-matrices instead of two-dimensional supply of extracellular ligands. These studies reveal some distinct differences [110], but still, most of the new lit-erature is based upon two dimensional-substrate experiments.

Sometimes (but only sometimes) colleagues in the field distinguish be-tween different forms of these protein aggregates, which can vary in size, composition, morphology, and localisation (as described in [111]). Briefly, the early formations of small dot-like adhesion plaques are termed focal complexes and include only a few components apart from integrins, like paxillin, vinculin and some tyrosine-phosphorylated proteins. These focal complexes are found mainly at the edge of lamellipodia and are induced by the Rho-family GTPase Rac. Focal complexes are associated with cell mi-gration and promote actin assembly. They are then thought to mature into focal adhesions, including additional components like -actinin and FAK. Focal adhesions are elongated and oval in shape and found at the periphery of the cell. These structures are more stable than focal complexes and linked to formation of actin stress fibres. Development of focal adhesions is stimu-lated by the small GTPase RhoA and is driven by the actomyosin contractil-ity. More rarely, a third kind of adhesion structure is mentioned as fibrillaradhesions, containing tensin and predominantly 5 1 integrins. They have fibrillar or beaded morphology as they typically are anchoring extracellular fibronectin fibrils. These structures are located in the central regions of cells. This is an attempt to clarify the literature, but of course there is no clear dis-tinction between the three adhesion structures.

Molecular components of focal adhesion sites There are more than 30 proteins identified in interactions with the integrin cytoplasmic domains (although the function and relevance of those are not always clear) and over 50 molecules suggested to localise in focal adhesions. I will here briefly describe some of the major constituents of focal adhesion ensembles.

30

Cytoskeletal proteins One striking feature of the focal adhesion associated cytoskeletal proteins is their multiple interactions with each other and other focal adhesion proteins [112], which potentially facilitates the assembly of protein complexes. Actinis the major cytoskeletal protein and one of the most abundant proteins in the cell. Actin monomers, of 43 kDa, are polymerised into long filamentous chains under the control of a large set of regulatory proteins [113]. Connec-tions from integrin cytoplasmic domains to the actin cytoskeleton occur mainly via the interaction of four actin-binding proteins, talin, -actinin, filamin, and tensin (Figure 6) [114-117].

Figure 6. The integrin-actin connection. Illustration of different pathways by which integrin can link to the actin cytoskeleton. The molecules are not drawn to scale. Adapted from [118].

Talin is an essential component of focal adhesions that couples filamentous actin to -integrin cytoplasmic domains (as shown for 1A, 1D, 2, 3, 5, and 7) [119]. Talin is a large protein, with a 50-kDa head region followed by a 220-kDa long rod region, and it assembles into an anti-parallel homodimer. Via the FERM domain in the head region talin binds to in-tegrins, FAK, PIPK-I , PIP2, non-integrin TM receptors, and weakly to actin

31

[70-72, 115, 120]. The talin rod domain contains a low affinity site for in-tegrin [76], three vinculin binding sites and a major actin-binding site at the C-terminus [121]. The talin-integrin interaction was discussed in more detail earlier (see page 22).

Talin associates with the PIP2 -producing enzyme PIPK-I and thereby regulates the local concentration of PIP2 (see page 23). This in turn, regulates the association of talin to vinculin [122], a ubiquitous 117-kDa cytoskeletal protein found at cell-cell and cell-matrix contacts. Binding of PIP2 to vincu-lin unfolds the inactive molecule and exposes binding sites for talin and

-actinin in the vinculin head domain and for the vasodilator-stimulated phosphoprotein (VASP) in the tail domain [123, 124]. Thus, the integrin-actin connection is further stabilised by the recruitment and activation of vinculin. Vinculin links to the second -tail associated actin binding protein,

-actinin, an antiparallel homodimer, which holds the actin filaments to-gether in loose bundles. The 103-kDa -actinin also serves as an important scaffolding protein by interacting with multiple kinases (Erk1/2, MEKK1, and the p85 subunit of PI3K). These interactions are modulated by PIP2 and PIP3 concentrations, and FAK dependent phosphorylation of -actinin de-creases its association with actin [118].

The third -tail associated actin binding protein filamin is, in contrast to talin and -actinin, a homodimer of parallel composition [125]. Filamin links actin filaments in orthogonal networks through an actin-binding site in the head domain. Via different repeat units in the rod, filamin can associate with integrins, kinases (e.g. MEKK1 and PAK1), Rho-GTPases and GEFs (guanine nucleotide exchange factors) [118]. Thus, by providing a platform for simultaneous binding of both activators and effectors, filamin can induce reorganisation of the cytoskeleton. As mentioned, the binding of filamin and talin to integrin is competitive [69].

Paxillin is an adaptor protein early recruited to nascent focal complexes, with important functions in matrix adhesion assembly, turnover as well as signalling. Although relatively small, this 68-kDa protein contains many protein binding modules (five N-terminal LD-repeats and four C-terminal LIM-domains) in addition to proline-rich sequences and several phosphory-lation sites [126]. FAK, in association with Src, phosphorylates paxillin at Y31 and Y118 that become high affinity binding sites for the adaptor protein Crk [127]. Crk couples paxillin to regulation of both motility and gene ex-pression (via mitogen-activated (MAP) kinases) [128]. Paxillin is also phos-phorylated on several serine and threonine residues, which are important for localisation [129]. The LIM-domains interacts with both kinases and phos-phatases which regulate paxillin targeting, for example PTP-PEST (protein tyrosine phosphatase-proline, glutamate, serine, and threonine sequences) is transiently targeted to focal adhesions by paxillin thereby facilitating the dephosphorylation of the main PTP-PEST substrate, Cas (Crk-associated substrate), an important event for focal adhesion disassembly [126].

32

The fourth -integrin and actin binding protein tensin is a 185-kDa adap-tor protein that is translocated to the assembling focal contact. Tensin, via its PTB-domain, binds well to the NPXY-motif of 3 and 5 cytoplasmic tails and weakly to 1 and 7 [87]. Tensin also contains an Src homology 2 (SH2) domain, which allows it to bind tyrosine-phosphorylated phosphatidy-linositol-3-kinase (PI3K), Cas and FAK [114].

Interestingly, the muscle-specific 1D-integrin cytoplasmic domain binds talin more tightly than either the more widely expressed 1A subunit or the leukocyte-specific integrin 7. To add complexity, the 7 tail binds filamin more strongly than the 1A tail, and the 7 tail supports less migration than either the 1A or the 1D tails, although all three support similar extent of cell adhesion [130]. Such differences in strength of integrin binding to dif-ferent cytoskeletal proteins may partly explain the diverging abilities of dis-tinct integrins to mediate processes such as cell migration, ECM assembly, and focal adhesion formation.

Signalling proteins and their regulators The lack of intrinsic catalytic activity forces the integrins to cooperate with cytosolic as well as receptor kinases. Among the cytosolic kinases I will focus on two tyrosine kinases, FAK and Src, and three serine-threonine kinases, ILK, PAK and PKC.

Tyrosine Kinases In 1992, the non-receptor tyrosine kinase FAK (focal adhesion kinase) was identified as an adhesion-activated kinase [131, 132]. FAK is localized to focal adhesions and rapidly becomes tyrosine-phosphorylated following integrin ligation. The activation of FAK is thought to involve autophos-phorylation at Y397, which creates a binding site for the SH2-domain con-taining tyrosine kinase Src [133]. The tight proximity of FAK and Src facili-tates phosphorylation of additional tyrosine residues of FAK, including Y576 and Y577 in the activation loop, which enhances the catalytic activity of FAK [134]. Phosphorylation of Y397 also appears to be important for the recruitment of other SH2-containing proteins like PI3K, PLC- , and the adaptor protein Grb7 [135-137]. In addition, Grb2 binds both to phosphory-lated Y397 and Y925, coupling FAK to activation of the MAPK cascade. FAK also contains proline-rich sequences that serve as binding sites for the Src homology 2 (SH3) domains of Cas [138] and GAPs (GTPase-activating protein for Rho and Arf1/6) [139]. Cas is a multifunctional adaptor protein with several tyrosine residues phosphorylated by FAK and Src (see page 37), and the GAPs couples focal adhesions to the regulation of cytoskeletal reor-ganisation. The C-terminal region of FAK holds a sequence designated “FAT” for focal adhesion targeting, which directs FAK to newly formed and existing adhesion complexes [140].

33

The N-terminal region of FAK contains a FERM domain, of which the crystal structure was recently solved [141]. Briefly, the tertiary folding of the three subdomains (F1, F2, and F3) is similar to those of known FERM struc-tures despite the low sequence conservation. Differences in sequence and relative orientation of the F3 subdomain result in loss of the phosphoinosi-tide-binding site found in other FERM structures. Associative features of the adjacent linker region, containing Src-SH3 and -SH2 binding sites, with the FERM domain were found. Such binding suggests the possibility that bind-ing of Src kinases to the linker region can regulate protein interactions of the FAK-FERM domain. The multitude of interacting partners of FAK raises the question of when and how do different effectors bind FAK? Perhaps, activa-tion of FAK in different contexts, i.e. depending of the structural composi-tion of FAK surrounding molecules, can control the physiological outcome of FAK signalling.

The functional importance of FAK in integrin mediated signalling has been extensively studied, some of the signalling pathways have been men-tioned elsewhere in this thesis, and the field is thoroughly reviewed in [139, 142, 143]. The generation of FAK-deficient mice and cell line in 1995, re-vealed the importance of FAK in cell migration and that this was likely due to defective focal adhesion turnover leading to an elevated number of focal adhesion sites [144].

The FAK homologous kinase proline-rich tyrosine kinase 2 (PYK2; also known as CAK , RAFTK, FAK2, or CadTK) has high structural similarity to FAK except that PYK2 lacks the C-terminal FAT domain of FAK. PYK2 does not localise to focal adhesions to a high degree, it rather has a diffuse distribution throughout the cytoplasm and is concentrated in the perinuclear region, but still it becomes activated upon integrin-mediated adhesion [145]. This protein is primarily expressed in the central nervous system and in cells derived from hematopoietic lineages as opposed to the ubiquitously ex-pressed FAK [146]. PYK2 is responsive to intracellular calcium levels and stress signals, and is involved in several signalling pathways (JNK/SAPK, Erk), some of which are intersecting with those induced by FAK. Together with Src-family kinases, PYK2 can activate the Erk-pathway in the absence of FAK, but cannot fully compensate for FAK to overcome migratory de-fects [147]. Expression of PYK2 was elevated in FAK-deficient fibroblasts, compared with cells expressing FAK, and integrin binding to fibronectin was reported to induce tyrosine phosphorylation of PYK2 in the FAK-deficient cells [147]. Interestingly, we observed up-regulated PYK2 protein levels in cells expressing 1B-integrins, defective in activating FAK, but efficient adhesion-induced PYK2-phosphorylation in 1A-cells (unpublished results). These data suggest that the C-terminal half of the 1 subunit is important for PYK2 activation.

In addition to FAK, members of the Src-family of nonreceptor protein-tyrosine kinases (SFKs) also associate with focal adhesions, mainly through recruitment by FAK [136]. These tyrosine kinases are also widely involved

34

in integrin signalling. The proto-oncogene src was the first gene to be identi-fied that is potentially capable of inducing cell transformation [148, 149]. Unlike the viral form of Src, the cellular Src kinase activity is regulated by an inhibitory binding mechanism, where Y527 is phosphorylated, by the Src-related tyrosine kinase Csk (C-terminal Src kinase), providing a medium affinity-binding site for the SH2 domain of Src itself [150]. The intramolecu-lar interaction folds the Src protein in a closed inactive conformation. In addition, the SH3 domain interacts with the proline-rich linker region be-tween the SH2 and kinase domains [150]. Close presentation and competi-tion of high affinity sites for the SH2 domain releases the inhibitory binding and facilitates the dephosphorylation of Y527 by RPTP , PTP-1B, or SHP2 (see page 35). In addition, phosphorylation of Y416 in the activation loop is necessary for complete activation of the kinase, and the list of target proteins of its kinase activity is long [151]. It should be noted that the above-mentioned mechanism not only precisely regulates the kinase activity of Src proteins, but also controls the interactions of the SH2 and SH3 domains with other molecules, thus providing different levels of Src regulation. Src is pro-posed to be necessary for optimal adhesion but not required for its initiation process [151]. Src and FAK appear to, as a result of the binding of the Src SH2 domain to the autophosphorylation site of FAK, act jointly in signalling from integrin receptors.

Abl is a tyrosine kinase that exists both in the cytoplasm and in the nu-cleus. The cytoplasmic pool of Abl become activated upon integrin-mediated adhesion and is then transiently located to focal contacts. Transforming vari-ants of Abl (Bcr/Abl) can overcome the anchorage-dependence for growth, but not the growth factor-dependence [152, 153].

Serine-Threonine Kinases Turning to the serine-threonine-kinases found at focal adhesion sites, ILK(integrin-linked kinase) binds directly to 1 and 3 cytoplasmic tails [127]. In addition to its kinase activity, ILK has an important function as an adaptor protein with binding sites for paxillin, parvins, and Pinch-1/2. Binding to paxillin is necessary but not sufficient to recruit ILK to focal adhesions. Fo-cal adhesion and cytoskeletal proteins phosphorylated by ILK include ILK itself, -parvin, and -integrin cytoplasmic tails [127]. The ILK-Pinch-parvin complex is pivotal to the functions of these proteins and is involved in focal adhesion formation and turnover [154].

The p21-activated kinases (PAKs) 1-3 are serine-threonine protein kinases whose activity is stimulated by the binding of active Rac and Cdc42 GTPases (but also by some GTPase-independent mechanisms) [155]. PAK serve as an important regulator of many cellular signalling pathways, such as cytoskeletal rearrangements, MAPK activation, and cell cycle progression. PAKs can interact via their proline-rich segments in the regulatory domain with SH3-containing adaptor proteins like Nck and Grb2, which recruit PAKs to activated tyrosine kinase receptors at the plasma membrane [155].

35

Membrane recruitment of PAK results in localisation in the proximity of the phosphoinositide-dependent kinase PDK1, which phosphorylates PAK pri-marily at the critical T423 residue and promotes its activation. PAK has been suggested to participate in the regulation of integrin signalling to Erk at the control steps of cRaf and Mek [156, 157]. In addition, protein kinase A (PKA), which has negative role in regulation of cRaf activity [158], was shown to phosphorylate PAK1 and this inhibits PAK activity [159].

PKC (protein kinase C) was one of the first signalling molecules identi-fied in focal adhesions [160], and it is established that PKC isoforms become activated following adhesion and cell spreading. In particular, the novel class of PKCs are highly phosphorylated as long as cells are adherent [161]. In addition, several PKC enzymes are under allosteric control of lipid second messengers (PI(3,4)P2 and PI(3,4,5)P3) which also recruit the enzymes to the plasma membrane; thus in this way the PKC and PI3K activities are coupled. Maximal phosphorylation and activation also depend on PDK1 [162]. The integrin-PKC relationship seems to be bi-directional since PKC has been reported to participate both in inside-out and outside-in signalling. The rele-vance of PKC for integrin functions is supported by the finding of a direct association of the 1-integrin tail with the V3 variable region of PKC [163, 164], which correlates with 1-integrin dependent migration. Interestingly, the association occurs preferentially to integrins with active conformation (12G10-positive) rather than with integrins unoccupied by ligand (Mab13-positive), and the association is promoted by PKC activation [164]. More-over, cellular trafficking of 1-integrins is suggested to be under the control of the PKC and PKC isoforms [161, 165]. Interestingly, receptor for acti-vated C-kinase (Rack1) serves as intracellular receptor for some PKC iso-forms (as shown for PKC , , and ) and shuttles them to their appropriate cellular location. Rack1 also binds to the 1, 2, 3 and 5 cytoplasmic tails, but only after the activation of PKC [166]. As Rack1 is also shown to be part of the small ribosomal subunit [167, 168] as well as being localised to so called “spreading initiation centres” at the early stages of spreading (but not to mature focal adhesion sites, [169]). The data suggest that the docking of ribosomes to focal contacts may be needed for local translation of specific proteins in the early establishment of focal adhesions [170]. This latter example link PKC and integrins in yet another aspect.

Phosphatases In contrast to the extensive research about kinases and their activities, the counter-actors, the phosphatases and the dephosphorylation events, have not been as much in line of investigation. In controlling the strength and duration of integrin signalling the phosphatases must, of course, be of equal impor-tance in regulating protein interactions. Src-family kinases, FAK, Cas and paxillin are all dephosphorylated by a number of protein-tyrosine phosphata-

36

ses (PTPs), including PTP-PEST, SHP2, PTP-1B, and RPTP as discussed below.

One phosphatase stimulated by integrin-mediated adhesion is PTP-PEST(protein tyrosine phosphatase-proline, glutamate, serine, and threonine se-quences), which acts on focal adhesion targets such as Cas and paxillin, but not on FAK [171]. Paxillin is essential for PTP-PEST inhibition of cell spreading and membrane protrusion as well as inhibition of adhesion-induced Rac GTPase-activation [172]. PTP-PEST also requires paxillin for stimulation of cell migration. Interestingly, both overexpression and deletion of PTP-PEST inhibit cell migration [173, 174], suggesting that a fine-tuned balance of tyrosine phosphorylation of relevant substrates regulates cell mi-gration.

Phosphatases that have been implicated in regulation of FAK phosphory-lation include SHP2 (SH2-domain-containing protein tyrosine phosphatase 2) and PTEN (phosphatase and tensin homologue deleted on chromosome ten). SHP2 is a widely expressed cytoplasmic tyrosine phosphatase with two SH2 domains, associating with phosphorylated SHPS-1 (SHP2-substrate-1), a receptor-like glycoprotein, and translocated to the plasma membrane after integrin ligation. SHP2 regulates FAK-mediated migration but also the activ-ity of Rho GTPases [171]. In addition, SHP2 promotes Src-family kinase activation by dephosphorylation of the membrane-bound Csk regulator Cbp (Csk-binding protein), thereby controlling Csk access to Src-family kinases and in activation [175, 176]. Thus, this phosphatase seem to control the ac-tivity of the central FAK/Src complex at several levels.

PTEN functions mainly as a lipid phosphatase, but it also displays weak tyrosine phosphatase activity towards FAK and Shc [177], whereas a number of other tyrosine-phosphorylated proteins appear unaffected. It has been shown that PTEN, by negatively regulating FAK activity as well as dephos-phorylating PIP3, down-regulates the matrix-dependent PI3K-Akt cell sur-vival pathway [178].

Another PTP implicated in integrin signalling is PTP-1B (protein tyrosine phosphatase 1B), which was the first PTP that was identified and cloned [179, 180]. In addition to dephosphorylating insulin and growth factor recep-tors, PTP-1B is required for the efficient transduction of integrin-mediated signals. Loss of PTP-1B delays fibronectin-induced cell signalling and cell spreading. In PTP-1B-deficient fibroblasts, a transient hyperphosphorylation of Y527 on Src was observed, suggesting that PTP-1B is involved in Src activation during adhesion [181]. PTP-1B associates with the ER through a C-terminal hydrophobic sequence with the catalytic domain facing the cyto-sol [182]. PTP-1B can be localised to the distal tip of newly formed cell-matrix adhesion sites through the dynamic activity of the ER and associated microtubules, and may play a positive role in the maturation of new focal adhesions [182]. Recently, PTP-1B was reported as an -actinin phos-phatase, which cooperatively with phosphorylated -actinin disrupts the

37

tight FAK-Src complex enabling dephosphorylation of FAK-Y397 by PTP-1B [183].

It is apparent from studies in knockout mice and diverse cell lines that the major action of RPTP (receptor protein tyrosine phosphatase ) is as a positive regulator of Src-family kinases. RPTP dephosphorylates and acti-vates Src [184] and is required for early integrin-proximal events that control the timing and efficiency of phosphorylation of FAK-Y397 [185]. The proc-esses of migration, haptotaxis, and spreading are impaired in RPTP -deficient cells, and the association of FAK with Src and Fyn is reduced [185]. In a recent study, two roles of RPTP in integrin signalling are distin-guished; an early role as an activator of Src-family kinases and FAK that does not require RPTP phosphorylation, and a later downstream role in cytoskeleton-associated events for which RPTP phosphorylation at Y789 is essential [186].

AdaptorsAdaptor proteins are signalling molecules without enzymatic activities. In-stead, they contain binding domains for other proteins, thereby linking the signal from one molecule to others. Common binding domains among adap-tor proteins, as well as enzymes, are SH2, SH3, PTB, and PH domains (sev-eral others are identified but appear more restricted to a certain family of proteins). In addition, adaptors can contain amino acid sequences of prefer-ence to the mentioned binding domains: motifs with phosphorylated tyrosi-nes (SH2 or PTB), motifs with non-phosphorylated tyrosines (PTB), or proline-rich motifs (SH3). I will here describe shortly four adaptor proteins important for integrin signalling found in focal adhesions: Crk, Cas, paxillin, and Shc.

As mentioned, Crk is recruited to focal adhesion by binding to tyrosine phosphorylated paxillin and Cas. Crk is a 33 kDa protein with one SH2 and two SH3 domains and an Abl phosphorylation site in between the two SH3 domains. Phosphorylation by Abl negatively regulates the interactions of Crk by causing an intramolecular binding of the Crk-SH2 domain to the phosphotyrosine, thereby preventing SH2 interactions with other proteins [187]. Via the two SH3 domains Crk bind several proteins to promote forma-tion of signalling complexes upon integrin ligation. Such a protein is the Rac1-GEF DOCK180, which in complex with Cas and Crk influences cell spreading [188].

Tightly associated with Crk is Cas (Crk-associated substrate) a scaffold-ing protein recruited to focal adhesions upon integrin ligation and important for cell spreading and migration [189]. This 130 kDa protein was named for its association with Crk (after its phosphorylation by Src). The N-terminal SH3 domain regulates localisation to the focal contact through interaction with FAK. The so called substrate domain contains 15 potential tyrosine phosphorylation sites and mediates interaction with the SH2 domains in e.g.

38

Crk, Nck, and Grb2. Cas is a target for phosphorylation both by FAK-dependent and -independent pathways after integrin ligation [190] and many mechanisms have been suggested. The FAK-independent pathway is inhib-ited by wortmannin, implicating that it requires an active PI3K. Whereas, the FAK-dependent phosphorylation of Cas is mainly mediated by Src [191]. In addition, Cas hold two proline-rich sequences mediating protein-protein interactions with SH3 domains and one serine-rich sequence.

It should be noted that the previously discussed cytoskeletal adaptor pro-tein paxillin also is of great importance as an adaptor in integrin signalling (reviewed in [126]). Paxillin is involved in several large protein signalling complexes such as the ILK-Pinch-Parvin (recently reviewed in [192]) and PKL-PIX-PAK [193, 194]. In addition, paxillin coordinates the proximity of the kinases Raf-1, Mek and Erk [195] by binding all three in response to hepatocyte growth factor (HGF) and Src phosphorylation of paxillin-Y118. Activated Erk phosphorylates paxillin-S83, which recruits FAK to the scaf-fold and subsequently PI3K and Rac are activated [195], which results in the extension of lamellipodia.

The adaptor protein Shc (Src-homology containing protein; isoform ShcA) is ubiquitously expressed and found in three splice variants (p46, p52 and p66), all three possess two distinct domains that bind phospho-tyrosine containing sequences (a PTB and a SH2 domain) and a central region (CH1) holding critical tyrosine phosphorylation sites [196]. The p66 isoform con-tains an additional CH-like region at the N-terminal. Shc has been implicated in signalling downstream of many different types of receptors, e.g. receptors for growth factors and cytokines, G-protein coupled receptors, and integrins, and it is one of the main substrates phosphorylated in Src-transformed cells or cells stimulated with growth factors. Shc interacts via its proline-rich re-gion with the SH3 domain of the Src-family proteins. Independent of the Src/FAK complex, Shc plays a role in signalling from integrin receptors to the Ras-MAP kinase signal transduction pathway, where it is suggested to connect the signal from the integrin subunit, via the Src-family kinase Fyn and provide binding site(s) for the Grb2/Sos Ras-activating complex [197, 198].

Additional proteins found at focal adhesion sites are modulators of small GTPases and proteases [199].

Signalling lipids and their regulators The phosphatidylinositides PIP2 and PIP3 serve as substrates for phospho-inositide kinases, phosphatases, and phospholipases, in addition to affecting the activities of several enzymes. PIP2 is well known to promote actin po-lymerisation directly in several ways, e.g. by interacting with profilin, an actin monomer sequestering protein, and by interfering with the action of capping proteins [162]. Also PIP3 is an important regulator of actin polym-erisation, but PIP2 and PIP3 are suggested to differ in respect to distribution, concentration levels, and timing. PIP2 is suggested to be largely restricted to

39

the plasma membrane, where it is present at relatively high levels and ho-mogenously distributed in most cells. However, PIP3 levels are low in un-stimulated cells and a rise in PIP3 levels alone is sufficient to induce actin assembly. The kinetics and spatial dynamics of PIP3 production correlate well with that of stimulus-induced actin polymerisation [200].

Phosphatidylinositide Kinases Among the large family of kinases phosphorylating different positions in the inositol ring of phosphatidylinositides, two kinases of importance for in-tegrin signalling are PIPK-I and PI3K. Phosphoinositide phospholipids are ligands for plextrin homology (PH) domains and recruit PH domain-containing proteins, such as PKB/Akt, to the plasma membrane and are im-portant for their activation [201, 202].

PIPK-I (phosphatidylinositol phosphate kinase type I) is responsible for PIP2 formation at nascent focal adhesions and is recruited and activated by talin. Src regulates PIPK-I activity by tyrosine phosphorylation of PIPK-I, leading to PIP2 formation and elongation of actin filaments [75]. Interest-ingly, the Src-phosphorylated PIPK-I is a substrate for SHP1, and SHP1 modulates both the association between PIPK-I and talin and the targeting of PIPK-I to focal adhesions [203].

PI3K (phosphatidylinositol-3-kinase) phosphorylates the 3 position hy-droxyl group of the inositol ring of phosphatidylinositides to produce PI(3)P, PI(3,4)P2, and mainly PI(3,4,5)P3, which are potential second messengers acting on effector molecules of diverse types (e.g. protein kinases, phosphol-ipases, and G-proteins) [204]. The classical PI3Ks (class I) are heterodimers composed of a p110 catalytic and a regulatory subunit (typically p85). The latter subunit contains two SH2 domains and sometimes an SH3 domain [205]). PI3K promotes migration (through binding to tyrosine phosphory-lated FAK-Y397 [206]), cell survival, proliferation, and growth factor recep-tor signalling pathways [205].

Phosphatidylinositide Phosphatases Similarly to PI3Ks, there is a large family (growing) of phosphatidylinositol phosphatases, of which some are better characterised than others. SHIP1 and SHIP2 (SH2-containing inositol 5-phosphatase 1 and 2) are important regu-lators of PI(3,4,5)P3 as well as PI(4,5)P2, as these enzymes are capable of dephosphorylating both substrates by removing the phosphate group of the 5OH position [162]. SHIP2 participates in growth-factor dependent signal-ling downstream of receptor tyrosine kinases by coupling to the adaptor molecules Grb2 and Shc. SHIP1 has a restricted distribution to hematopoi-etic cells, compared to the more ubiquitously expressed SHIP2, and partici-pates in cytokine-dependent signalling [162]. Among the 3OH phosphatases the dual specificity phosphatase PTEN (phosphatase and tensin homologue deleted on chromosome ten) has gained major interest because of its central role in suppressing malignancy. PTEN dephosphorylates PI(3,4,5)P3 at the 3

40

position in the inositol ring and thereby control one of the central cell sur-vival pathways; in addition it has the previously mentioned protein tyrosine-phosphatase activity [162].

Altogether, one should note that sometimes phosphoinositides seem to regulate the integrin activation phase and sometimes they act in response to integrin-mediated adhesion. In the focal adhesion ensemble – who is really the conductor?

Adhesion dependent and integrin regulated cell functionsAdhesion of cells to the surrounding ECM and to other cells is essential for tissue integrity. In addition to this structural function, cell adhesion induces intracellular signals that regulate proliferation, apoptosis, cell polarity, dif-ferentiation and other processes. Some of the major effects of integrin out-side-in signalling are illustrated in Figure 7 and discussed below.

Figure 7. The major effects on cell behaviour mediated by integrins, often act-ing in concert with other cell surface receptors.

41

Spreading and organisation of the actin cytoskeletonFormation of focal adhesions mediated by integrins causes the cytoskeletal proteins to become organised. The organisation of actin into networks and higher-order structures is crucial to both form and function of cells. Integrins activate signalling cascades that affect both formation and turnover of the actin filaments by stimulation of Rho-GTPases. These molecules are essen-tial and promote specialized actin structures. Cdc42, Rac, and RhoA three members of the Rho family of small GTPases, each control a cascade. Cdc42 triggers filopodial extensions at the cell periphery, Rac regulates the forma-tion of lamellipodia protrusions and membrane ruffles, and RhoA regulates stress fibre and focal adhesion assembly [207].

Integrins can stimulate Rho-GTPases via different pathways, of which those via FAK and Src-like kinases seem to be most important [208, 209]. In addition, the integrin-associated molecules ICAP-1 and ILK are suggested to stimulate Rho-GTPases [210, 211] and both molecules affect the linkage of integrins to the cytoskeleton [80, 212].

Migration & integrin cell surface recycling The most obvious role for integrins in cell motility involves their ability to interact with the ECM and thus allow formation of new contacts. The proc-ess of cell migration is initiated by protrusion (driven by actin polymerisa-tion) at the leading edge of the cell, the formation of peripheral adhesions, the exertion of force on these adhesions, and finally, the release of the adhe-sions at the rear of the cell. Inevitably, this complex process is regulated by multiple signalling mechanisms. Members of the Rho family of GTPases play a central role in the regulation of the cytoskeleton as mentioned above. Coordination of Rho GTPase activity is required for efficient migration where Rac specifically controls cell protrusion, while Cdc42 controls cell polarity [207]. In addition, many other signalling components influence the actin-myosin machinery, including FAK, Src, and PI3-kinases and the Erk and JNK pathways [213]. Integrins modulate many of these signalling com-ponents, thus the role of integrins in cell movement is not merely structural but also includes as signalling aspect. For coverage of the known details in this complex issue, the reader is referred to the following reviews [214-216].

It is well established that membrane trafficking pathways, which recycle adhesion receptors, contribute to cell migration. The mechanisms to move integrins from the trailing edge toward the advancing lamellipodia are likely keys to the speed and directionality of cell migration. The integrin distribu-tion is clearly polarised in migrating cells and there is good evidence that endosomal transport contributes to this [217]. However, the precise way in which integrin vesicular transport acts to drive the heterodimers toward the advancing lamellipodia is not yet clear, nor is there definite evidence for increased internalisation of integrins at the cell rear. Instead, data support a

42

model in which the rates of endocytosis and recycling of integrins are both elevated at the leading edge, and possibly contributing to the high rate of focal contact assembly/disassembly that occurs in this region [217].

Integrins are continually internalised from the plasma membrane into en-dosomal compartments and then recycled back to the cell surface [217]. The integrin recycling is a highly efficient process where most integrins are re-turned, rather than being targeted for degradation, and is a rapid process with a completed endo-exocytic cycle once every 30 minutes [208]. Furthermore, the process seems to be selective with certain heterodimers being rapidly cycled whilst others remain longer at the cell surface [218]. The internalisa-tion mechanisms also differ; the V 5 receptor is recruited to clathrin-coated pits (as for the low density lipoprotein (LDL) receptor), whereas the recycling of 1, 2, and 3 families of integrins involve caveolin or com-partments related to the detergent-resistant membrane microdomains (DRMMs) [219-222].

The return of receptors to the plasma membrane can occur by at least two distinct routes: a short-loop or a long-loop back from the early endosomes where the receptors are first delivered. In the early endosomes, key decisions are made concerning the fate of internalised proteins. The short-loop recy-cling path immediately returns receptors to the membrane, while others are passed to the perinuclear recycling compartment (PNRC) and from there returned to the plasma membrane. The short-loop and long-loop pathways are under the control of Rab4 and Rab11 GTPases, respectively [217]. Inter-estingly, certain signalling pathways act to impose selectivity on the Rab11-dependent transport of integrins. PKB/Akt acts via the phosphorylation of GSK-3 to promote the return 5 1 and V 3 integrins from the Rab11 compartment to the plasma membrane but has no effect on transferrin recep-tor (TFN-R) recycling that passes through the same compartments [223]. Similarly, PKC is suggested to regulate a selective return of 1 integrins from PNRC without influencing TNF-R transport [165]. Directed selective return of V 3 (but not 5 1) integrin via the Rab4-regulated short-loop is promoted by platelet derived growth factor (PDGF)-treatment of fibroblasts and requires the activity of PKD1. Whereas, mutants of V 3 that cannot bind to PKD1 do not recycle in response to PDGF [208, 224].

Survival Integrins mediate two different kinds of signals resulting in “adhesion-dependent growth”. First, integrin ligation will intersect the control of mito-gen activated signalling pathways (see next section “Crosstalk with other receptors”). Second, many cells will undergo a form of apoptosis, known as anoikis, if their integrin-mediated adhesion is lost. The process of anoikis is thought to prevent cells that have not reached, or lost, their proper location within the tissue from surviving. Anoikis is one of the control mechanisms

43

overcome by oncogenic transformation in cancer cells [225]. The capacity of individual integrin receptors to protect cells from anoikis is different (and

6 4 even promote growth arrest and apoptosis) [226]. Even though the mechanism which cells achieve survival seems to vary with cell type and integrin receptor involved, it seems to require phosphorylation of Y397 in FAK and its kinase activity [227]. Active FAK then induces different sur-vival pathways e.g. via the activation of PI3K-Akt in glioblastoma cell line [228] or the JNK pathway in primary fibroblasts [229]. Importantly, anoikis can be prevented by expression of constitutively active FAK mutants.

Crosstalk with other cell surface receptors Integrin binding to ECM starts multiple signalling cascades, such as activa-tion of FAK, PI3K, the Ras-Erk pathway, and Rho-family proteins [230, 231]. These signals overlap to a large extent signals activated by receptor protein tyrosine kinases (RPTKs) after binding of soluble growth factors and cytokines. Crosstalk between integrins and RPTKs control the extent and duration of each signal. The outcome depends on whether the integrin, the RTPK, or both are engaged by ligand. Integrin ligation induces cell motility and some protection against apoptosis. However, the receptors need to coop-erate to promote cell proliferation and optimal cell migration and survival [232]. For example, input from both integrins and growth factor receptors is required to stimulate the progression through the G1 phase of the cell cycle, via the induction of G1 cyclins and suppression of inhibitors of the G1 cyclin-dependent kinases [233]. The effects of RPTK signals are often transient, due to down regulation by endocytosis. Whereas integrin signals are more likely to persist over time, since most integrin ligands are not diffusible and integrins have relatively long half-life at the plasma membrane. Interestingly, the cell’s “choice” of engaged integrin receptor can alter the response to soluble factors very dra-matically, as for example, endothelial cells were reported to proliferate in response to mitogenic growth factors if they adhere to the 5 1 ligand fi-bronectin, whereas they undergo growth arrest if they adhere to the 2 1ligand laminin-1 [197]. Coordination or integration of adhesion and growth factor signals is facilitated by the close physical proximity of key molecules in the signalling cascades. Indeed, integrins co-precipitates with several growth-factor receptors, including the epidermal growth factor receptor (EGFR) and PDGF-receptor [234, 235]. In addition, several of the focal ad-hesion proteins have a role in both signalling cascades from integrins as well as growth factor receptors, e.g. paxillin [126].

44

Present investigation

Aims of thesis This project concerns structural and functional aspects of integrin 1-subunit mediated signalling. The aim has been to identify structures in the cytoplas-mic domain that regulate activation of the receptor and adhesion-induced signal transduction. For the latter aim, the contribution of two 1-integrinvariants, 1A and 1B, in two major signalling pathways was investigated, with focus on the importance of focal adhesion kinase (FAK).

The present studies were performed using GD25 cells, which are 1-integrin knockout mouse fibroblasts [236], and stably transfected clones expressing wildtype isoforms or mutant variants of the integrin 1 subunit. The GD25 cells provide an excellent model system for detailed studies of variant 1-integrin properties and function in the absence of endogenously expressed 1 subunits.

Results and Discussion Integrin activation– depending on common features of subunits? The suggested common activation model of integrins involving integrin cy-toplasmic domain separation is widely accepted. It proposes that the and cytoplasmic tails of the inactive integrin associate to form a clasp, maintain-ing the receptor at a low-affinity state. The cytoplasmic tails get separated upon binding of talin, or possibly other similar cellular activators, during inside-out activation. The separation of the membrane-proximal regions of the cytoplasmic tails is thought to induce a conformational rearrangement of the integrin extracellular domains leading to high affinity for ligand. This model was a big step forward in our understanding of integrin activation, but still it explains only a part of the process. In particular, it does not include the mechanism(s) by which binding of cellular activators (e.g. talin) or in-hibitors is regulated.

Phosphorylation of integrin cytoplasmic tails has been proposed as means of regulating integrin activity, since a reversible and tight regulation can be achieved through kinases and phosphatases. The leukocyte specific 2 in-

45

tegrins are well studied in the context of phosphorylation-induced integrin activation. The non-adherent T-cells respond to extracellular stimuli, like CD3-ligation or phorbol esters, by phosphorylating the threonine cluster and at the same time activating their 2 integrins [89]. For the platelet-specific

IIb 3 integrin, phosphorylation of the TST-motif in 3 has been shown, and phosphorylated and dephosphorylated residues correlate closely with the exposure and closure of ligand binding sites, respectively [37].

In contrast to 2 and 3, there is so far no direct evidence for threonine phosphorylation of the conserved cluster in 1 integrins, why is this? Does the lack of strong evidence for phosphorylation of the threonine cluster in 1reflect a non-occurring event, a very rare event, or a transient event that is strictly spatio- and temporally regulated? The difficulties of capturing a phosphorylated 1-threonine cluster have lead to alternative approaches to study the possible functional importance of these residues.

Paper I – Extracellular activity of 1 integrins is regulated by the intracellular residue threonine 788 We, and others, have tackled the question of possible 1-threonine cluster phosphorylation by mutational analysis. 1 integrins with a double threonine to alanine substitution in the cytoplasmic tail were found to be defective in mediating cell attachment and did not contribute to fibronectin fibril forma-tion [36]. The conformation of the 1 extracellular domain was shifted to-wards an inactive state as measured by binding of the 9EG7 antibody to a conformation-sensitive epitope. 1-antibody induced clustering of 1-T788/789A demonstrated that the mutant cytoplasmic part was functional in mediating activation of FAK [36]. In the thesis work, we have addressed the question whether the drastic effect on extracellular conformation and cell attachment could be attributed to one of the threonine substitutions or both, and investigated the functional role of the suggested phosphorylation of the conserved threonines. We established in Paper I that T788, but not T789, is a site of critical importance for inside-out activation of integrins. In contrast to an alanine substitution, introduction of a phospho-mimicking mutation, 1-T788D, resulted in integrins expressing the 9EG7 epitope and fully active in promoting cell adhesion. In line with these data, the T788D mutation did not interfere with binding of the talin head domain. Thus, we concluded that phosphorylation of T788 in integrin 1 is compatible with inside-out recep-tor activation. Furthermore, cells expressing 1-T788D contained exagger-ated numbers of focal contacts and migrated slowly. It is possible that the phospho-mimetic mutation stimulated formation of adhesion sites by pro-moting integrin activation. Alternatively, the mutation reduced the turnover of focal contacts.

Is phosphorylation alone sufficient for inducing conformational rear-rangements in the extracellular domain or is it mediated indirectly through regulation of binding to cytoplasmic proteins? A recent publication by

46

Kiema and co-workers [69] extensively describe the binding of the actin-crosslinking protein filamin A to the 7-cytoplasmic domain peptide, and the solved structure of the complex is thought to be representative for multiple integrin-filamin interactions. The binding site in the tails partially overlaps with that of talin, and the binding domains of filamin and talin can compete for binding to integrin tails. Phospho-mimicking mutations of the 7-threonine cluster strongly reduced filamin binding, but, consistent with our data, did not interfere with talin binding. Therefore, phosphorylation of the

7 tail may regulate integrin-filamin interactions, and thereby indirectly the talin-dependent integrin activation. In analogy, the muscle specific 1D iso-form, which binds poorly to filamin but is strongly associated with actin filaments [130], holds two asparagine residues in the position of the 1Athreonine residues. The filamin- 1-T788D interaction would be of great interest to investigate. Taken together, even though threonine phosphoryla-tion of 1 still has not been clearly shown, several lines of data suggest that phosphorylation contributes to the regulation of inside-out activation of the receptor in a similar way as shown for 2 and 3 integrins. However, further investigations are required to establish whether threonine phosphorylation is part of a common model of integrin activation.

1 mediated signalling – relying on activation of the FAK/Src complex? FAK is activated after ligand binding to most integrins, and is thought to be, together with Src, one of the central players in integrin-mediated events. A number of important signalling molecules, including Src family kinases, phosphatidylinositol 3-kinase (PI3K), protein kinase B (PKB/Akt), mitogen-activated protein (MAP) kinases, and Rho GTPases, can be activated in a FAK-dependent manner after integrin ligation [237]. However, the forma-tion of focal complexes and certain early signalling events are FAK-independent [144, 190], whereas FAK facilitates focal adhesion turnover and promotes migration [134, 237].

From previous studies using 1B-expressing cells it is known that this al-ternatively spliced isoform combines with available subunits and is dis-played on the cell surface. However, the 3 1B, 5 1B, and 6 1B recep-tors, which form when expressed in GD25 cells, are presented on the cell surface in an inactive conformation with low affinity for ligand. Adhesion to immobilised 1-antibodies through 1A, but not 1B, receptors trigger phosphorylation of FAK, indicating that the 1B cytoplasmic tail lacks the adequate sequence information necessary for FAK activation [238, 239]. The bacterial protein invasin is a very potent ligand of 1 integrins, including

1B receptors, but does not bind V 3 integrins [240]. Thus, 1A- and 1B-expressing GD25 cells, together with invasin, are well suited for de-

47

tailed analysis of 1-integrin mediated signalling pathways to determine what regulatory steps really are dependent on active FAK.

Figure 8. Outline of how integrins and growth factor receptors crosstalk in the pathways to PI3K-Akt and Erk.

Paper II – FAK and PI3K activation as two parallel key events Adhesion-induced phosphorylation of tyrosine 397 in FAK serves as a bind-ing site for the Src-SH2 domain as well as the p85-subunit of class Ia PI3K and the interaction results in activation of these kinases [133, 137]. Phos-phorylation of FAK, and downstream targets like paxillin, is abolished in

1B-cells. Interestingly, integrin-induced phosphorylation of Cas, another downstream target of the FAK/Src-complex, is only partially reduced in these cells [107, 241]. PI3K inhibitors abolish the remaining Cas-phosphorylation, suggesting a PI3K-dependent pathway activated by integrin engagement upstream of FAK activation. In addition, mutation of K756 in the TM-cytoplasmic interface of the 1 integrin causes reduced Cas-phosphorylation and PI3K activity, although the FAK activity is unaffected [190].

The aim of the study in Paper II was to address these previous indications of a non-common pathway from 1 integrins to FAK and PI3K activation. The well-established PI3K-dependent phosphorylation of Akt-S473 was used as a read-out of PI3K activity. Unexpectedly, also ligation of 1B-integrins induced strong phosphorylation of Akt-S473 and this was not due to differences in the activity of the serine-threonine phosphatase PP2A. The FAK independence of PI3K activation was confirmed in FAK-deficient cells. The possible contribution of Src-family kinases was addressed by the use of two Src-family inhibitors, PP2 and SU6656, and in SYF (Src, Yes, and Fyn)-deficient cells. Surprisingly, the phosphorylation of Akt-S473 oc-curred in the presence of SU6656 and in SYF-deficient cells whereas it was prevented by PP2. Together with previous studies, the results of Paper II demonstrated the presence of a FAK-independent pathway, which also ex-

48

cluded the involvement of Src-family kinases, from ligand-engaged 1integrins to the activation of PKB/Akt via PI3K. The membrane-proximal half of 1 was necessary and sufficient to promote this pathway. The path-way involves a PP2-sensitive step, a yet unidentified tyrosine kinase, sugges-tively acting upstream of PI3K.

Recently, we found FAK-Y397 to be phosphorylated in SYF-deficient cells after adhesion (unpublished results). The finding is in contrast to the previous suggestion that FAK was not tyrosine phosphorylated in SYF-deficient cells in response to adhesion [242], and strengthens the idea that FAK-Y397 is an autophosphorylation site and phosphorylation here is not dependent on Src-activity [134]. This finding does not contradict our previ-ous conclusion of a FAK- and Src-independent activation pathway to PI3K, since autophosphorylation of FAK-Y397 in SYF-deficient cells was not sensitive to PP2, in contrast to PI3K activation.

Thus, PI3K can apparently be activated by integrins both downstream of FAK [137, 243-245] and independently of FAK. The possibility that FAK activation would be dependent on PI3K activity is unlikely because phos-phorylation of FAK-Y397 was previously found to be wortmannin-insensitive [190]. This suggests that parallel pathways serving as two key events in integrin-mediated signalling can activate PI3K and FAK.

Paper III – Adhesion via 1 integrins controls activation of Erk at several checkpoints, with and without FAK involvement The classic MAPK cascade begins with the activation and auto-phosphorylation of a growth factor-receptor tyrosine kinase (GF-RTK) on the cell surface [246]. Briefly, the following cascade of events takes place. The phosphorylation of cytoplasmic RTK-tyrosine residues creates binding sites for SH2-domain containing adaptor proteins such as Shc and Grb2, which in turn recruit Sos, an exchange factor for Ras GTPase. Sos activates Ras at the membrane, which there can associate with the serine-threonine kinase cRaf. Here starts the MAP kinase part of the cascade. In order to lo-calise the enzyme/substrate molecules next to each other cRaf interacts with large scaffolding protein complexes e.g. KSR (kinase suppressor of Ras) or paxillin (reviewed in [195]). These scaffolds are constitutively associated with Mek1 or Mek2 that becomes phosphorylated and activated by cRaf. Mek1/2 then in turn phosphorylate Erk1 and Erk2 on a critical TEY-motif, leading to the activation of Erk. Activated Erk is released from Mek and acts on a diverse set of substrates in both the cytoplasm and in the nucleus. Im-portantly, the GF-induced stimulation of Erk is largely linked to, and de-pendent on, integrin mediated adhesion [234] for maximal and sustained effect. The coordinated signalling is required for the cascade reaction to pass checkpoints downstream of Ras; the recent reviews [213, 247, 248] describe the multiple levels of regulation by integrins.

49

To add complexity to the cooperative signalling cascades, the integrin signals are divergent already at early steps, with both FAK-dependent and -independent mechanisms. The aim of the study in Paper III was to character-ize the mechanisms by which specifically 1-integrins contribute to Erk-activation and, in particular, what regulatory steps are dependent on activa-tion of FAK. The study was carried out in the absence of growth factors. First, we determined that the two commonly co-expressed fibronectin recep-tors 5 1A and V 3 could induce the Erk-pathway independently. The

1A-integrin mechanism was then compared with the signals induced by 1B-receptors to pinpoint FAK-dependent phosphorylation sites on the

MAP kinases and their effector proteins. The results indicated that 1-integrin ligation promote Erk activation by FAK-dependent mechanisms at the levels of both cRaf and Mek, and in addition, by a FAK-independent pathway requiring Mek. The latter pathway may be the quantitatively most significant for Erk activation under these defined conditions. All adhesion-induced phosphorylation of Erk was abolished by the use of the MEK-specific inhibitor PD98059, thus both the FAK-dependent and -independent pathways require Mek to phosphorylate Erk.

By using small interfering RNA (siRNA) directed against all three ShcA splice variants, we showed that the mechanism for FAK-independent 1-induced Mek and Erk activation does not involve ShcA. ShcA was previ-ously reported to connect certain integrins (including 5 1) to the activation of Ras independently of FAK [196, 197]. Nevertheless, ShcA has not a ma-jor role in the FAK-independent activation of Erk in 1B-cells.

Phosphorylation of Mek in the active loop (S217/S221) was mediated to similar extent in both 1A- and 1B-cells, but phosphorylation of two up-stream checkpoints, at Mek-S298, a site important for close localisation of Mek to cRaf, and cRaf-S338 (required for cRaf activation) were only weakly induced in 1B-cells. These results reflect very poor conditions for efficient phosphorylation of Mek-S217/S221 by cRaf and opens up for the suggestion of an alternative kinase in the activation of Mek. This could be addressed by use of siRNA targeting of candidate kinases, e.g. B-Raf. Interestingly, 1-integrin ligation strongly induced phosphorylation of the PKC kinase do-main, primarily on the conserved site corresponding to PKC -T514. Phos-phorylation at this site is thought to be mediated by PDK1, intriguingly, a kinase found constitutively active in the cytoplasm [249]. Another link be-tween PKC and Mek activation is the Raf kinase inhibitory protein (RKIP) [250], which after phosphorylation by PKCs (in response to phorbol ester or EGF) dissociates from cRaf [251]. Thus, there is a correlation between the

1B-induced phosphorylation of PKC and the Mek and Erk activation, and direct linking mechanisms are suggested in the literature but not yet tested in response to adhesion.

50

Future Perspectives

The investigations included in this thesis have generated new issues to ad-dress – as always in science…

Concerning the 1 cytoplasmic threonine cluster and its role in integrin acti-vation I would like to point out the following questions:

Does phosphorylation of the threonine residues take place? Generation of phospho-site specific antibodies for each phosphorylated threonine resi-due may allow clarification of this issue. The set of 1-mutants we hold (T788A, T789A, T788/789A) could serve as negative controls for binding specificity. Is the binding of intact talin (rather than the isolated head domain) or only talin rod domain possibly influenced by T788A or T788D substitutions? Does binding of the 1 cytoplasmic tail to other FERM-domain contain-ing proteins than talin influence integrin activation? If so, do these pro-teins promote activation or inhibition? Are such interactions possibly regulated by threonine substitutions/phosphorylation?

The newly identified 1-integrin mediated pathway to PI3K suggests to: search for the PP2-sensitive tyrosine kinase upstream of PI3K; there are

90 tyrosine kinases in the human genome, almost all with orthologs in mouse, thus the number of PP2-sensitive kinases is reasonable to screen. compare the known regulatory steps in the pathway from integrin to Akt, such as the activity of the phospholipid phosphatase PTEN in 1A- and

1B-cells. map in more detail what part of the membrane-proximal region of 1 that is necessary and sufficient for PI3K activation.

The FAK-dependent and -independent mechanisms of integrin-mediated Erk activation require more investigation regarding:

the contribution of B-Raf to Mek activation in GD25 cells. if not involving B-Raf, how is the FAK-independent activation of Mek taking place? the role of the distinct adhesion-dependent PKC phosphorylation in the regulation of Erk activation.

51

Concluding remarks Although I have not addressed the involvement of integrins in human dis-ease, my hope is that insights obtained from this and further detailed studies will, in the long run, lead to better therapeutic agents targeting integrin sig-nalling. If it could contribute to treatment of any of the diverse diseases in-volving integrins, from thrombosis, inflammation, atherosclerosis, osteopo-rosis, cancer, to infectious disease, it would be a great reward. It is after all the biological importance of these receptors that make them particularly interesting; the elegance of their allosteric signal transduction mechanisms is merely an extra bonus and thrill during everyday struggle in the lab for bio-chemists like me.

52

Acknowledgement

This doctoral thesis project has been carried out at the department of Medical Biochemistry and Microbiology at Uppsala Universitet, Sweden. I would like to thank the department for providing good research facilities and reliable working conditions for PhD students.

Ett stort tack till dig Staffan, min handledare! Din breda och djupa kunskap är imponerande och guld värd, men minst lika viktigt är att du alltid tar dig tid, alltid hittar positiva saker i ens resultat (oavsett hur det gått), alltid bryr dig om hur man mår och är lika bra på att alltid vara lika sen som jag är! ;-)

Tack Teet för att du ville vara min biträdande handledare, för gott samarbete på labbet, för dina underfundiga skämt och hjälpsamma kartor över Tallinn med omnejd…

Mitt varmaste tack går till:Birgitta – vad mer finns att säga än: vad skulle jag göra utan dig? Tack för all hjälp genom åren, inte bara nu på slutet…

Alla som varit och är i gruppen: Anne, Annika, Dorota, Gunbjörg, Fredrik, Lars, Nathalie, Paola, Sophie och Åsa. Tack till er studenter som påmint mig om vad framförhållning är och som gjort ett stort jobb i mina projekt: Rosa-lien, Priya, Hanna, David, Sofia och Karin.

Lena – för att du är en sådan varm person! För alla stunder kring ritningar, korridorer som villor…

Katarina – för att jag fick flytta in hos dig och dela stressen & mathyllan med någon som var i samma sits som jag, både i skrivandet & hemma. Snart är det vår och vi är fria!

Alla ni andra som finns, kommit och gått i korridoren genom åren. Speciellt tack till Inger – du är en klippa! Blå soffan rules!

Klipporna som får kugghjulen att rulla: Kerstin, Olav, Piotr, Tony, nya och gamla sekreterare!

53

Alla trevliga kollegor på IMBIM, speciellt Ahn-Tri & Maarten för att ni alltid lättar upp stämningen!

Åsa L – Skönt att ha vänner som du att vända sig till! Tack för sällskap i IPhAB och alla andra diskussioner om livet som doktorand och inte…

Jenny P & Frida H – Minns ni hur det började i det grå & slitna D3:4? Vil-ken tur att ni två funnits vid min sida då, på Stallet, Grindis, och delat alla åren här på BMC med mig!

Jenny HM – tänk vad långa dagar på Kemikum kan leda fram till! Vad glad jag är att du, den där gången på V-Dala puben, tyckte att jag skulle börja hos ”den jättesnälla professorn” i din korridor! Tack för att du är så positiv och glad! Längtar tills ni kommer hem!

Ridgruppen – som får mig att glömma allt annat en timme i veckan och komma tillbaka med ny energi.

Ulrika, Johan & Oliver – Tack för alla avspända familjestunder! Tänk Ulri-ka, vad mycket den där, nu nästan bortglömda, golfkursen våren 2002 kunde ge! Tack vännen för att du finns!

Min, uti landet vidsträckta, familj! Tack för allt stöd och all hjälp med Adam! Speciellt tack till dig pappa, för att du orkat ta dig igenom och rätta denna smått obegripliga, men mycket interesting, skrift!

Micael – för allt slit med husbyggnadsplanerna som jag inte har kunnat vara delaktig i så mycket som jag egentligen ville! Du är och har varit min trygg-het under åren!

Adam – min lilla gullegubbe – Glimten i dina ögon gör mig alltid varm inombords!

54

References

1. Hynes, R.O., Integrins: versatility, modulation, and signaling in cell adhesion. Cell, 1992. 69(1): p. 11-25.

2. Horwitz, A.F., Integrins and health. Sci Am, 1997. 276(5): p. 68-75. 3. Juliano, R.L., Signal transduction by cell adhesion receptors and the

cytoskeleton: functions of integrins, cadherins, selectins, and immu-noglobulin-superfamily members. Annu Rev Pharmacol Toxicol, 2002. 42: p. 283-323.

4. Tamkun, J.W., et al., Structure of integrin, a glycoprotein involved in the transmembrane linkage between fibronectin and actin. Cell, 1986. 46(2): p. 271-82.

5. Hynes, R.O., Integrins: bidirectional, allosteric signaling machines.Cell, 2002. 110(6): p. 673-87.

6. de Melker, A.A. and A. Sonnenberg, Integrins: alternative splicing as a mechanism to regulate ligand binding and integrin signaling events. Bioessays, 1999. 21(6): p. 499-509.

7. Fornaro, M. and L.R. Languino, Alternatively spliced variants: a new view of the integrin cytoplasmic domain. Matrix Biol, 1997. 16(4): p. 185-93.

8. Delwel, G.O., I. Kuikman, and A. Sonnenberg, An alternatively spliced exon in the extracellular domain of the human alpha 6 in-tegrin subunit--functional analysis of the alpha 6 integrin variants.Cell Adhes Commun, 1995. 3(2): p. 143-61.

9. Ziober, B.L., et al., Alternative extracellular and cytoplasmic do-mains of the integrin alpha 7 subunit are differentially expressed during development. J Biol Chem, 1993. 268(35): p. 26773-83.

10. Schober, S., et al., The role of extracellular and cytoplasmic splice domains of alpha7-integrin in cell adhesion and migration on laminins. Exp Cell Res, 2000. 255(2): p. 303-13.

11. Belkin, A.M., et al., Muscle beta1D integrin reinforces the cy-toskeleton-matrix link: modulation of integrin adhesive function by alternative splicing. J Cell Biol, 1997. 139(6): p. 1583-95.

12. Armulik, A., Splice variants of human beta 1 integrins: origin, bio-synthesis and functions. Front Biosci, 2002. 7: p. d219-27.

13. Kumar, C.S., et al., Cloning and characterization of a novel integrin beta3 subunit. J Biol Chem, 1997. 272(26): p. 16390-7.

14. van Kuppevelt, T.H., et al., An alternative cytoplasmic domain of the integrin beta 3 subunit. Proc Natl Acad Sci U S A, 1989. 86(14): p. 5415-8.

55

15. Iacovacci, S., et al., Genomic organization of the human integrin beta4 gene. Mamm Genome, 1997. 8(6): p. 448-50.

16. van Leusden, M.R., I. Kuikman, and A. Sonnenberg, The unique cytoplasmic domain of the human integrin variant beta4E is pro-duced by partial retention of intronic sequences. Biochem Biophys Res Commun, 1997. 235(3): p. 826-30.

17. Zhang, H., S.M. Tan, and J. Lu, cDNA cloning reveals two mouse beta5 integrin transcripts distinct in cytoplasmic domains as a result of alternative splicing. Biochem J, 1998. 331 ( Pt 2): p. 631-7.

18. Heino, J., et al., Regulation of cell adhesion receptors by transform-ing growth factor-beta. Concomitant regulation of integrins that share a common beta 1 subunit. J Biol Chem, 1989. 264(1): p. 380-8.

19. Nermut, M.V., et al., Electron microscopy and structural model of human fibronectin receptor. Embo J, 1988. 7(13): p. 4093-9.

20. Humphries, M.J., et al., Integrin structure: heady advances in ligand binding, but activation still makes the knees wobble. Trends Bio-chem Sci, 2003. 28(6): p. 313-20.

21. Humphries, M.J., E.J. Symonds, and A.P. Mould, Mapping func-tional residues onto integrin crystal structures. Curr Opin Struct Biol, 2003. 13(2): p. 236-243.

22. Gu, J. and N. Taniguchi, Regulation of integrin functions by N-glycans. Glycoconj J, 2004. 21(1-2): p. 9-15.

23. Bellis, S.L., Variant glycosylation: an underappreciated regulatory mechanism for beta1 integrins. Biochim Biophys Acta, 2004. 1663(1-2): p. 52-60.

24. Seales, E.C., et al., A PKC/ras/ERK signaling pathway activates myeloid fibronectin receptors by altering beta 1 integrin sialylation.J Biol Chem, 2005.

25. Springer, T.A., Folding of the N-terminal, ligand-binding region of integrin alpha-subunits into a beta-propeller domain. Proc Natl Acad Sci U S A, 1997. 94(1): p. 65-72.

26. Humphries, M.J., Integrin Structure. Biochem Soc Trans, 2000. 28(4): p. 311-339.

27. Gailit, J. and E. Ruoslahti, Regulation of the fibronectin receptor affinity by divalent cations. J Biol Chem, 1988. 263(26): p. 12927-32.

28. Stefansson, A., et al., Determination of N- and C-terminal borders of the transmembrane domain of integrin subunits. J Biol Chem, 2004. 279(20): p. 21200-5.

29. Adair, B.D. and M. Yeager, Three-dimensional model of the human platelet integrin alpha IIbbeta 3 based on electron cryomicroscopy and x-ray crystallography. Proc Natl Acad Sci U S A, 2002. 99(22):p. 14059-64.

30. Sastry, S.K. and A.F. Horwitz, Integrin cytoplasmic domains: me-diators of cytoskeletal linkages and extra- and intracellular initiated

56

transmembrane signaling. Curr Opin Cell Biol, 1993. 5(5): p. 819-31.

31. Travis, M.A., J.D. Humphries, and M.J. Humphries, An unraveling tale of how integrins are activated from within. Trends Pharmacol Sci, 2003. 24(4): p. 192-7.

32. Ylanne, J., Conserved functions of the cytoplasmic domains of in-tegrin beta subunits. Front Biosci, 1998. 3: p. d877-86.

33. Reszka, A.A., Y. Hayashi, and A.F. Horwitz, Identification of amino acid sequences in the integrin beta 1 cytoplasmic domain implicated in cytoskeletal association. J Cell Biol, 1992. 117(6): p. 1321-30.

34. Sampath, R., P.J. Gallagher, and F.M. Pavalko, Cytoskeletal interac-tions with the leukocyte integrin beta2 cytoplasmic tail. Activation-dependent regulation of associations with talin and alpha-actinin. J Biol Chem, 1998. 273(50): p. 33588-94.

35. Hibbs, M.L., et al., The cytoplasmic domain of the integrin lympho-cyte function-associated antigen 1 b subunit: sites required for bind-ing to intercellular adhesion molecule 1 and the phorbol ester-stimulated phosphorylation site. J. Exp. Med., 1991. 174(5): p. 1227-1238.

36. Wennerberg, K., et al., Mutational analysis of the potential phos-phorylation sites in the cytoplasmic domain of integrin beta1A. Re-quirement for threonines 788- 789 in receptor activation. J Cell Sci, 1998. 111(Pt 8): p. 1117-26.

37. van Willigen, G., et al., Exposure of ligand-binding sites on platelet integrin alpha IIB/beta 3 by phosphorylation of the beta 3 subunit.Biochem J, 1996. 314(Pt 3): p. 769-79.

38. Fagerholm, S.C., T.J. Hilden, and C.G. Gahmberg, P marks the spot: site-specific integrin phosphorylation regulates molecular interac-tions. Trends Biochem Sci, 2004. 29(9): p. 504-12.

39. Xiong, J.P., et al., Crystal structure of the extracellular segment of integrin alpha Vbeta3. Science, 2001. 294(5541): p. 339-45.

40. Takagi, J., et al., Global conformational rearrangements in integrin extracellular domains in outside-in and inside-out signaling. Cell, 2002. 110(5): p. 599.

41. Beglova, N., et al., Cysteine-rich module structure reveals a fulcrum for integrin rearrangement upon activation. Nat Struct Biol, 2002. 9(4): p. 282-7.

42. O'Neill, S., et al., The platelet integrin alpha IIbbeta 3 has an en-dogenous thiol isomerase activity. J Biol Chem, 2000. 275(47): p. 36984-90.

43. Walsh, G.M., et al., Redox modulation of integrin [correction of integin] alpha IIb beta 3 involves a novel allosteric regulation of its thiol isomerase activity. Biochemistry, 2004. 43(2): p. 473-80.

44. Li, R., et al., Activation of integrin alphaIIbbeta3 by modulation of transmembrane helix associations. Science, 2003. 300(5620): p. 795-8.

57

45. Li, R., et al., Dimerization of the transmembrane domain of Integrin alphaIIb subunit in cell membranes. J Biol Chem, 2004. 279(25): p. 26666-73.

46. Li, R., J.S. Bennett, and W.F. Degrado, Structural basis for integrin alphaIIbbeta3 clustering. Biochem Soc Trans, 2004. 32(Pt3): p. 412-5.

47. Li, R., et al., Characterization of the monomeric form of the trans-membrane and cytoplasmic domains of the integrin beta 3 subunit by NMR spectroscopy. Biochemistry, 2002. 41(52): p. 15618-24.

48. Vinogradova, O., et al., A Structural Mechanism of Integrin al-pha(IIb)beta(3) "Inside-Out" Activation as Regulated by Its Cyto-plasmic Face. Cell, 2002. 110(5): p. 587.

49. Armulik, A., et al., Determination of the border between the trans-membrane and cytoplasmic domains of human integrin subunits. J Biol Chem, 1999. 274(52): p. 37030-4.

50. Liddington, R.C. and M.H. Ginsberg, Integrin activation takes shape. J Cell Biol, 2002. 158(5): p. 833-9.

51. Calderwood, D.A., Integrin activation. J Cell Sci, 2004. 117(Pt 5): p. 657-66.

52. Hughes, P.E., et al., The conserved membrane-proximal region of an integrin cytoplasmic domain specifies ligand binding affinity. J Biol Chem, 1995. 270(21): p. 12411-7.

53. O'Toole, T.E., et al., Modulation of the affinity of integrin alpha IIb beta 3 (GPIIb-IIIa) by the cytoplasmic domain of alpha IIb. Science, 1991. 254(5033): p. 845-7.

54. O'Toole, T.E., et al., Integrin cytoplasmic domains mediate inside-out signal transduction. J Cell Biol, 1994. 124(6): p. 1047-59.

55. Hughes, P.E., et al., Breaking the integrin hinge. A defined struc-tural constraint regulates integrin signaling. J Biol Chem, 1996. 271(12): p. 6571-4.

56. Vinogradova, O., et al., Membrane-mediated structural transitions at the cytoplasmic face during integrin activation. Proc Natl Acad Sci U S A, 2004. 101(12): p. 4094-9.

57. Takagi, J., H.P. Erickson, and T.A. Springer, C-terminal opening mimics 'inside-out' activation of integrin alpha5beta1. Nat Struct Biol, 2001. 8(5): p. 412-6.

58. Lu, C., J. Takagi, and T.A. Springer, Association of the membrane proximal regions of the alpha and beta subunit cytoplasmic domains constrains an integrin in the inactive state. J Biol Chem, 2001. 276(18): p. 14642-8.

59. Kim, M., C.V. Carman, and T.A. Springer, Bidirectional transmem-brane signaling by cytoplasmic domain separation in integrins. Sci-ence, 2003. 301(5640): p. 1720-5.

60. Tadokoro, S., et al., Talin binding to integrin beta tails: a final common step in integrin activation. Science, 2003. 302(5642): p. 103-6.

58

61. Calderwood, D.A., et al., The Talin head domain binds to integrin beta subunit cytoplasmic tails and regulates integrin activation. J Biol Chem, 1999. 274(40): p. 28071-4.

62. Calderwood, D.A., et al., The phosphotyrosine binding-like domain of talin activates integrins. J Biol Chem, 2002. 277(24): p. 21749-58.

63. Xing, B., A. Jedsadayanmata, and S.C. Lam, Localization of an in-tegrin binding site to the C terminus of talin. J Biol Chem, 2001. 276(48): p. 44373-8.

64. Kaapa, A., K. Peter, and J. Ylanne, Effects of mutations in the cyto-plasmic domain of integrin beta(1) to talin binding and cell spread-ing. Exp Cell Res, 1999. 250(2): p. 524-34.

65. Patil, S., et al., Identification of a talin-binding site in the integrin beta(3) subunit distinct from the NPLY regulatory motif of post-ligand binding functions. The talin n-terminal head domain interacts with the membrane-proximal region of the beta(3) cytoplasmic tail. J Biol Chem, 1999. 274(40): p. 28575-83.

66. Garcia-Alvarez, B., et al., Structural determinants of integrin recog-nition by talin. Mol Cell, 2003. 11(1): p. 49-58.

67. Ulmer, T.S., et al., Domain-specific interactions of talin with the membrane-proximal region of the integrin beta3 subunit. Biochem-istry, 2003. 42(27): p. 8307-12.

68. Shattil, S.J., et al., Beta 3-endonexin, a novel polypeptide that inter-acts specifically with the cytoplasmic tail of the integrin beta 3 sub-unit. J Cell Biol, 1995. 131(3): p. 807-16.

69. Kiema, T., et al., The molecular basis of filamin binding to integrins and competition with talin. Mol Cell, 2006. 21(3): p. 337-47.

70. Martel, V., et al., Conformation, localization, and integrin binding of talin depend on its interaction with phosphoinositides. J Biol Chem, 2001. 276(24): p. 21217-27.

71. Di Paolo, G., et al., Recruitment and regulation of phosphatidy-linositol phosphate kinase type 1 gamma by the FERM domain of ta-lin. Nature, 2002. 420(6911): p. 85-9.

72. Ling, K., et al., Type I gamma phosphatidylinositol phosphate kinase targets and regulates focal adhesions. Nature, 2002. 420(6911): p. 89-93.

73. Calderwood, D.A., et al., Competition for talin results in trans-dominant inhibition of integrin activation. J Biol Chem, 2004. 279(28): p. 28889-95.

74. Barsukov, I.L., et al., Phosphatidylinositol phosphate kinase type 1gamma and beta1-integrin cytoplasmic domain bind to the same region in the talin FERM domain. J Biol Chem, 2003. 278(33): p. 31202-9.

75. Ling, K., et al., Tyrosine phosphorylation of type I{gamma} phos-phatidylinositol phosphate kinase by Src regulates an integrin-talin switch. J Cell Biol, 2003. 163(6): p. 1339-1349.

59

76. Yan, B., et al., Calpain cleavage promotes talin binding to the beta 3 integrin cytoplasmic domain. J Biol Chem, 2001. 276(30): p. 28164-70.

77. Sato, K. and S. Kawashima, Calpain function in the modulation of signal transduction molecules. Biol Chem, 2001. 382(5): p. 743-51.

78. Chang, D.D., et al., Molecular basis for interaction between Icap1aalpha PTB domain and bbeta1 integrin. J Biol Chem, 2001. 7: p. 7.

79. Chang, D.D., et al., ICAP-1, a novel beta1 integrin cytoplasmic do-main-associated protein, binds to a conserved and functionally im-portant NPXY sequence motif of beta1 integrin. J Cell Biol, 1997. 138(5): p. 1149-57.

80. Bouvard, D., et al., Disruption of focal adhesions by integrin cyto-plasmic domain-associated protein-1 alpha. J Biol Chem, 2003. 278(8): p. 6567-74.

81. Bouvard, D. and M.R. Block, Calcium/calmodulin-dependent pro-tein kinase II controls integrin alpha5beta1-mediated cell adhesion through the integrin cytoplasmic domain associated protein-1alpha.Biochem Biophys Res Commun, 1998. 252(1): p. 46-50.

82. Bouvard, D., A. Molla, and M.R. Block, Calcium/calmodulin-dependent protein kinase II controls alpha5beta1 integrin-mediated inside-out signaling. J Cell Sci, 1998. 111 ( Pt 5): p. 657-65.

83. Blystone, S.D., et al., Inducible tyrosine phosphorylation of the beta3 integrin requires the alphaV integrin cytoplasmic tail. J Biol Chem, 1996. 271(49): p. 31458-62.

84. Buyon, J.P., et al., Constitutive and induced phosphorylation of the alpha- and beta-chains of the CD11/CD18 leukocyte integrin family. Relationship to adhesion-dependent functions. J Immunol, 1990. 144(1): p. 191-7.

85. Parise, L.V., et al., Glycoprotein IIIa is phosphorylated in intact human platelets. Blood, 1990. 75(12): p. 2363-2368.

86. Sakai, T., et al., Role of the cytoplasmic tyrosines of beta 1A in-tegrins in transformation by v-src. Proc Natl Acad Sci U S A, 2001. 98(7): p. 3808-13.

87. Calderwood, D.A., et al., Integrin beta cytoplasmic domain interac-tions with phosphotyrosine-binding domains: a structural prototype for diversity in integrin signaling. Proc Natl Acad Sci U S A, 2003. 100(5): p. 2272-7.

88. Xi, X., et al., Critical roles for the COOH-terminal NITY and RGT sequences of the integrin beta3 cytoplasmic domain in inside-out and outside-in signaling. J Cell Biol, 2003. 162(2): p. 329-39.

89. Hilden, T.J., et al., Threonine Phosphorylation Sites in the beta(2) and beta(7) Leukocyte Integrin Polypeptides. J Immunol, 2003. 170(8): p. 4170-7.

90. Perez, O.D., et al., Leukocyte functional antigen 1 lowers T cell acti-vation thresholds and signaling through cytohesin-1 and Jun-activating binding protein 1. Nat Immunol, 2003. 4(11): p. 1083-92.

60

91. Fagerholm, S., et al., Phosphorylation of the cytoplasmic domain of the integrin CD18 chain by protein kinase C isoforms in leukocytes.J Biol Chem, 2002. 277(3): p. 1728-38.

92. Valmu, L., et al., The cytoskeletal association of CD11/CD18 leuko-cyte integrins in phorbol ester-activated cells correlates with CD18 phosphorylation. Eur J Immunol, 1999. 29(7): p. 2107-18.

93. Valmu, L. and C.G. Gahmberg, Treatment with okadaic acid reveals strong threonine phosphorylation of CD18 after activation of CD11/CD18 leukocyte integrins with phorbol esters or CD3 anti-bodies. J Immunol, 1995. 155(3): p. 1175-83.

94. Mastrangelo, A.M., et al., Amino acid motifs required for isolated beta cytoplasmic domains to regulate 'in trans' beta1 integrin con-formation and function in cell attachment. J Cell Sci, 1999. 112 ( Pt 2): p. 217-29.

95. Bodeau, A.L., et al., A functional comparison of mutations in in-tegrin beta cytoplasmic domains: effects on the regulation of tyro-sine phosphorylation, cell spreading, cell attachment and beta1 in-tegrin conformation. J Cell Sci, 2001. 114(Pt 15): p. 2795-807.

96. Kim, S.M., et al., Modulation of Thr phosphorylation of integrin beta1 during muscle differentiation. J Biol Chem, 2004. 279(8): p. 7082-90.

97. Suzuki, K. and K. Takahashi, Reduced cell adhesion during mitosis by threonine phosphorylation of beta1 integrin. J Cell Physiol, 2003. 197(2): p. 297-305.

98. Mulrooney, J., et al., Phosphorylation of the beta1 integrin cyto-plasmic domain: toward an understanding of function and mecha-nism. Exp Cell Res, 2000. 258(2): p. 332-41.

99. Barreuther, M.F. and L.B. Grabel, The role of phosphorylation in modulating beta 1 integrin localization. Exp Cell Res, 1996. 222(1):p. 10-5.

100. Miyamoto, S., S.K. Akiyama, and K.M. Yamada, Synergistic roles for receptor occupancy and aggregation in integrin transmembrane function. Science, 1995. 267(5199): p. 883-5.

101. Law, D.A., et al., Integrin cytoplasmic tyrosine motif is required for outside-in alphaIIbbeta3 signalling and platelet function. Nature, 1999. 401(6755): p. 808-11.

102. Cowan, K.J., D.A. Law, and D.R. Phillips, Identification of SHC as the primary protein binding to the tyrosine phosphorylated beta{sub3} subunit of alphalIbbeta{sub3} during outside- in integrin platelet signaling. J Biol Chem, 2000.

103. Jenkins, A.L., et al., Tyrosine phosphorylation of the beta3 cyto-plasmic domain mediates integrin-cytoskeletal interactions. J Biol Chem, 1998. 273(22): p. 13878-85.

104. Chandhoke, S.K., et al., Beta 3 integrin phosphorylation is essential for Arp3 organization into leukocyte alpha V beta 3-vitronectin ad-hesion contacts. J Cell Sci, 2004. 117(Pt 8): p. 1431-41.

61

105. Butler, B., et al., Purified integrin adhesion complexes exhibit actin-polymerization activity. Curr Biol, 2006. 16(3): p. 242-51.

106. Sakai, T., et al., Modulation of beta1A integrin functions by tyrosine residues in the beta1 cytoplasmic domain. J Cell Biol, 1998. 141(2):p. 527-38.

107. Wennerberg, K., et al., The cytoplasmic tyrosines of integrin subunit beta1 are involved in focal adhesion kinase activation. Mol Cell Biol, 2000. 20(15): p. 5758-65.

108. Lerea, K.M., et al., Phosphorylation sites in the integrin beta3 cyto-plasmic domain in intact platelets. J Biol Chem, 1999. 274(4): p. 1914-9.

109. Kirk, R.I., M.R. Sanderson, and K.M. Lerea, Threonine phosphory-lation of the beta 3 integrin cytoplasmic tail, at a site recognized by PDK1 and Akt/PKB in vitro, regulates Shc binding. J Biol Chem, 2000. 275(40): p. 30901-6.

110. Cukierman, E., et al., Taking cell-matrix adhesions to the third di-mension. Science, 2001. 294(5547): p. 1708-12.

111. Geiger, B., et al., Transmembrane crosstalk between the extracellu-lar matrix--cytoskeleton crosstalk. Nat Rev Mol Cell Biol, 2001. 2(11): p. 793-805.

112. Burridge, K. and M. Chrzanowska-Wodnicka, Focal adhesions, contractility, and signaling. Annu Rev Cell Dev Biol, 1996. 12: p. 463-518.

113. Winder, S.J. and K.R. Ayscough, Actin-binding proteins. J Cell Sci, 2005. 118(Pt 4): p. 651-4.

114. Lo, S.H., Tensin. Int J Biochem Cell Biol, 2004. 36(1): p. 31-4. 115. Horwitz, A., et al., Interaction of plasma membrane fibronectin re-

ceptor with talin--a transmembrane linkage. Nature, 1986. 320(6062): p. 531-3.

116. Pfaff, M., et al., Integrin beta cytoplasmic domains differentially bind to cytoskeletal proteins. J Biol Chem, 1998. 273(11): p. 6104-9.

117. Critchley, D.R., Focal adhesions - the cytoskeletal connection. Curr Opin Cell Biol, 2000. 12(1): p. 133-9.

118. Brakebusch, C. and R. Fassler, NEW EMBO MEMBER'S REVIEW: The integrin-actin connection, an eternal love affair. Embo J, 2003. 22(10): p. 2324-33.

119. Calderwood, D.A., Talin controls integrin activation. Biochem Soc Trans, 2004. 32(Pt3): p. 434-7.

120. Chen, H.C., et al., Interaction of focal adhesion kinase with cy-toskeletal protein talin. J Biol Chem, 1995. 270(28): p. 16995-9.

121. Hemmings, L., et al., Talin contains three actin-binding sites each of which is adjacent to a vinculin-binding site. J Cell Sci, 1996. 109 ( Pt 11): p. 2715-26.

122. Izard, T., et al., Vinculin activation by talin through helical bundle conversion. Nature, 2004. 427(6970): p. 171-5.

62

123. Huttelmaier, S., et al., The interaction of the cell-contact proteins VASP and vinculin is regulated by phosphatidylinositol-4,5-bisphosphate. Curr Biol, 1998. 8(9): p. 479-88.

124. Gilmore, A.P. and K. Burridge, Regulation of vinculin binding to talin and actin by phosphatidyl-inositol-4-5-bisphosphate. Nature, 1996. 381(6582): p. 531-5.

125. van der Flier, A. and A. Sonnenberg, Structural and functional as-pects of filamins. Biochim Biophys Acta, 2001. 1538(2-3): p. 99-117.

126. Turner, C.E., Paxillin and focal adhesion signalling. Nat Cell Biol, 2000. 2(12): p. E231-6.

127. Wiesner, S., K.R. Legate, and R. Fassler, Integrin-actin interactions.Cell Mol Life Sci, 2005.

128. Turner, C.E., Paxillin interactions. J Cell Sci, 2000. 113 Pt 23: p. 4139-40.

129. Panetti, T.S., TYROSINE PHOSPHORYLATION OF PAXILLIN, FAK, AND p130CAS: EFFECTS ON CELL SPREADING AND MI-GRATION. Front Biosci, 2002. 7: p. D143-50.

130. Calderwood, D.A., et al., Increased filamin binding to beta-integrin cytoplasmic domains inhibits cell migration. Nat Cell Biol, 2001. 3(12): p. 1060-8.

131. Schaller, M.D., et al., pp125FAK a structurally distinctive protein-tyrosine kinase associated with focal adhesions. Proc Natl Acad Sci U S A, 1992. 89(11): p. 5192-6.

132. Hanks, S.K., et al., Focal adhesion protein-tyrosine kinase phos-phorylated in response to cell attachment to fibronectin. Proc Natl Acad Sci U S A, 1992. 89(18): p. 8487-91.

133. Schaller, M.D., et al., Autophosphorylation of the focal adhesion kinase, pp125FAK, directs SH2-dependent binding of pp60src. Mol Cell Biol, 1994. 14(3): p. 1680-8.

134. Owen, J.D., et al., Induced focal adhesion kinase (FAK) expression in FAK-null cells enhances cell spreading and migration requiring both auto- and activation loop phosphorylation sites and inhibits adhesion-dependent tyrosine phosphorylation of Pyk2. Mol Cell Biol, 1999. 19(7): p. 4806-18.

135. Han, D.C. and J.L. Guan, Association of focal adhesion kinase with Grb7 and its role in cell migration. J Biol Chem, 1999. 274(34): p. 24425-30.

136. Zhang, X., et al., Focal adhesion kinase promotes phospholipase C-gamma1 activity. Proc Natl Acad Sci U S A, 1999. 96(16): p. 9021-6.

137. Chen, H.C., et al., Phosphorylation of tyrosine 397 in focal adhesion kinase is required for binding phosphatidylinositol 3-kinase. J Biol Chem, 1996. 271(42): p. 26329-34.

138. O'Neill, G.M., S.J. Fashena, and E.A. Golemis, Integrin signalling: a new Cas(t) of characters enters the stage. Trends Cell Biol, 2000. 10(3): p. 111-9. _00001714 _00001714.

63

139. Parsons, J.T., Focal adhesion kinase: the first ten years. J Cell Sci, 2003. 116(Pt 8): p. 1409-16.

140. Martin, K.H., S.A. Boerner, and J.T. Parsons, Regulation of focal adhesion targeting and inhibitory functions of the FAK related pro-tein FRNK using a novel estrogen receptor "switch". Cell Motil Cy-toskeleton, 2002. 51(2): p. 76-88.

141. Ceccarelli, D.F., et al., Crystal structure of the FERM domain of focal adhesion kinase. J Biol Chem, 2006. 281(1): p. 252-9.

142. Mitra, S.K., D.A. Hanson, and D.D. Schlaepfer, Focal adhesion kinase: in command and control of cell motility. Nat Rev Mol Cell Biol, 2005. 6(1): p. 56-68.

143. Schlaepfer, D.D., C.R. Hauck, and D.J. Sieg, Signaling through focal adhesion kinase. Prog Biophys Mol Biol, 1999. 71(3-4): p. 435-78.

144. Ilic, D., et al., Reduced cell motility and enhanced focal adhesion contact formation in cells from FAK-deficient mice. Nature, 1995. 377(6549): p. 539-544.

145. Avraham, H., et al., RAFTK/Pyk2-mediated cellular signalling. Cell Signal, 2000. 12(3): p. 123-33.

146. Andreev, J., et al., Identification of a new Pyk2 target protein with Arf-GAP activity. Mol Cell Biol, 1999. 19(3): p. 2338-50.

147. Sieg, D.J., et al., Pyk2 and Src-family protein-tyrosine kinases com-pensate for the loss of FAK in fibronectin-stimulated signaling events but Pyk2 does not fully function to enhance FAK- cell migra-tion. Embo J, 1998. 17(20): p. 5933-47.

148. Jove, R. and H. Hanafusa, Cell transformation by the viral src onco-gene. Annu Rev Cell Biol, 1987. 3: p. 31-56.

149. Frame, M.C., Newest findings on the oldest oncogene; how activated src does it. J Cell Sci, 2004. 117(Pt 7): p. 989-98.

150. Thomas, S.M. and J.S. Brugge, Cellular functions regulated by Src family kinases. Annu Rev Cell Dev Biol, 1997. 13: p. 513-609.

151. Tatosyan, A.G. and O.A. Mizenina, Kinases of the Src family: struc-ture and functions. Biochemistry (Mosc), 2000. 65(1): p. 49-58.

152. Renshaw, M.W., J.R. McWhirter, and J.Y. Wang, The human leu-kemia oncogene bcr-abl abrogates the anchorage requirement but not the growth factor requirement for proliferation. Mol Cell Biol, 1995. 15(3): p. 1286-93.

153. Lewis, J.M., et al., Integrin regulation of c-Abl tyrosine kinase activ-ity and cytoplasmic-nuclear transport. Proc Natl Acad Sci U S A, 1996. 93(26): p. 15174-9.

154. Wu, C., The PINCH-ILK-parvin complexes: assembly, functions and regulation. Biochim Biophys Acta, 2004. 1692(2-3): p. 55-62.

155. Bokoch, G.M., Biology of the p21-activated kinases. Annu Rev Bio-chem, 2003. 72: p. 743-81.

156. Dhillon, A.S., et al., A Raf-1 mutant that dissociates MEK/extracellular signal-regulated kinase activation from malig-

64

nant transformation and differentiation but not proliferation. Mol Cell Biol, 2003. 23(6): p. 1983-93.

157. Slack-Davis, J.K., et al., PAK1 phosphorylation of MEK1 regulates fibronectin-stimulated MAPK activation. J Cell Biol, 2003. 162(2): p. 281-91.

158. Dhillon, A.S. and W. Kolch, Untying the regulation of the Raf-1 kinase. Arch Biochem Biophys, 2002. 404(1): p. 3-9.

159. Howe, A.K. and R.L. Juliano, Regulation of anchorage-dependent signal transduction by protein kinase A and p21-activated kinase.Nat Cell Biol, 2000. 2(9): p. 593-600.

160. Jaken, S., K. Leach, and T. Klauck, Association of type 3 protein kinase C with focal contacts in rat embryo fibroblasts. J Cell Biol, 1989. 109(2): p. 697-704.

161. Ivaska, J., et al., Integrin-protein kinase C relationships. Biochem Soc Trans, 2003. 31(Pt 1): p. 90-3.

162. Toker, A., Phosphoinositides and signal transduction. Cell Mol Life Sci, 2002. 59(5): p. 761-79.

163. Parsons, M., et al., Site-directed perturbation of protein kinase C- integrin interaction blocks carcinoma cell chemotaxis. Mol Cell Biol, 2002. 22(16): p. 5897-911.

164. Ng, T., et al., PKCalpha regulates beta1 integrin-dependent cell motility through association and control of integrin traffic. Embo J, 1999. 18(14): p. 3909-23.

165. Ivaska, J., et al., PKCepsilon controls the traffic of beta1 integrins in motile cells. Embo J, 2002. 21(14): p. 3608-19.

166. Liliental, J. and D.D. Chang, Rack1, a receptor for activated protein kinase C, interacts with integrin beta subunit. J Biol Chem, 1998. 273(4): p. 2379-83.

167. Sengupta, J., et al., Identification of the versatile scaffold protein RACK1 on the eukaryotic ribosome by cryo-EM. Nat Struct Mol Biol, 2004. 11(10): p. 957-62.

168. Link, A.J., et al., Direct analysis of protein complexes using mass spectrometry. Nat Biotechnol, 1999. 17(7): p. 676-82.

169. de Hoog, C.L., L.J. Foster, and M. Mann, RNA and RNA binding proteins participate in early stages of cell spreading through spreading initiation centers. Cell, 2004. 117(5): p. 649-62.

170. Nilsson, J., et al., Regulation of eukaryotic translation by the RACK1 protein: a platform for signalling molecules on the ribosome.EMBO Rep, 2004. 5(12): p. 1137-41.

171. DeMali, K.A., K. Wennerberg, and K. Burridge, Integrin signaling to the actin cytoskeleton. Curr Opin Cell Biol, 2003. 15(5): p. 572-82.

172. Jamieson, J.S., et al., Paxillin is essential for PTP-PEST-dependent regulation of cell spreading and motility: a role for paxillin kinase linker. J Cell Sci, 2005. 118(Pt 24): p. 5835-47.

65

173. Angers-Loustau, A., et al., Protein tyrosine phosphatase-PEST regu-lates focal adhesion disassembly, migration, and cytokinesis in fi-broblasts. J Cell Biol, 1999. 144(5): p. 1019-31.

174. Garton, A.J. and N.K. Tonks, Regulation of fibroblast motility by the protein tyrosine phosphatase PTP-PEST. J Biol Chem, 1999. 274(6): p. 3811-8.

175. Zhang, S.Q., et al., Shp2 regulates SRC family kinase activity and Ras/Erk activation by controlling Csk recruitment. Mol Cell, 2004. 13(3): p. 341-55.

176. Kawabuchi, M., et al., Transmembrane phosphoprotein Cbp regu-lates the activities of Src-family tyrosine kinases. Nature, 2000. 404(6781): p. 999-1003.

177. Yamada, K.M. and M. Araki, Tumor suppressor PTEN: modulator of cell signaling, growth, migration and apoptosis. J Cell Sci, 2001. 114(Pt 13): p. 2375-82.

178. Tamura, M., et al., PTEN interactions with focal adhesion kinase and suppression of the extracellular matrix-dependent phosphatidy-linositol 3-kinase/Akt cell survival pathway. J Biol Chem, 1999. 274(29): p. 20693-703.

179. Tonks, N.K., C.D. Diltz, and E.H. Fischer, Purification of the major protein-tyrosine-phosphatases of human placenta. J Biol Chem, 1988. 263(14): p. 6722-30.

180. Tonks, N.K., C.D. Diltz, and E.H. Fischer, Characterization of the major protein-tyrosine-phosphatases of human placenta. J Biol Chem, 1988. 263(14): p. 6731-7.

181. Cheng, A., et al., Attenuation of adhesion-dependent signaling and cell spreading in transformed fibroblasts lacking protein tyrosine phosphatase-1B. J Biol Chem, 2001. 276(28): p. 25848-55.

182. Hernandez, M.V., et al., ER-bound PTP1B is targeted to newly forming cell-matrix adhesions. J Cell Sci, 2006.

183. Zhang, Z., et al., Phosphorylated alpha-actinin and protein-tyrosine phosphatase 1B coregulate the disassembly of the focal adhesion kinase x Src complex and promote cell migration. J Biol Chem, 2006. 281(3): p. 1746-54.

184. Pallen, C.J., Protein tyrosine phosphatase alpha (PTPalpha): a Src family kinase activator and mediator of multiple biological effects.Curr Top Med Chem, 2003. 3(7): p. 821-35.

185. Zeng, L., et al., PTP alpha regulates integrin-stimulated FAK auto-phosphorylation and cytoskeletal rearrangement in cell spreading and migration. J Cell Biol, 2003. 160(1): p. 137-46.

186. Chen, M., S.C. Chen, and C.J. Pallen, Integrin-induced tyrosine phosphorylation of PTPalpha is required for cytoskeletal reorgani-zation and cell migration. J Biol Chem, 2006.

187. Rosen, M.K., et al., Direct demonstration of an intramolecular SH2-phosphotyrosine interaction in the Crk protein. Nature, 1995. 374(6521): p. 477-9.

66

188. Gustavsson, A., M. Yuan, and M. Fallman, Temporal dissection of beta1-integrin signaling indicates a role for p130Cas-Crk in filopo-dia formation. J Biol Chem, 2004. 279(22): p. 22893-901.

189. Cary, L.A., et al., Identification of p130Cas as a mediator of focal adhesion kinase-promoted cell migration. J Cell Biol, 1998. 140(1): p. 211-21.

190. Armulik, A., T. Velling, and S. Johansson, The integrin beta1 sub-unit transmembrane domain regulates phosphatidylinositol 3-kinase-dependent tyrosine phosphorylation of Crk-associated sub-strate. Mol Biol Cell, 2004. 15(6): p. 2558-67.

191. Ruest, P.J., et al., Mechanisms of CAS substrate domain tyrosine phosphorylation by FAK and Src. Mol Cell Biol, 2001. 21(22): p. 7641-52.

192. Legate, K.R., et al., ILK, PINCH and parvin: the tIPP of integrin signalling. Nat Rev Mol Cell Biol, 2006. 7(1): p. 20-31.

193. Hashimoto, S., et al., Interaction of paxillin with p21-activated Kinase (PAK). Association of paxillin alpha with the kinase-inactive and the Cdc42-activated forms of PAK3. J Biol Chem, 2001. 276(8): p. 6037-45.

194. Turner, C.E., et al., Paxillin LD4 motif binds PAK and PIX through a novel 95-kD ankyrin repeat, ARF-GAP protein: A role in cy-toskeletal remodeling. J Cell Biol, 1999. 145(4): p. 851-63.

195. Kolch, W., Coordinating ERK/MAPK signalling through scaffolds and inhibitors. Nat Rev Mol Cell Biol, 2005. 6(11): p. 827-37.

196. Ravichandran, K.S., Signaling via Shc family adapter proteins. On-cogene, 2001. 20(44): p. Shc.

197. Wary, K.K., et al., The adaptor protein Shc couples a class of in-tegrins to the control of cell cycle progression. Cell, 1996. 87(4): p. 733-43.

198. Barberis, L., et al., Distinct roles of the adaptor protein Shc and focal adhesion kinase in integrin signaling to ERK. J Biol Chem, 2000. 275(47): p. 36532-40.

199. Zamir, E. and B. Geiger, Components of cell-matrix adhesions. J Cell Sci, 2001. 114(Pt 20): p. 3577-9.

200. Insall, R.H. and O.D. Weiner, PIP3, PIP2, and cell movement--similar messages, different meanings? Dev Cell, 2001. 1(6): p. 743-7.

201. Tonks, N.K. and M.P. Myers, Structural assets of a tumor suppres-sor. Science, 1999. 286(5447): p. 2096-7.

202. Vazquez, F. and W.R. Sellers, The PTEN tumor suppressor protein: an antagonist of phosphoinositide 3- kinase signaling. Biochim Bio-phys Acta, 2000. 1470(1): p. M21-35.

203. Bairstow, S.F., K. Ling, and R.A. Anderson, Phosphatidylinositol phosphate kinase type Igamma directly associates with and regu-lates Shp-1 tyrosine phosphatase. J Biol Chem, 2005. 280(25): p. 23884-91.

67

204. Toker, A., Protein kinases as mediators of phosphoinositide 3-kinase signaling. Mol Pharmacol, 2000. 57(4): p. 652-8.

205. Foster, F.M., et al., The phosphoinositide (PI) 3-kinase family. J Cell Sci, 2003. 116(Pt 15): p. 3037-40.

206. Reiske, H.R., et al., Requirement of phosphatidylinositol 3-kinase in focal adhesion kinase-promoted cell migration. J Biol Chem, 1999. 274(18): p. 12361-6.

207. Hall, A., Rho GTPases and the control of cell behaviour. Biochem Soc Trans, 2005. 33(Pt 5): p. 891-5.

208. Roberts, M., et al., PDGF-regulated rab4-dependent recycling of alphavbeta3 integrin from early endosomes is necessary for cell ad-hesion and spreading. Curr Biol, 2001. 11(18): p. 1392-402.

209. Schaller, M.D., Biochemical signals and biological responses elic-ited by the focal adhesion kinase. Biochim Biophys Acta, 2001. 1540(1): p. 1-21.

210. Rosenberger, G., et al., Interaction of alphaPIX (ARHGEF6) with beta-parvin (PARVB) suggests an involvement of alphaPIX in in-tegrin-mediated signaling. Hum Mol Genet, 2003. 12(2): p. 155-67.

211. Degani, S., et al., The integrin cytoplasmic domain-associated pro-tein ICAP-1 binds and regulates Rho family GTPases during cell spreading. J Cell Biol, 2002. 156(2): p. 377-88.

212. Sakai, T., et al., Integrin-linked kinase (ILK) is required for polariz-ing the epiblast, cell adhesion, and controlling actin accumulation.Genes Dev, 2003. 17(7): p. 926-40.

213. Juliano, R.L., et al., Integrin regulation of cell signalling and motil-ity. Biochem Soc Trans, 2004. 32(Pt3): p. 443-6.

214. Brakebusch, C. and R. Fassler, beta 1 integrin function in vivo: ad-hesion, migration and more. Cancer Metastasis Rev, 2005. 24(3): p. 403-11.

215. Webb, D.J., J.T. Parsons, and A.F. Horwitz, Adhesion assembly, disassembly and turnover in migrating cells -- over and over and over again. Nat Cell Biol, 2002. 4(4): p. E97-100.

216. Hood, J.D. and D.A. Cheresh, Role of integrins in cell invasion and migration. Nat Rev Cancer, 2002. 2(2): p. 91-100.

217. Caswell, P.T. and J.C. Norman, Integrin trafficking and the control of cell migration. Traffic, 2006. 7(1): p. 14-21.

218. Bretscher, M.S., Circulating integrins: alpha 5 beta 1, alpha 6 beta 4 and Mac-1, but not alpha 3 beta 1, alpha 4 beta 1 or LFA-1. Embo J, 1992. 11(2): p. 405-10.

219. Upla, P., et al., Clustering induces a lateral redistribution of alpha 2 beta 1 integrin from membrane rafts to caveolae and subsequent protein kinase C-dependent internalization. Mol Biol Cell, 2004. 15(2): p. 625-36.

220. Fabbri, M., et al., Dynamic partitioning into lipid rafts controls the endo-exocytic cycle of the alphaL/beta2 integrin, LFA-1, during leu-kocyte chemotaxis. Mol Biol Cell, 2005. 16(12): p. 5793-803.

68

221. De Deyne, P.G., et al., The vitronectin receptor associates with clathrin-coated membrane domains via the cytoplasmic domain of its beta5 subunit. J Cell Sci, 1998. 111 ( Pt 18): p. 2729-40.

222. Wary, K.K., et al., A requirement for caveolin-1 and associated kinase Fyn in integrin signaling and anchorage-dependent cell growth. Cell, 1998. 94(5): p. 625-34.

223. Roberts, M.S., et al., Protein kinase B/Akt acts via glycogen syn-thase kinase 3 to regulate recycling of alpha v beta 3 and alpha 5 beta 1 integrins. Mol Cell Biol, 2004. 24(4): p. 1505-15.

224. Woods, A.J., et al., PKD1/PKCmu promotes alphavbeta3 integrin recycling and delivery to nascent focal adhesions. Embo J, 2004. 23(13): p. 2531-43.

225. Danen, E.H. and K.M. Yamada, Fibronectin, integrins, and growth control. J Cell Physiol, 2001. 189(1): p. 1-13.

226. Clarke, A.S., et al., Activation of the p21 pathway of growth arrest and apoptosis by the beta 4 integrin cytoplasmic domain. J Biol Chem, 1995. 270(39): p. 22673-6.

227. Frisch, S.M., et al., Control of adhesion-dependent cell survival by focal adhesion kinase. J Cell Biol, 1996. 134(3): p. 793-9.

228. Sonoda, Y., et al., FAK is the upstream signal protein of the phos-phatidylinositol 3-kinase-Akt survival pathway in hydrogen perox-ide-induced apoptosis of a human glioblastoma cell line. J Biol Chem, 1999. 274(15): p. 10566-70.

229. Almeida, E.A., et al., Matrix survival signaling: from fibronectin via focal adhesion kinase to c-Jun NH(2)-terminal kinase. J Cell Biol, 2000. 149(3): p. 741-54.

230. Parsons, J.T. and S.J. Parsons, Src family protein tyrosine kinases: cooperating with growth factor and adhesion signaling pathways.Curr Opin Cell Biol, 1997. 9(2): p. 187-92.

231. Schlaepfer, D.D. and T. Hunter, Integrin signalling and tyrosine phosphorylation: just the FAKs? Trends Cell Biol, 1998. 8(4): p. 151-7.

232. Giancotti, F.G. and G. Tarone, Positional control of cell fate through joint integrin/receptor protein kinase signaling. Annu Rev Cell Dev Biol, 2003. 19: p. 173-206.

233. Walker, J.L. and R.K. Assoian, Integrin-dependent signal transduc-tion regulating cyclin D1 expression and G1 phase cell cycle pro-gression. Cancer Metastasis Rev, 2005. 24(3): p. 383-93.

234. Miyamoto, S., et al., Integrins can collaborate with growth factors for phosphorylation of receptor tyrosine kinases and MAP kinase activation: roles of integrin aggregation and occupancy of recep-tors. J Cell Biol, 1996. 135(6 Pt 1): p. 1633-42.

235. Giancotti, F.G. and E. Ruoslahti, Integrin signaling. Science, 1999. 285(5430): p. 1028-32.

236. Fassler, R., et al., Lack of beta 1 integrin gene in embryonic stem cells affects morphology, adhesion, and migration but not integra-

69

tion into the inner cell mass of blastocysts. J Cell Biol, 1995. 128(5):p. 979-88.

237. Sieg, D.J., C.R. Hauck, and D.D. Schlaepfer, Required role of focal adhesion kinase (FAK) for integrin-stimulated cell migration. J Cell Sci, 1999. 112(Pt 16): p. 2677-91.

238. Balzac, F., et al., Expression of beta 1B integrin isoform in CHO cells results in a dominant negative effect on cell adhesion and mo-tility. J Cell Biol, 1994. 127(2): p. 557-65.

239. Retta, S.F., et al., 1-integrin cytoplasmic subdomains involved in dominant negative function. Mol Biol Cell, 1998. 9(4): p. 715-31.

240. Isberg, R.R. and J.M. Leong, Multiple beta 1 chain integrins are receptors for invasin, a protein that promotes bacterial penetration into mammalian cells. Cell, 1990. 60(5): p. 861-71.

241. Gustavsson, A., et al., Role of the beta1-integrin cytoplasmic tail in mediating invasin-promoted internalization of Yersinia. J Cell Sci, 2002. 115(Pt 13): p. 2669-78.

242. Klinghoffer, R.A., et al., Src family kinases are required for integrin but not PDGFR signal transduction. Embo J, 1999. 18(9): p. 2459-71.

243. Reif, S., et al., The role of focal adhesion kinase-phosphatidylinositol 3-kinase-akt signaling in hepatic stellate cell proliferation and type I collagen expression. J Biol Chem, 2003. 278(10): p. 8083-90.

244. Shen, T.L. and J.L. Guan, Differential regulation of cell migration and cell cycle progression by FAK complexes with Src, PI3K, Grb7 and Grb2 in focal contacts. FEBS Lett, 2001. 499(1-2): p. 176-81.

245. Akagi, T., et al., v-Crk activates the phosphoinositide 3-kinase/AKT pathway in transformation. Proc Natl Acad Sci U S A, 2000. 97(13):p. 7290-5.

246. Schlessinger, J., Cell signaling by receptor tyrosine kinases. Cell, 2000. 103(2): p. 211-25.

247. Assoian, R.K. and M.A. Schwartz, Coordinate signaling by in-tegrins and receptor tyrosine kinases in the regulation of G1 phase cell-cycle progression. Curr Opin Genet Dev, 2001. 11(1): p. 48-53.

248. Howe, A.K., A.E. Aplin, and R.L. Juliano, Anchorage-dependent ERK signaling--mechanisms and consequences. Curr Opin Genet Dev, 2002. 12(1): p. 30-5.

249. Biondi, R.M., Phosphoinositide-dependent protein kinase 1, a sen-sor of protein conformation. Trends Biochem Sci, 2004. 29(3): p. 136-42.

250. Trakul, N., et al., Raf kinase inhibitory protein regulates Raf-1 but not B-Raf kinase activation. J Biol Chem, 2005. 280(26): p. 24931-40.

251. Corbit, K.C., et al., Activation of Raf-1 signaling by protein kinase C through a mechanism involving Raf kinase inhibitory protein. J Biol Chem, 2003. 278(15): p. 13061-8.

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 136

Editor: The Dean of the Faculty of Medicine

A doctoral dissertation from the Faculty of Medicine, UppsalaUniversity, is usually a summary of a number of papers. A fewcopies of the complete dissertation are kept at major Swedishresearch libraries, while the summary alone is distributedinternationally through the series Digital ComprehensiveSummaries of Uppsala Dissertations from the Faculty ofMedicine. (Prior to January, 2005, the series was publishedunder the title “Comprehensive Summaries of UppsalaDissertations from the Faculty of Medicine”.)

Distribution: publications.uu.seurn:nbn:se:uu:diva-6747

ACTAUNIVERSITATISUPSALIENSISUPPSALA2006