Differential regulation of cell adhesive functions by

JCS ePress online publication date 23 April 2003

Research Article

2333

Differential regulation of cell adhesive functions by integrin α subunit cytoplasmic tails in vivo Jie Na, Mungo Marsden and Douglas W. DeSimone* Department of Cell Biology, University of Virginia School of Medicine, PO Box 800732, Charlottesville, VA 22908, USA *Author for correspondence (e-mail: [email protected])

Accepted 27 February 2003 Journal of Cell Science 116, 2333-2343 © 2003 The Company of Biologists Ltd doi:10.1242/jcs.00445

Summary Cell adhesion to fibronectin (FN) is crucial for early vertebrate morphogenesis. In Xenopus gastrulae, several distinct integrin-dependent adhesive behaviors can be identified: adhesion of cells to FN, assembly of FN fibrils, and initiation of cell spreading and migration in response to mesoderm inducing signals. We have taken a chimeric integrin approach to investigate the role of the integrin α cytoplasmic tail in the specification of these developmentally significant adhesive functions. Cytoplasmic tail-deleted α4 constructs and α4ectodomain/α-cytoplasmic tail chimeras were generated and expressed in whole embryos. Normal gastrula cells lack integrin α4 and, correspondingly, are unable to adhere to the α4 ligand, the V-region of FN. The ability of α4 constructs to promote adhesive behaviors was established by placing tissue explants or dissociated cells on an FN V-

Key words: Integrin, Cell migration, Cell adhesion, Xenopus, Fibronectin

Introduction Integrin-mediated cell adhesion is crucial for many physiological and pathological processes such as embryo development, tissue differentiation, immune response, wound healing, tumorigenesis and metastasis (Adams and Watt, 1993; Brown et al., 2000; Dzamba et al., 2002; Hynes, 1992). Integrin cytoplasmic domains are targets of intracellular signals that modulate adhesion (Dedhar and Hannigan, 1996; Liu et al., 2000; Sastry and Horwitz, 1993). The cytoplasmic domains of integrin α subunits have been shown to confer specific receptor activation states and trigger unique post-ligand binding events (Ginsberg et al., 1992; Hemler et al., 1992; Ivaska et al., 1999; Sastry et al., 1996; Weber et al., 1999). Integrin activation is hypothesized to involve the propagation of a conformational change from the cytoplasmic domain, resulting in high-affinity ligand binding. This is aptly illustrated by the well-studied mammalian platelet integrin αIIbβ3 (Hughes and Pfaff, 1998; Shattil et al., 1994). Normally, αΙΙb β3 is expressed at the surface of resting platelets in a low-affinity state. On platelet stimulation, αΙΙb β3 is activated by intracellular (‘inside-out’) signals, binds fibrinogen with high affinity and can be detected by ligand-induced binding site antibodies, which specifically recognize high-affinity conformations of the receptor (Shattil and Brass, 1987). The conserved membrane proximal GFFKR motif of αΙΙb is important for maintaining the receptor in a default low-affinity state; deletion or mutation of key amino acids within this motif resets the integrin to a high-affinity state

(Hughes et al., 1996; O’Toole et al., 1991). Swapping the αIIb cytoplasmic tail with the α2, α5, α6A or α6B tails leads to constitutive receptor activation; however, swapping it with the αM, αL, or αv tails results in a low-affinity receptor (O’Toole et al., 1994). Thus, α cytoplasmic tail sequences can specify different receptor activation states. Alterations in integrin activation state can trigger distinct cell adhesive functions. For example, when the αIIb cytoplasmic tail is replaced with an integrin α5 tail or carries deletions in the GFFKR motif, the receptor becomes competent to mediate the assembly of fibronectin (FN) fibrils (Wu et al., 1995b). The cytoplasmic domains of α subunits can also initiate unique post-ligand-binding signaling events (Dedhar and Hannigan, 1996; Sastry and Horwitz, 1993). The α4 tail can bind directly to the adapter protein paxillin, and this property is important for α4β1 to support strong migration (Liu and Ginsberg, 2000). Nischarin, a novel protein that can affect Rho family GTPase function, is reported to associate with the α5 tail (Alahari et al., 2000). Cells expressing chimeric integrins with the α2 extracellular and transmembrane domains fused to an α4 tail showed enhanced migration, whereas strong collagen gel contraction was observed with the α2 and α5 tails (Chan et al., 1992). In myoblasts, enhanced paxillin and mitogen-activated protein kinase (MAPK) activation is associated with integrin α5, whereas focal adhesion kinase (FAK) and MAPK are suppressed on ectopic expression of α6A. Ιn each case, the cytoplasmic tails of these subunits are

region fusion protein that lacks the RGD (Arg-GlyAsp)/synergy sites or treating whole embryos with antibodies that block endogenous integrin-FN interactions. We found that each α4 cytoplasmic domain deletion mutant and α-tail chimera examined could support cell attachment; however, activin induction-dependent cell spreading, mesoderm cell and explant motility, and the ability to assemble FN matrix on the blastocoel roof varied with specific α subunit tail sequences. These data suggest that α cytoplasmic tail signaling and changes in integrin activation state can regulate a variety of developmentally significant adhesive behaviors in both space and time.

2334

Journal of Cell Science 116 (11)

believed to act by regulating β1 cytoplasmic tail functions (Sastry et al., 1999). Differences in integrin signaling can also influence cell fate decisions. Cytoplasmic domain-swapping experiments using quail skeletal muscle cells show that the α5 tail can promote proliferation, whereas the α6 tail affects differentiation (Sastry et al., 1996). Although studies such as these have led to important general insights regarding the functional importance of integrin α cytoplasmic tails, specific results vary depending on the integrin and/or cell type under investigation (Kassner and Hemler, 1993; O’Toole et al., 1994). To reveal the contribution of integrin α tail-specific signaling in various physiological and pathological processes, therefore, it is important to conduct these studies in systems amenable to in vivo analyses. Studies of α tail function in Xenopus gastrulae can provide useful information about the molecular machinery that controls position-specific changes in adhesive behaviors in development. The advantage of the Xenopus system is that embryonic cells or tissue fragments can be dissected and cultured under conditions that allow the progression of normal in vivo behaviors and developmental fates. In early Xenopus embryos, integrins act to assemble FN fibrils along the blastocoel roof (BCR), mediate the adhesion and migration of mesendodermal cells (Davidson et al., 2002), and are involved in regulating the polarity of cells engaged in cell intercalation movements (Marsden and DeSimone, 2001). At least three integrins, α5β1, α3β1 and an αV-containing integrin (Joos et al., 1995; Joos et al., 1998; Meng et al., 1997; Whittaker and DeSimone, 1993), are expressed at these stages of development and may potentiate the adhesive behaviors of cells on FN. Moreover, integrin adhesive activity is regulated in both space and time by inductive interactions that help to specify mesoderm and initiate gastrulation movements (Ramos and DeSimone, 1996; Ramos et al., 1996). Our results establish that α cytoplasmic tail sequences following the membrane-proximal GFFK/RR motif are required to confer position-specific cell-adhesive behaviors in embryos at the gastrula stage. Distinct α cytoplasmic tails differ in their ability to support FN fibrillogenesis on the BCR and the migration of mesodermal cell and tissue explants. Our data also suggest that integrin α cytoplasmic domains carry distinct, although often overlapping, functional properties.

Materials and Methods Preparation of Xenopus embryos Adult Xenopus laevis were purchased from Nasco (Fort Atkinson, WI). Eggs and embryos were obtained as described by Newport and Kirschner (Newport and Kirschner, 1982). Embryos were cultured in 0.1× Modified Barth’s Saline (MBS) (Gurdon, 1977) and staged according to Nieuwkoop and Faber (Nieuwkoop and Faber, 1967). Construction of α4 chimeric integrins The Xenopus integrin α4 cDNA (Whittaker and DeSimone, 1998) and the two cytoplasmic tail truncation mutants were prepared by Charles Whittaker. The α4 chimeras were engineered using PCR-based methods. In brief, the DNA fragment that encodes the α4 extracellular and transmembrane domains, as well as the amino acids KVGFF of the cytoplasmic tail was amplified from the Xenopus integrin α4 cDNA (basepairs 1 to 3113). The cytoplasmic domain segments of different integrin α subunits were amplified starting with sequence

encoding K or R (α2 Κ760, α3 R977, α5 K1028, α6 R1057, αV K1007) of the conserved GFFK/RR domain. The two DNA fragments were then digested with restriction enzymes, ligated and subcloned into the pCS2+ vector. In vitro transcription of capped RNA and micro-injection Capped RNA was transcribed in vitro from linearized plasmid DNA using SP6 RNA polymerase (Promega, Madison, WI). Following the transcription reaction, unincorporated nucleotides were removed using ProbeQuant G-50 Micro Columns (Amersham Biosciences, Piscataway, NJ) and resuspended in distilled water at desired concentrations. The amount of transcript injected varied according to the specific construct, but it was typically in the range of 1.0-5.0 ng/ embryo. Micro-injection was performed by targeting either the animal pole or the future dorsal marginal zone regions of eggs or two-cellstage blastomeres, depending on the experiment. Antibody generation Two PCR-amplified fragments encoding the C- and N-termini of the proteolytically cleaved extracellular domain of Xenopus α4 were each subcloned into the pGEX-KG vector (Guan and Dixon, 1991) to generate two glutathione-S-transferase (GST)-α4 fusion proteins; α4EX-S contained amino acids K557 to E606, and α4EX-LS contained amino acids G607 to E726. GST fusion proteins were prepared as described in Guan and Dixon (Guan and Dixon, 1991) and used to immunize rabbits for polyclonal antibody production. Immune-sera were first precleared with a GST cross-linked glutathione-agarose bead column and then affinity purified on a column containing a mixture of α4EX-S and α4EX-LS fusion proteins. GST fusion protein columns were prepared by crosslinking GST to glutathione agarose beads using the method described by Koff et al. (Koff et al., 1992). The affinity-purified polyclonal antibody (PcAb) directed against the Xenopus α4 extracellular domain is named A4EX. Biotinylation of Xenopus embryonic cells Xenopus embryos were dissociated in Ca2+-free, Mg2+-free MBS (CMF) in agarose-coated petri dishes. EZ-linkTM Sulfo-NHS-LCBiotin (Sulfosuccinimidyl-6-biotinamido Hexanoate) from Pierce (Rockford, IL) was used to biotinylate cell-surface proteins according to the method described by Alfandari et al. (Alfandari et al., 1995). Immunoprecipitation and western blot Xenopus embryos were extracted in embryo solublization buffer (ESB) consisting of 100 mM NaCl, 50 mM Tris-HCl (pH=8.0), 1% NP-40 or Triton X-100, 2 mM PMSF (phenylmethylsulphonylfluoride), 1 µg/ml aprotinin, 1 µg/ml leupeptin, 1 µg/ml, pepstatin A. For immunoprecipitation, embryo lysates were first precleared with nonimmune mouse or pre-immune rabbit serum coupled to protein G or protein A agarose beads, followed by immunoprecipitation with desired antibodies. Immunoprecipitated proteins were then separated on 8% polyacrylamide gels and electrotransferred onto nitrocellulose membranes for western blot and ECL detection. Horseradish peroxidase (HRP)-labeled streptavidin was used to detect biotinylated proteins by western blot. Cell adhesion and migration assays GST-FN fusion proteins were prepared and coated onto Falcon petri dishes at a concentration of 0.23 µM according to Ramos et al. (Ramos et al., 1996). The GST fusion proteins used in these experiments included 9.11-GST, which contains the synergy/RGD site of Xenopus FN, and V-GST, which contains only the alternatively spliced V-region of FN (Ramos and DeSimone, 1996). Xenopus

Integrin α cytoplasmic tail functions animal cap cell adhesion assays and activin-induced cell spreading assays were performed as described by Ramos et al. (Ramos et al., 1996). Phase-contrast images were obtained using a Zeiss Axiovert microscope and an MTI VE1000 CCD camera using the NIH Image software package. For dorsal involuted marginal zone (DIMZ) cell migration assays, GFP mRNA was co-injected with the α4 constructs targeting the future dorsal marginal zone region. GFP-expressing cells from the DIMZ region of stage 11 embryos were dissected and dissociated in CMF-MBS. These cells were then transferred onto coated substrates. Digital time-lapse images were recorded with an Orca Camera (Hamamatsu, Bridgewater, NJ) using a Zeiss Axiophot microscope equipped with GFP filter set, and the ISEE (Inovision, ISee Imaging, Raleigh, NC) software package. Paths and parameters of cell migration were tracked using the Nanotrack function within ISEE. Each experiment was repeated with three different batches of embryos; a total of 40-65 cells were tracked and analyzed. Dorsal marginal zone explants and assays Dorsal marginal zone (DMZ) explants were prepared as described previously (Davidson et al., 2002). For migration assays, DMZ explants expressing α4 constructs were first allowed to adhere to FN or V-GST coated substrates for 1 hour before time-lapse image recording. The culture chamber was then placed on a motorized stage (Ludl, Hawthorne, NY) attached to a Zeiss Axiovert microscope. Phase-contrast, time-lapse images were recorded using the MTI VE1000 CCD camera and NIH Image at 1 minute intervals for 1 hour. FN fibril assembly rescue assay Embryos expressing α4 constructs targeted to the animal pole region were injected into the blastocoel at blastula stage 9 with 400 ng of monocloncal antibodies (mAbs) that block integrin α5β1 (P8D4) (Davidson et al., 2002) or the RGD site of Xenopus FN (4B12) (Ramos et al., 1996), as previously described (Ramos et al., 1996). Embryos were cultured until stage 12, then animal caps were dissected and washed in MBS, followed by fixation in 2% trichloroacetic acid (TCA) at 4°C overnight. For detection of α4 constructs, A4EX was used as the primary antibody followed by FITC-labeled goat anti-rabbit immunoglobulin G (IgG) (Jackson Labs, Bar Harbor, ME). For FN, mAb 4H2 (Ramos and DeSimone, 1996) and Rhodamine-labeled goat anti-mouse IgG (Jackson Labs) were used as primary and secondary antibodies, respectively. Florescence was detected using a Zeiss Axiophot microscope and recorded using a Hamamastu Orca Digital Camera and the ISEE software. Each experiment was repeated with three different batches of embryos, and more than ten animal caps for each α4 construct were examined from every experiment.

Results α4 Chimeras and tail truncation mutants are expressed on the surface of embryonic cells and form functional heterodimeric receptors with the β1 subunit In Xenopus embryos, α4 protein is normally not expressed until late neurula stages (Whittaker and DeSimone, 1998) and, correspondingly, cells dissected from gastrula-stage embryos are unable to attach to the V-region of FN, which is a ligand for α4β1 (Ramos and DeSimone, 1996; Ramos et al., 1996). Injection of RNA transcript encoding α4 leads to ectopic expression of the protein by gastrulation and confers the ability of embryonic cells to attach to the V-region (Ramos et al., 1996). Thus, it is possible to study the functions of a single integrin (i.e. α4β1) by providing embryonic cells and explants

2335

Table 1. Amino acid sequences of the cytoplasmic domains of α4 constructs Amino-acid sequences following the α4 transmembrane domain α4wt α4KV α4RR α4α2 α4α3 α4α5 α4α6 α4αV

---α4--KVGFFRRKYQPIGTEETSRRESWNYLNKDEKEVK ---α4--KV ---α4--KVGFFRR ---α4--KVGFFKRKYQKLQKTEEEIAETTQLN ---α4--KVGFFRRANTRAMYEVRGQKAEMKVQPSETERLTQDY ---α4--KVGFFKRSYQYGTAMEKAELKPQAASEA ---α4--KVGFFRRDKKDQFDATYHKAEIHAQPSDKERLTSDA ---α4--KVGFFKRVRPPQEETEREQLQPQENGEGITDT

α4wt is the full-length Xenopus integrin α4. α4KV and α4RR are cytoplasmic tail truncation mutants; α4KV had cytoplasmic tail deleted up to the two amino acids KV immediately after the transmembrane domain, and α4RR had truncation up to the conserved GFFRR domain. Five α4 chimeric integrins, α4α2, α4α3, α4α5, α4α6 and α4αv, had the cytoplasmic tail of Xenopus integrin α2, α3, α5, α6 and αv each joined to α4KV.

that express α4 with the V-region alone as substrate. Alternatively, cells expressing α4β1 can be exposed to intact FN in the presence of antibodies that block endogenous integrin recognition of the FN central cell-binding domain (CCBD; RGD and synergy site-containing region of FN). In this study, Xenopus integrin α4 cytoplasmic tail truncations and chimeric constructs were designed to investigate the functions of different integrin α cytoplasmic tails in the regulation of position-specific adhesive behaviors that occur during Xenopus gastrulation. The cytoplasmic tail sequences of these constructs are listed in Table 1. A rabbit polyclonal antibody (PcAb) A4EX was generated against the Xenopus integrin α4 extracellular domain. Similar to mammalian α4, Xenopus integrin α4 is expressed in two forms – a full-length form (140 kDa) and a proteolytically cleaved form consisting of two products, the partial N-terminal extracellular domain (80 kDa) and the remaining C-terminal portion that contains the cytoplasmic domain (60 kDa) (Hemler et al., 1990). As shown in Fig. 1A, α4 is not expressed in control embryos injected with H2O. A4EX recognizes three bands in the embryos injected with α4wt transcript, representing the full-length (140 kDa) and the cleaved form (80 kDa and 60 kDa) of the α4 protein. In the α4RR lane, the size of the full-length protein and the C-terminal fragment containing the cytoplasmic domain is smaller than that in the α4wt lane because of the tail truncation. D2AP is a PcAb raised against the α4 cytoplasmic tail (Whittaker and DeSimone, 1998) and, thus, only recognizes the 140 kDa and 60 kDa forms of α4wt representing the full-length protein and the cleaved product that has the cytoplasmic tail. As predicted, PcAb D2AP did not recognize α4RR. To confirm that the α4 constructs were expressed on the surface of embryonic cells, Xenopus embryos injected with RNA transcripts encoding α4 constructs were cultured until stage 15 and dissociated in CMF-MBS. Cells were surface labeled with biotin, solublized and subjected to immunoprecipitation using the A4EX PcAb (Fig. 1B). With the exception of α4α2, each of the α4 chimeras and tail truncation mutants was expressed on the embryonic cell surface at comparable levels (i.e. less than a twofold difference based on quantification of pixel densities) and primarily in the cleaved form. The β1 subunit does not label as efficiently as the α4 constructs and is therefore only detected on longer exposures

2336

Journal of Cell Science 116 (11) Fig. 1. Expression of α4 constructs. (A) A4EX recognizes the extracellular domain of Xenopus integrin α4. Embryos injected with full-length Xenopus α4 (α4wt) or partial cytoplasmic tail deletion α4 mutant (α4RR) were lysed and subjected to western blot with A4EX (raised against the α4 extracellular domain) and D2AP (raised against the α4 cytoplasmic tail). Protein equivalent to one embryo was loaded per lane. Arrows indicated bands of 140 kDa, 80 kDa and 60 kDa. A4EX can detect both α4wt and α4RR, whereas D2AP only recognizes α4wt. (B) α4 constructs are expressed on embryonic cell surface. Cell-surface biotinylated α4 tail truncation and chimera proteins (as indicated) are immunoprecipitated by A4EX. Each lane contains protein precipitated from seven embryos. Arrows indicate bands of 80 kDa and 60 kDa. Note that the α4 cytoplasmic tail deletion mutants and chimeras are present at the cell surface predominantly in the cleaved form. α4α2 is expressed on the surface less well than other constructs. Quantification of the pixel densities of the α4 bands indicates a less-than-twofold variation in surface expression from sample to sample, with the exception of α4α2. (C) α4 constructs form heterodimers with the β1 subunit. Embryos co-injected with RNA transcripts of α4 constructs and β1 were lysed at stage 15 and immunoprecipitated with mAb 8c8. The immunoprecipitated proteins were then separated on an 8% polyacrylamide gel and western blotted with A4EX. Precipitate equivalent to ten embryos was loaded per lane. Both full-length and cleaved forms of α4 can associate with β1 subunit.

spread. The interaction between the α4 extracellular domain and the V-region is specific, because cells expressing α4 cannot adhere to GST alone nor to bovine serum albumin (BSA)coated surfaces (not shown). These experiments confirm that α4 constructs are expressed on the embryonic cell surface and act as functional heterodimeric receptors by binding to the V-region of FN. The following studies investigate the importance of αsubunit-specific tail sequences in supporting morphogenetic cell adhesion behaviors in Xenopus. They include cell spreading in response to mesoderm inducing signal, mesodermal cell and tissue migration, and the assembly of FN matrix on the BCR. or at later stages when receptor levels at the surface are increased (Whittaker and DeSimone, 1998) (data not shown). Because we were unable to normalize the surface expression level of α4α2 with other α4 constructs, this chimera is not included in the functional studies described in this paper. To confirm that each α4 construct formed heterodimeric receptors with the β1 subunit, mAb 8c8 (directed against Xenopus β1 integrin) (Gawantka et al., 1992) was used to immunoprecipitate β1 integrins from embryos injected with α4 transcripts. All α4 chimeras and tail deletion mutants were found to co-immunoprecipitate with the β1 subunit (Fig. 1C). Cell adhesion assays were performed to test whether the α4β1 constructs form functional receptors. Animal cap cells isolated from normal gastrula-stage embryos attach to FN (Fig. 2A). These cells do not express integrin α4 and are unable to attach to the V-region of FN (Fig. 2B). Animal cap cells expressing α4 constructs attach to a V-region GST fusion protein substrate, similar to non-injected cells on intact FN (Fig. 2C-I). The α4 cytoplasmic tail is not required for cell attachment; cells expressing either α4KV or α4RR were still able to attach to V-GST (Fig. 2D,E). In all cases, attached cells formed short filopodial protrusions on the substrate, but did not

The GFFK/RR motif is sufficient to support cell spreading in response to activin induction Cells dissected from the animal cap region can only attach to FN, whereas involuted marginal zone cells can spread on FN (Ramos et al., 1996). The spreading behavior of marginal zone cells can be replicated in animal cap cells following exposure to the mesoderm inducing factor activin (Smith et al., 1990). Previous studies suggest that activin induction leads to changes in the functional ‘activation states’ of integrins, which affect adhesive behavior (Ramos and DeSimone, 1996). Because integrin α cytoplasmic tails play important roles in the regulation of integrin activation, we investigated whether specific α subunit tail sequences are required for activin responsiveness. Animal cap cells from embryos injected with α4 construct transcript were treated with 30 units/ml of activin and cultured on V-GST. Spread cells were scored as in Ramos et al. (Ramos et al., 1996) on the basis of polygonal cell shape and the presence of lamellipodial membrane protrusions. With the exception of α4KV, each of the α4 constructs was able to support cell spreading (Fig. 3C,E-I). Activin-treated α4KV cells failed to elongate their cell bodies (Fig. 3D) but were able

Integrin α cytoplasmic tail functions

2337

Fig. 2. α4 Constructs are able to mediate the attachment of dissociated animal cap cells to the V-region of FN. The adhesive substrate and the α4 construct injected are as indicated on each panel. Scale bar for all panels is shown in A.

Fig. 4. ‘Spider’ graphs of individual DIMZ cell migration tracks. The α4 constructs expressed are as indicated. The ‘control’ is GFP transcript-injected DIMZ cells and the substrate used is 9.11-GST (A). For all the α4 constructs, the substrate is V-GST (B-H). Each ‘spider’ graph contains 15 representative cell migration tracks with the start point all set to ‘0,0’. The scale is as indicated in (G). Note that α4KV and α4 RR have reduced paths compared with α4 chimeras with full-length cytoplasmic tails. Among all the α4 chimeras, α4α5 has slightly shorter and more coiled migration tracks.

Fig. 3. Animal cap cell spreading in response to activin induction. The α4 construct injected and the substrate are indicated on each panel. Scale bar for all panels is shown in A.

to form some short ruffle-like protrusions, which were small and highly transient. The α4RR construct was able to support cell spreading in a manner similar to each α4 chimera with a full-length cytoplasmic tail (Fig. 3E). Thus, the conserved GFFK/RR domain is the minimal sequence required to mediate cell spreading in response to activin induction.

Tracking single mesodermal cell migration mediated by α4 constructs Individual mesodermal cells isolated from the dorsal involuted marginal zone (DIMZ) region, unlike untreated animal cap cells, are able to migrate on FN (Ramos et al., 1996; Winklbauer, 1990). To study the migration behaviors of single mesoderm cells expressing different integrin cytoplasmic domains, α4 transcripts were injected into the future dorsal marginal zone region at the four-cell stage. GFP RNA transcripts were co-injected as a marker to trace expression. At stage 10.5, DIMZ tissue with GFP fluorescence was dissected, dissociated to single cells and plated on artificial substrates. The migration of these cells was recorded over a 1 hour period, then tracked and measured using the Nanotrack function within the ISEE software package. ‘Spider’ graphs were used to represent the migration behaviors of these cells (Fig. 4). DIMZ cells migrate actively on 9.11-GST (Fig. 4A), as do DIMZ cells expressing α4 on V-GST (Fig. 4B). Many of these cells are

2338

Journal of Cell Science 116 (11) Table 2. The statistical significance of the travel rate (A) and the radial displacement rate (B) among different α4 constructs A. Travel rate α4KV α4KV α4RR α4α5 α4α3 α4α6 α4wt α4αv

α4RR

α4α5

α4α3

α4α6

α4wt

α4αv

夹

夹 夹

夹 夹 –

夹 夹 夽 –

夹 夹 夽 – –

夹 夹 夽 – – –

B. Radial displacement rate α4KV α4KV α4RR α4α5 α4αv α4wt α4α6 α4α3

α4RR

α4α5

α4αv

α4wt

α4α6

α4α3

夹

夹 夹

夹 夹 夽

夹 夹 夹 –

夹 夹 夹 – –

夹 夹 夹 夽 – –

The statistical significance is decided by first using analysis of variance within different groups followed by paired t test. 夹 indicates significance at 0.01 level. 夽 indicates significance at 0.05 level. – indicates no significance.

Fig. 5. DIMZ cell migration rate mediated by α4 constructs. The travel rate (A) and the radial displacement rate (B) were measured using the Nanotrack function within the ISEE software package, then plotted as bar graphs. The α4 constructs expressed and substrates used are as indicated. For each construct, 40 to 65 cells from three different batches of embryos were measured. Error bars indicate standard deviations. Normal DIMZ cells cannot adhere to V-GST, thus their travel rate and radial displacement are not available. α4KV and α4RR cells migrate significantly slower than α4 chimeras, whereas α4α5 cells migrate the slowest compared with other α4 chimeras. The statistical significance among different α4 constructs is summarized in Table 2.

able to move persistently and translocate significant distances from their points of origin, whereas other cells change their direction more frequently (Fig. 4A,B). Cells expressing α4KV remain close to the point of origin during the 1 hour timelapse recording (Fig. 4C). α4RR (Fig. 4D) supports slightly increased migration compared with α4KV, but the tracks of these cells remain significantly shorter than α4 chimeras containing a full-length cytoplasmic tail (Fig. 4B,E-H). Among all of the chimeras, α4α5 has slightly shorter and less persistent migration tracks (Fig. 4F). The travel distance (the length of a cell migration track) and radial displacement (the straight-line distance from the start point to the end point of the migration) within 1 hour were

quantified to compare the migration behaviors of cells expressing different α4 constructs (Fig. 5A,B). Statistical analyses confirm and extend the data represented by the ‘spider’ graphs (Table 2). Thus, the sequences following the conserved GFFK/RR motif are important for cell motility. Although the interactions between integrin α5 and the synergy/RGD site of FN is required for normal DIMZ cell adhesion and migration on FN (Davidson et al., 2002; Ramos and DeSimone, 1996; Ramos et al., 1996), the α5 cytoplasmic domain alone did not support the highest migration rates in these assays. A full-length α cytoplasmic domain is required for intact dorsal marginal zone tissue explant migration on V-GST At the onset of gastrulation, dorsal marginal zone cells involute at the site of the dorsal blastopore lip and migrate towards the animal pole along the FN matrix deposited on the BCR (Winklbauer and Nagel, 1991). This directional migration is important for gastrulation movements and the formation of the three-layered basic body plan. When isolated dorsal marginal zone (DMZ) tissue explants are placed on FN-coated substrates, the explants are able to reproduce the in vivo behaviors of the DMZ tissue by spreading and migrating out from the dorsal lip as a coherent sheet (Davidson et al., 2002). The DMZ explants are schematically represented in Fig. 6A. These explants were used to investigate the ability of α4 constructs to support the migration of intact mesodermal tissue. DMZ explants injected with GFP RNA transcripts spread and migrate on FN but not the V-region GST fusion protein (Fig. 6B, GFP). DMZ explants expressing α4 chimeras containing a full-length cytoplasmic domain were each able to migrate out on V-GST, similar to explants on FN (Fig. 6C; α4α5 is shown as a representative example in Fig. 6B). The

Integrin α cytoplasmic tail functions

2339

explants on V-GST. The leading edges of both α4KV and α4RR explants were able to send out some membrane protrusions; however, these protrusions were not capable of forming stable anchors with the substrate. The distance of leading edge advancement in 1 hour was measured as the migration rate (Fig. 6C). The α4KV and α4RR explants failed to migrate, whereas α4α5 explants migrated the slowest among each of the α4 chimera explants observed, which is consistent with our previous observation of single DIMZ cell migration behaviors (Fig. 5). Thus, the α cytoplasmic domain sequences beyond the GFFK/RR motif play a general positive role in supporting mesodermal tissue explant migration. The adhesive and migratory behaviors of α4-expressing explants and dissociated cells on V-region substrates (e.g. Figs 2, 3 and 6) are similar to those on intact FN in the presence of antibodies that block endogenous α5β1 binding to the CCBD of FN (data not shown).

Fig. 6. DMZ explant migration assay. (A) A DMZ explant. The bottom layer is the mesoderm tissue sheet (red), which encounters the substrate, and the top layer is ectoderm tissue (blue). The dorsal lip edge is positioned up and the arrow indicates the direction of mesoderm tissue sheet migration. (B) The morphology of DMZ explants expressing GFP or α4α5, shown as a representative example. The construct injected and the adhesive substrate are as indicated. Arrows point to the migrating sheet of mesodermal tissue. (C) Migration rate of DMZ explants. The distance of leading edge cell advancement in 1 hour is measured and plotted as a bar graph. The substrate and construct injected are as indicated.

leading edge cells from the dorsal lip margin of the explants send out frequent lamellipodial protrusions followed by advancement of their cell bodies. This eventually leads to the spreading and migration of the whole mesodermal tissue sheet. In contrast to explants derived from embryos injected with α4 chimera transcripts, DMZ explants expressing α4KV were unable to migrate on V-GST. Despite the ability of α4RR to support cell spreading in response to activin induction and some cell motility in single DIMZ cells (Fig. 3E, Fig. 5D), this construct also failed to mediate migration of intact DMZ

Only α5, α6 and α3 cytoplasmic domains are able to mediate FN fibril assembly on the blastocoel roof (BCR) During gastrulation, cells lining the inner surface of the BCR assemble FN into a fibrillar network (Winklbauer and Stoltz, 1995). This process is dependent on integrin α5β1 in vivo. Antibodies that block either α5β1 or its ligand, the RGD/synergy domain of FN, can abolish FN fibrillogenesis, leading to defects in later development (Marsden and DeSimone, 2001; Ramos and DeSimone, 1996). Integrin α4β1 is competent to assemble FN fibrils in vitro when exogenously activated with Mn2+ or activating antibody (Sechler et al., 2000). Because ectopically expressed α4β1 is able to bind to FN through the V-region, we investigated whether α4 constructs could assemble FN into a fibrillar matrix in vivo under conditions that block endogenous α5β1-dependent assembly. Early cleavage-stage embryos injected in the animal pole with transcripts encoding the α4 constructs were cultured to blastula stages and then injected with mAb P8D4 (Fig. 7A), which blocks integrin α5β1 function and FN fibrillogenesis (Davidson et al., 2002). By the end of gastrulation (stage 12), a dense FN fibrillar network is assembled normally on the BCR of control embryos (Fig. 7B). Injection of P8D4 into the blastocoel completely abolished the formation of FN fibrils (Fig. 7C). The α4wt and two cytoplasmic tail truncation mutants failed to rescue fibril assembly (Fig. 7D,E,F). However, α4 with the α5 or α6 cytoplasmic domains was able to assemble significant numbers of fibrils (Fig. 7H,I). A few short fibrils were also frequently observed on the BCR injected with the α4α3 transcript (Fig. 7G). No fibrils were observed on animal caps expressing α4αv (Fig. 7J). To further confirm that the observed FN fibril assembly by α4 chimeras was independent of interactions of endogenous integrins with the RGD site of FN, we repeated the above experiments with mAb 4B12, which blocks the activity of the RGD motif in the 10th type III repeat of Xenopus FN (Marsden and DeSimone, 2001; Ramos and DeSimone, 1996). Similar results were obtained with this antibody, thus we conclude that the α4 chimera-dependent rescues of fibril assembly occur in the absence of any contribution of endogenous α5β1 binding to FN (data not shown). To exclude the possibility that the observed differences in fibril assembly by various α4 constructs were due to differences in cell-surface expression

2340

Journal of Cell Science 116 (11)

Fig. 7. Only some α cytoplasmic domains are able to support FN fibrillogenesis. (A) FN fibril assembly rescue assay. α4 Constructs were injected into the animal pole regions of embryos at the two-cell stage. At stage 9, mAb P8D4 were injected into the blastocoel of these embryos. The embryos were cultured until the end of gastrulation. Then animal caps were dissected and subjected to immunostaining for FN. (B-J) Immunofluorescence of FN on the blastocoel roof. The α4 construct expressed is indicated on each panel. Endogenous FN fibril network is formed on the blastocoel roof at the end of gastrulation (B). P8D4 blocked fibril formation (C). In the presence of P8D4, α4 chimeras with the α5 and α6 cytoplasmic domains were able to mediate the assembly of significant amount of FN fibrils (H,I). A few fibrils were also observed on the BCR injected with α4α3 (G). The mAb 4H2 was used to immunostain FN.

Table 3. Summary of cell adhesive functions mediated by α4 constructs Cytoplasmic tail Animal cap cell attachment on V-GST Activin-induced animal cap cell spreading and migration on V-GST DMZ explant migration on V-GST FN fibril assembly on the BCR

KV

RR

α2

α3

α4

α5

α6

αv

+ – – –

+ + – –

+ + + –

+ + + +

+ + + –

+ + + ++

+ + + +++

+ + + –

+ indicates that the cytoplasmic tail is able to support a certain adhesive function; – means it is incapable of supporting an adhesive function.

levels, α4 transcripts were injected over a range of concentrations. Differences in cell-surface expression were confirmed by biotinylation and immunoprecipitation with the A4EX PcAb. When the amount of transcript injected was varied within a fourfold range, the same results were obtained (data not shown). Thus, we can conclude that the α5, α6 and α3 cytoplasmic tails enable an RGD-independent fibrillogenesis by the integrin α4 extracellular domain, whereas the α4wt, α4αv and α4 cytoplasmic tail truncation mutants are unable to mediate fibril assembly. Discussion In this study, we show the functional significance of integrin α subunit cytoplasmic domains in supporting position-specific adhesive behaviors in Xenopus gastrulae. An α subunit tail is not necessary for the basic ligand-binding properties of integrin α4β1 in Xenopus embryonic cells because the tail deletion mutant that lacks the GFFK/RR domain and sequences Cterminal to it is still able to support cell attachment to fusion proteins that contain the V-region. Retaining only the

GFFK/RR motif within the cytoplasmic tail further restores the ability to mediate cell spreading in response to activin induction. This confirms the importance of this highly conserved motif to integrin adhesive function, as shown in many cell culture studies (Kassner et al., 1994; Kawaguchi et al., 1994; Kawaguchi and Hemler, 1993; Vossmeyer et al., 2000). However, cells expressing α4RR still displayed significantly reduced motility and failed to assemble FN fibrils. Thus, the sequences following the GFFK/RR motif are needed to specify appropriate receptor activation states and related adhesive behaviors. Moreover, our results also show that the post-GFFK/RR sequences of various α tails are able to differentially regulate the migration of mesodermal cell and tissue explants, and FN fibrillogenesis on the BCR. These results are summarized in Table 3. Integrin α cytoplasmic domains and mesodermal cell migration Integrin α4 chimeras with a full-length cytoplasmic domain are each able to support the migration of mesoderm cells

Integrin α cytoplasmic tail functions and tissue explants. This suggests that α subunit tails play a general positive role in promoting cell migration. Quantitative analysis of mesoderm cell and tissue explant migration reveals that α4α5 has a slightly reduced migration ability compared with α4 chimeras containing the α6, α4 and αv cytoplasmic domains. Cell migration is a multistep dynamic process involving cell attachment and subsequent detachment from the substrate (Lauffenburger and Horwitz, 1996). Migration speed is directly related to receptor and ligand concentrations, as well as the affinity of the receptor for its ligand (Palecek et al., 1997). Maximal migration speed is achieved at an intermediate force:adhesiveness ratio (DiMilla et al., 1993; Palecek et al., 1997). In our experiments, the substrate concentration and the amount of receptor expressed on the cell surface were normalized. Thus, it is possible that the α5 tail confers the α4 extracellular domain with an activation state not optimized for rapid cell migration. High-affinity ligand binding may delay cell detachment from the substrate, hence resulting in a reduction in the migration speed. The α5 cytoplasmic domain is reported to confer integrin receptors with ‘high-affinity’ binding in a variety of contexts (O’Toole et al., 1994; Wu et al., 1995b). Moreover, an α2 chimera with an α5 tail is reported to support strong collagen gel contraction, but not enhanced migration (Chan et al., 1992). α Cytoplasmic domains have different abilities to support FN fibril assembly In this study, we show that α4 chimeras with α5, α6 or α3 cytoplasmic tails, but not α4 or αv tails, can mediate RGDindependent FN fibril assembly in vivo during Xenopus gastrulation. Fibrillogenesis is important to many biological processes including development, wound healing and tumorigenesis; thus, mechanisms of matrix assembly are of considerable interest. Integrin activation and cytoskeletal interaction have been shown to be essential for FN matrix assembly (Wu et al., 1995b). Integrin α4 is reported to exist in multiple activation states (Masumoto and Hemler, 1993). In the case of FN fibrillogenesis, it cannot initiate matrix assembly unless exogenous stimuli such as Mn2+ or ‘activating’ antibodies are present (Sechler et al., 2001; Wu et al., 1995a). In our experiments, it is highly likely that the α5, α6 and α3 tails are able to confer the α4 extracellular domain with the appropriate conformation needed to promote FN matrix assembly on the BCR, whereas the α4 and αv tails can not. In support of this hypothesis, O’Toole et al. (O’Toole et al., 1994) have shown that the α5, α6A, α6B and α2 tails are all able to confer an energy-dependent high-affinity state to the αIIb extracellular domain in CHO cells, whereas αIIb, αv, αL and αM tails lack such activities. Thus, α cytoplasmic domainspecific inside-out signaling is probably essential for FN matrix assembly on the BCR during Xenopus gastrulation. This may help to explain why the matrix is preferentially deposited on the BCR and not the blastocoel floor (although cells in both regions express α5β1). Cytoskeletal tension is also crucial for FN fibrillogenesis. Integrin α tails can directly interact with some cytoplasmic scaffolding proteins. For example, the α4 tail associates with paxillin (Liu et al., 1999), and the α5 tail has been show to bind Nischarin, a novel protein that can affect Rho GTPase

2341

function (Alahari et al., 2000). These α tail-specific interactions probably result in differential composition and/or dynamics of adhesive contacts organized by these integrins. Ultimately, this may affect the organization of the cytoskeleton. It has been proposed that the α5 tail preferentially regulates prolonged force, which can help to ‘pull open’ the coiled FN dimer and expose cryptic sites that promote FN self-assembly (Chan et al., 1992; Schwarzbauer and Sechler, 1999). Thus, it is possible that, aside from defining integrin activation states, α cytoplasmic tails can also differentially modulate cytoskeletal interactions that are crucial for the preferential assembly of FN matrix. Other possible roles of integrin α cytoplasmic domains Aside from their signaling functions, integrin α subunit cytoplasmic domains may also play a role in controlling integrin trafficking at the cell surface. When the α2 cytoplasmic tail is joined to the α4 extracellular and transmembrane domains, most of the protein synthesized is retained intracellularly, even when high concentrations of α4α2 mRNA are injected. Biochemical analyses suggest that the α4α2 chimeric protein is translated in proportion to the amount of RNA injected (Fig. 1C) but is not efficiently mobilized to the cell surface (Fig. 1B and data not shown). The α4 chimeras with other α cytoplasmic tails are expressed on the cell surface approximately as efficiently as the two cytoplasmic tail truncation mutants. Thus, the sequences after the GFFKR motif within the α2 cytoplasmic domain probably account for the low cell-surface expression of the α4α2 chimera. During Xenopus development, α2 mRNA is detected at late gastrula/neurula stages (Whittaker and DeSimone, 1993) but the α2 protein is not expressed at the cell surface until much later in development (Meng et al., 1997). Our findings suggest the existence of a trafficking checkpoint in embryonic cells, which may limit the transit of some integrins to the cell surface. Such checkpoint machinery may recognize information within specific α cytoplasmic tail sequences. Conclusions The cytoplasmic domains of integrin α subunits are generally short with diverse sequences C-terminal to the conserved membrane proximal GFFK/RR motif (Liu et al., 2000; Sastry and Horwitz, 1993). They are thought to impart unique signaling properties to the receptor. However, to our knowledge, no studies have been conducted to investigate the functions of α cytoplasmic tails in vertebrate embryos. Research on Drosophila integrins suggests that the cytoplasmic tails of αPS1 and αPS2 are functionally indistinguishable (Martin-Bermudo et al., 1997). In this study, we provide evidence that integrin α subunit cytoplasmic domains can contribute to the modulation of integrin functions in gastrulating Xenopus embryos. We also show that, in some cases, the α subunit tail can confer specificity to select cell adhesive functions, such as the assembly of FN fibrils. Future research will be needed to identify the nature of the positionspecific intracellular signals that regulate integrins in vivo and to identify sequence motifs within α cytoplasmic tails that are involved in specifying the distinct functions carried out by these adhesion molecules.

2342

Journal of Cell Science 116 (11)

The authors are very grateful to Lance Davidson, Betty Dzamba, Dominique Alfandari, Robert Bloodgood, Ann Sutherland, Rick Horwitz, Judith White, Theresa Curtis, Katherine Smith and Benjamin Hoffstrom for helpful discussions and technical support throughout this research. We also thank Charles Whittaker for preparing the α4KV and α4RR cDNA constructs. This work was supported by United States Public Health Service grant #HD26402.

References Adams, J. C. and Watt, F. M. (1993). Regulation of development and differentiation by the extracellular matrix. Development 117, 1183-1198. Alahari, S. K., Lee, J. W. and Juliano, R. L. (2000). Nischarin, a novel protein that interacts with the integrin alpha5 subunit and inhibits cell migration. J. Cell Biol. 151, 1141-1154. Alfandari, D., Whittaker, C. A., DeSimone, D. W. and Darribere, T. (1995). Integrin alpha v subunit is expressed on mesodermal cell surfaces during amphibian gastrulation. Dev. Biol. 170, 249-261. Brown, N. H., Gregory, S. L. and Martin-Bermudo, M. D. (2000). Integrins as mediators of morphogenesis in Drosophila. Dev. Biol. 223, 1-16. Chan, B. M., Kassner, P. D., Schiro, J. A., Byers, H. R., Kupper, T. S. and Hemler, M. E. (1992). Distinct cellular functions mediated by different VLA integrin alpha subunit cytoplasmic domains. Cell 68, 1051-1060. Davidson, L. A., Hoffstrom, B. G., Keller, R. and DeSimone, D. W. (2002). Mesendoderm extension and mantle closure in Xenopus laevis gastrulation: combined roles for integrin alpha(5)beta(1), fibronectin, and tissue geometry. Dev. Biol. 242, 109-129. Dedhar, S. and Hannigan, G. E. (1996). Integrin cytoplasmic interactions and bidirectional transmembrane signalling. Curr. Opin. Cell Biol. 8, 657-669. DiMilla, P. A., Stone, J. A., Quinn, J. A., Albelda, S. M. and Lauffenburger, D. A. (1993). Maximal migration of human smooth muscle cells on fibronectin and type IV collagen occurs at an intermediate attachment strength. J. Cell Biol. 122, 729-737. Dzamba, B. J., Bolton, M., A. and DeSimone, D. W. (2002). The integrin family of cell adhesion molecules. In Cell Adhesion: Frontiers in Molecular Biology (ed. M. C. Beckerle), pp. 100-154. Oxford: Oxford University Press. Gawantka, V., Ellinger-Ziegelbauer, H. and Hausen, P. (1992). Beta 1integrin is a maternal protein that is inserted into all newly formed plasma membranes during early Xenopus embryogenesis. Development 115, 595605. Ginsberg, M. H., Du, X. and Plow, E. F. (1992). Inside-out integrin signalling. Curr. Opin. Cell Biol. 4, 766-771. Guan, K. L. and Dixon, J. E. (1991). Eukaryotic proteins expressed in Escherichia coli: an improved thrombin cleavage and purification procedure of fusion proteins with glutathione S-transferase. Anal. Biochem. 192, 262267. Gurdon, J. B. (1977). Methods for nuclear transplantation in amphibia. Methods Cell Biol. 16, 125-139. Hemler, M. E., Elices, M. J., Parker, C. and Takada, Y. (1990). Structure of the integrin VLA-4 and its cell-cell and cell-matrix adhesion functions. Immunol. Rev. 114, 45-65. Hemler, M. E., Kassner, P. D. and Chan, B. M. (1992). Functional roles for integrin alpha subunit cytoplasmic domains. Cold Spring Harb. Symp. Quant. Biol. 57, 213-220. Hughes, P. E. and Pfaff, M. (1998). Integrin affinity modulation. Trends Cell Biol. 8, 359-364. Hughes, P. E., Diaz-Gonzalez, F., Leong, L., Wu, C., McDonald, J. A., Shattil, S. J. and Ginsberg, M. H. (1996). Breaking the integrin hinge. A defined structural constraint regulates integrin signaling. J. Biol. Chem. 271, 6571-6574. Hynes, R. O. (1992). Integrins: versatility, modulation, and signaling in cell adhesion. Cell 69, 11-25. Ivaska, J., Reunanen, H., Westermarck, J., Koivisto, L., Kahari, V. M. and Heino, J. (1999). Integrin alpha2beta1 mediates isoform-specific activation of p38 and upregulation of collagen gene transcription by a mechanism involving the alpha2 cytoplasmic tail. J. Cell Biol. 147, 401-416. Joos, T. O., Whittaker, C. A., Meng, F., DeSimone, D. W., Gnau, V. and Hausen, P. (1995). Integrin alpha 5 during early development of Xenopus laevis. Mech. Dev. 50, 187-199. Joos, T. O., Reintsch, W. E., Brinker, A., Klein, C. and Hausen, P. (1998). Cloning of the Xenopus integrin alpha(v) subunit and analysis of its distribution during early development. Int. J. Dev. Biol. 42, 171-179.

Kassner, P. D. and Hemler, M. E. (1993). Interchangeable alpha chain cytoplasmic domains play a positive role in control of cell adhesion mediated by VLA-4, a beta 1 integrin. J. Exp. Med. 178, 649-660. Kassner, P. D., Kawaguchi, S. and Hemler, M. E. (1994). Minimum alpha chain cytoplasmic tail sequence needed to support integrin-mediated adhesion. J. Biol. Chem. 269, 19859-19867. Kawaguchi, S. and Hemler, M. E. (1993). Role of the alpha subunit cytoplasmic domain in regulation of adhesive activity mediated by the integrin VLA-2. J. Biol. Chem. 268, 16279-16285. Kawaguchi, S., Bergelson, J. M., Finberg, R. W. and Hemler, M. E. (1994). Integrin alpha 2 cytoplasmic domain deletion effects: loss of adhesive activity parallels ligand-independent recruitment into focal adhesions. Mol. Biol. Cell 5, 977-988. Koff, A., Giordano, A., Desai, D., Yamashita, K., Harper, J. W., Elledge, S., Nishimoto, T., Morgan, D. O., Franza, B. R. and Roberts, J. M. (1992). Formation and activation of a cyclin E-cdk2 complex during the G1 phase of the human cell cycle. Science 257, 1689-1694. Lauffenburger, D. A. and Horwitz, A. F. (1996). Cell migration: a physically integrated molecular process. Cell 84, 359-369. Liu, S. and Ginsberg, M. H. (2000). Paxillin binding to a conserved sequence motif in the α4 integrin cytoplasmic domain. J. Biol. Chem. 275, 2273622742. Liu, S., Thomas, S. M., Woodside, D. G., Rose, D. M., Kiosses, W. B., Pfaff, M. and Ginsberg, M. H. (1999). Binding of paxillin to alpha4 integrins modifies integrin-dependent biological responses. Nature 402, 676-681. Liu, S., Calderwood, D. A. and Ginsberg, M. H. (2000). Integrin cytoplasmic domain-binding proteins. J. Cell Sci. 113, 3563-3571. Marsden, M. and DeSimone, D. W. (2001). Regulation of cell polarity, radial intercalation and epiboly in Xenopus: novel roles for integrin and fibronectin. Development 128, 3635-3647. Martin-Bermudo, M. D., Dunin-Borkowski, O. M. and Brown, N. H. (1997). Specificity of PS integrin function during embryogenesis resides in the alpha subunit extracellular domain. EMBO J. 16, 4184-4193. Masumoto, A. and Hemler, M. E. (1993). Multiple activation states of VLA4. Mechanistic differences between adhesion to CS1/fibronectin and to vascular cell adhesion molecule-1. J. Biol. Chem. 268, 228-234. Meng, F., Whittaker, C. A., Ransom, D. G. and DeSimone, D. W. (1997). Cloning and characterization of cDNAs encoding the integrin alpha2 and alpha3 subunits from Xenopus laevis. Mech. Dev. 67, 141-155. Newport, J. and Kirschner, M. (1982). A major developmental transition in early Xenopus embryos: II. Control of the onset of transcription. Cell 30, 687-696. Nieuwkoop, P. and Faber, J. (1967). Normal Table of Xenopus laevis (Daudin). Amsterdam: North-Holland Publishing Company. O’Toole, T. E., Mandelman, D., Forsyth, J., Shattil, S. J., Plow, E. F. and Ginsberg, M. H. (1991). Modulation of the affinity of integrin alpha IIb beta 3 (GPIIb-IIIa) by the cytoplasmic domain of alpha IIb. Science 254, 845-847. O’Toole, T. E., Katagiri, Y., Faull, R. J., Peter, K., Tamura, R., Quaranta, V., Loftus, J. C., Shattil, S. J. and Ginsberg, M. H. (1994). Integrin cytoplasmic domains mediate inside-out signal transduction. J. Cell Biol. 124, 1047-1059. Palecek, S. P., Loftus, J. C., Ginsberg, M. H., Lauffenburger, D. A. and Horwitz, A. F. (1997). Integrin-ligand binding properties govern cell migration speed through cell-substratum adhesiveness. Nature 385, 537540. Ramos, J. W. and DeSimone, D. W. (1996). Xenopus embryonic cell adhesion to fibronectin: position-specific activation of RGD/synergy site-dependent migratory behavior at gastrulation. J. Cell Biol. 134, 227-240. Ramos, J. W., Whittaker, C. A. and DeSimone, D. W. (1996). Integrindependent adhesive activity is spatially controlled by inductive signals at gastrulation. Development 122, 2873-2883. Sastry, S. K. and Horwitz, A. F. (1993). Integrin cytoplasmic domains: mediators of cytoskeletal linkages and extra- and intracellular initiated transmembrane signaling. Curr. Opin. Cell Biol. 5, 819-831. Sastry, S. K., Lakonishok, M., Thomas, D. A., Muschler, J. and Horwitz, A. F. (1996). Integrin alpha subunit ratios, cytoplasmic domains, and growth factor synergy regulate muscle proliferation and differentiation. J. Cell Biol. 133, 169-184. Sastry, S. K., Lakonishok, M., Wu, S., Truong, T. Q., Huttenlocher, A., Turner, C. E. and Horwitz, A. F. (1999). Quantitative changes in integrin and focal adhesion signaling regulate myoblast cell cycle withdrawal. J. Cell Biol. 144, 1295-1309.

Integrin α cytoplasmic tail functions Schwarzbauer, J. E. and Sechler, J. L. (1999). Fibronectin fibrillogenesis: a paradigm for extracellular matrix assembly. Curr. Opin. Cell Biol. 11, 622627. Sechler, J. L., Cumiskey, A. M., Gazzola, D. M. and Schwarzbauer, J. E. (2000). A novel RGD-independent fibronectin assembly pathway initiated by alpha4beta1 integrin binding to the alternatively spliced V region. J. Cell Sci. 113, 1491-1498. Sechler, J. L., Rao, H., Cumiskey, A. M., Vega-Colon, I., Smith, M. S., Murata, T. and Schwarzbauer, J. E. (2001). A novel fibronectin binding site required for fibronectin fibril growth during matrix assembly. J. Cell Biol. 154, 1081-1088. Shattil, S. J. and Brass, L. F. (1987). Induction of the fibrinogen receptor on human platelets by intracellular mediators. J. Biol. Chem. 262, 992-1000. Shattil, S. J., Ginsberg, M. H. and Brugge, J. S. (1994). Adhesive signaling in platelets. Curr. Opin. Cell Biol. 6, 695-704. Smith, J. C., Symes, K., Hynes, R. O. and DeSimone, D. (1990). Mesoderm induction and the control of gastrulation in Xenopus laevis: the roles of fibronectin and integrins. Development 108, 229-238. Vossmeyer, D., Kaufmann, C., Loster, K., Lucka, L., Horstkorte, R., Reutter, W. and Danker, K. (2000). The cytoplasmic domain of the alpha1 integrin subunit influences stress fiber formation via the conserved GFFKR motif. Exp. Cell Res. 256, 321-327.

2343

Weber, K. S., Klickstein, L. B. and Weber, C. (1999). Specific activation of leukocyte beta2 integrins lymphocyte function-associated antigen-1 and Mac-1 by chemokines mediated by distinct pathways via the alpha subunit cytoplasmic domains. Mol. Biol. Cell 10, 861-873. Whittaker, C. A. and DeSimone, D. W. (1993). Integrin alpha subunit mRNAs are differentially expressed in early Xenopus embryos. Development 117, 1239-1249. Whittaker, C. A. and DeSimone, D. W. (1998). Molecular cloning and developmental expression of the Xenopus homolog of integrin alpha 4. Ann. N. Y. Acad. Sci. 857, 56-73. Winklbauer, R. (1990). Mesodermal cell migration during Xenopus gastrulation. Dev. Biol. 142, 155-168. Winklbauer, R. and Nagel, M. (1991). Directional mesoderm cell migration in the Xenopus gastrula. Dev. Biol. 148, 573-589. Winklbauer, R. and Stoltz, C. (1995). Fibronectin fibril growth in the extracellular matrix of the Xenopus embryo. J. Cell Sci. 108, 1575-1586. Wu, C., Fields, A. J., Kapteijn, B. A. and McDonald, J. A. (1995a). The role of alpha 4 beta 1 integrin in cell motility and fibronectin matrix assembly. J. Cell Sci. 108, 821-829. Wu, C., Keivens, V. M., O’Toole, T. E., McDonald, J. A. and Ginsberg, M. H. (1995b). Integrin activation and cytoskeletal interaction are essential for the assembly of a fibronectin matrix. Cell 83, 715-724.