Involvement of the adapter protein CRKL in integrin-mediated adhesion


CRKL, an SH2-SH3-SH3 adapter protein, is one of the major tyrosine phosphoproteins detected in primary leukemic neutrophils from patients with CML. CRKL binds directly to BCR/ABL through its N-terminal SH3 domain, suggesting it may be involved in BCR/ABL signal transduction. However, the biological function of CRKL in either normal or leukemic cells is still largely unknown. In this study, we have examined the effects of overexpressing full length or deletion mutants of CRKL in hematopoietic cell lines. Full length, SH2- and SH3(N)-domain deletion mutants of CRKL were transfected into an interleukin-3-dependent hematopoietic cell line, Ba/F3, and 3 – 5 individual sublines which stably overexpressed each transgene were obtained [Ba/F-CRKL, Ba/F-CRKLΔSH2, and Ba/F-CRKL ΔSH3(N)]. The growth properties of these transfected cells in the presence or absence of IL-3 were not different from mock transfected or untransfected Ba/F3 cells. However, Ba/F3 cells overexpressing full length CRKL, but not deletion mutants of CRKL, were found to have an increase in their ability to bind to fibronectin-coated surfaces. Further, expression of full length, but not ΔSH2- or ΔSH3-CRKL deletion mutants, was found to alter cell morphology on fibronectin-coated plates, an effect which was further enhanced by certain kinds of stress stimuli, such as ionizing radiation. Similar results were obtained when CRKL was transiently overexpressed in Ba/F3 cells, and were also obtained in a second IL-3 dependent hematopoietic cell line, 32Dcl3. Adhesion to fibronectin was blocked by anti-β1 integrin monoclonal antibody, but overexpression of CRKL did not affect surface expression of β1 integrins, nor did it spontaneously induce expression of the β1 integrin `activation' epitope recognized by the 9EG7 monoclonal antibody. These data suggest a role for CRKL in signaling pathways which regulate adhesion to fibronectin.


CRKL is a 39 kDa adapter protein containing one SH2 and two SH3 domains. The CRKL gene was originally identified by ten Hoeve et al. (1993) and found to be in close proximity to the BCR gene on human chromosome 22q11. The CRKL protein has an overall amino acid homology of 60% to CRK II, one of two proteins generated through alternative splicing of the human crk proto-oncogene. CRK I lacks the C-terminal SH3 domain of CRK II, but overexpression of either CRK I or CRKL, but not CRK II, leads to transformation of fibroblasts (Matsuda et al., 1992; Senechal et al., 1996). CRK was identified by its homology to v-crk, the oncogene in the avian retrovirus CT10, which differs from CRK by deletion of the C-terminal SH3 domain and the major tyrosine phosphorylation site at tyr221 (Mayer et al., 1988).

CRKL and/or CRK have been shown to be involved in a number of signaling pathways, including pathways activated by T cell stimulation (Sawasdikosol et al., 1995), EGF receptor activation (Birge et al., 1992), and hematopoietic cytokine receptor activation (Barber et al., 1997; Sattler et al., 1997a). Further, we have previously shown that CRKL is tyrosine phosphorylated after integrin cross-linking (Sattler et al., 1997b). However, the exact roles of either CRK and CRKL in any of these normal signaling pathways are unknown.

A number of cellular proteins have been shown to bind to the CRKL and CRK SH3 domains, including c-ABL, C3G, EPS15, DOCK180, and SOS (Feller et al., 1994; Ren et al., 1994; ten Hoeve et al., 1994b; Tanaka et al., 1994; Uemura et al., 1997; Schumacher et al., 1995; Hasegawa et al., 1996; Feller et al., 1995). Several proteins have also been identified which interact with the CRKL or CRK SH2 domains, including c-CBL, CAS, HEF1, and paxillin (Sawasdikosol et al., 1995; de Jong et al., 1995; Sattler et al., 1996, 1997b; Sakai et al., 1994; Salgia et al., 1996a,b; Birge et al., 1993). In vitro, the proteins which can interact with CRKL and CRK appear to be similar, although a few differences have been noted. Immunoprecipitation studies suggest that CRKL and CRK bind constitutively to proteins such as C3G which bind to their SH3 domains, while interaction with the SH2-binding proteins requires tyrosine phosphorylation of the target protein, typically involving a pY-X-X-P motif (Songyang et al., 1993; Salgia et al., 1996b; Andoniou et al., 1996).

CRKL is a prominent tyrosine phosphoprotein in cells transformed by the BCR/ABL oncogene, and we and others have previously shown that CRKL binds to BCR/ABL through its N-terminal SH3 domain (Oda et al., 1994; Nichols et al., 1994; ten Hoeve et al., 1994a). In these transformed cells, CRKL links BCR/ABL to p120c-CBL, which is also heavily phosphorylated on tyrosine residues. PI3K joins this complex through binding to CBL, thus forming a multimeric complex with BCR/ABL (Sattler et al., 1996). The site of tyrosine phosphorylation of CRKL in BCR/ABL-tranformed cells has been identified as Y207 (de Jong et al., 1997). The consequences of CRKL phosphorylation are unknown, although phosphorylation of the homologous tyrosine residue in CRK has been reported to form an intramolecular binding site for the CRK SH2 domain (Feller et al., 1994; Rosen et al., 1995).

In the present study, we sought to investigate the biological consequences of CRKL signaling by overexpressing CRKL or CRKL mutants in growth factor-dependent hematopoietic cell lines. The results indicate that CRKL may be involved in pathways which regulate adhesion to extracellular matrix proteins such as fibronectin. Interestingly, recent studies suggest that v-crk may have have a similar effect on adhesion (Altun-Gultekin et al., 1998).


Biological effects of overexpression of wild-type and deletion mutants of CRKL in hematopoietic cell lines

The IL-3 dependent hematopoietic cell line Ba/F3 was transfected with vectors expressing either wild-type CRKL, CRKLΔSH2 or CRKLΔSH3(N) cDNAs. Similarly, IL-3 dependent 32Dcl3 cells were transfected with either wild-type CRKL or CRKLΔSH3(N) cDNAs (Figure 1a). After selection in G418 and screening by Western blotting, 3 – 5 independent subclones which overexpressed each transgene were obtained. The expression of CRKL proteins in two representative clones of each line is shown in Figure 1b. Each transgene was typically expressed at a 5 – 10-fold increase over that of endogenous CRKL. Another monoclonal antibody, anti-CRKL #2-2, which recognizes an epitope located within the N-terminal SH3 domain (Uemura et al., 1997), detected equivalent overexpression of the wild-type and ΔSH2 CRKL mutant proteins, but but did not detect the ΔSH3(N) mutant CRKL protein, confirming expression of the appropriate transgene in the ΔSH3 mutants (data not shown). The two clones shown in Figure 1 of each line were those evaluated in all subsequent experiments.

Figure 1

Expression of wild-type and mutant CRKL proteins in Ba/F3 and 32Dcl3 cells. (a) Diagrams of structure of wild-type and mutant CRKL proteins. All constructs were tagged at the C-terminus with an HA tag. (b) Western blotting with anti-CRKL mAb (#5 – 6) of whole cell lysates (1×106 cells) from two representative clones of each line and mock-transfected cells. CRKL with HA-tag migrated at ≈40 kDa and CRKL-ΔSH2 or CRKL-ΔSH3(N) with HA-tag migrated at ≈34 kDa. Endogenous CRKL migrated at 39 kDa. (c) Lysates from 1×107 cells of a representative clone of each line and mock-transfected cells were immunoprecipitated with anti-CRKL mAb. Immunoprecipitates and whole cell lysates from 1×106 cells were subjected to Western blotting with antibodies against several CRKL-SH3 binding proteins (C3G, ABL, and p85P13K) and CRKL

The interaction of overexpressed CRKL proteins with cellular proteins was examined by looking for coprecipitation of specific cellular proteins in anti-CRKL immune complexes. Coprecipitation of CRKL with three proteins known to bind to CRKL-SH3 domains, C3G, ABL and the p85 subunit of PI3K, was determined. Coprecipitation of CRKL with C3G, ABL and p85PI3K was increased in Ba/F3 overexpressing wild-type CRKL and the CRKLΔSH2 as compared to mock transfected cells, but not in cells overexpressing a ΔSH3(N) mutant of CRKL (Figure 1c). Similar results were seen in 32Dcl3 cells. The coprecipitation of small amounts of C3G and p85Pl3K with CRKL in cells overexpressing CRKLΔSH3 was attributed to the presence of endogenous CRKL in the immune complex. We also looked for CRKL-SH2 binding proteins in Ba/F3 cells, but found only trace amounts of tyrosine phosphoproteins in anti-CRKL immunoprecipitations in these cells, consistent with our previous findings in IL-3-stimulated Ba/F3 cells using CRKL fusion proteins as probes in far Western blots (Uemura et al., 1997). However, Chin et al. (1997) reported that several proteins including CBL are coprecipitated with CRKL after IL-3 stimulation in another cell line, 32D.

All lines overexpressing wild-type CRKL or CRKL deletion mutants were analysed to determine if these proteins affected IL-3 dependence, growth rates, or viability. All lines grew at similar rates in the presence of IL-3, and died within two days in the absence of IL-3 (data not shown). The Ba/F3 cell lines were also tested for radiation sensitivity by serially measuring viability after 10 – 40 Gy ionizing radiation. Again, overexpression of CRKL or the CRKL deletion mutants tested here had no demonstrable effects on radiation-induced apoptosis (data not shown).

Effects of CRKL on cell adhesion to fibronectin

CRKL is known to bind to several focal adhesion proteins, including paxillin, thought to be involved in regulating integrin signaling or function. Therefore, we next examined cell adhesion of these lines to fibronectin-coated surfaces using a short-term (30 min) adhesion assay. Cell lines were in log phase of growth in IL-3 containing medium when tested. Adhesion of Ba/F-CRKL cells was significantly increased as compared to Ba/F-mock cells. In contrast, overexpression of either CRKLΔSH3(N) or CRKLΔSH2 mutants in Ba/F3 cells did not lead to increased adhesion in multiple assays, and cells overexpressing CRKLΔSH3(N) had somewhat reduced adhesion compared to mock transfected cells (Figure 2a). Similar results were obtained on each of two independent clones of each transfectant, as shown in Figure 2a, where each experiment was repeated six times. In studies not shown, similar results were also obtained with these cell lines either with or without prior IL-3 deprivation.

Figure 2

Overexpression of CRKL, but not CRKL-ΔSH2 or CRKL-ΔSH3N, increases cell adhesion to fibronectin. (a, b) Two representative clones of Ba/F3 (a) or 32Dcl3 (b) were subjected to short-term adhesion assays as described in Materials and methods. (c) Transient expression of CRKL alters adhesion in Ba/F3 cells. Ba/F3 cells were transfected with pE-IRES-EGFP, pE-CRKL-IRES-EGFP or pE-CRKLΔSH3(N)-IRES-EGFP. Twenty-four hours after transfection, GFP-expressing cells were selected by flow cytometric cell sorting. Short-term adhesion to fibronectin coated surfaces of GFP-positive cells (more than 95% of cells were positive) was measured. Shown are results from three independent transfections. All results are mean±s.d. of triplicate wells. (d) Adhesion of Ba/F-mock and two representative clones of BaF-CRKL was determined in a short term adhesion assay in the presence or absence of an anti-β1 integrin antibody (Ha2/5) which can inhibit β1 integrin function. Ha2/5 or hamster lgM (as a control) was added to cell suspensions for 30 min at 37°C prior to the adhesion assay

Adhesion assays were also performed on surfaces coated with other extracellular matrix proteins (each at 5 μg/cm2), including collagen I, collagen IV, laminin, and vitronectin. The Ba/F3 derived lines did not adhere to collagen I, collagen IV, or laminin at all. However, short-term adhesion of Ba/F3 derived lines to vitronectin was similarly enhanced by overexpression of wild-type CRKL, but not by overexpression of the ΔSH3(N) mutant of CRKL (data not shown). The effects of CRKL overexpression on cell adhesion to fibronectin were also examined using 32Dcl3 cells (Figure 2b). Wild-type CRKL enhanced adhesion to fibronectin, while overexpression of CRKLΔSH3(N) had no effect or was modestly inhibitory.

The effects of CRKL overexpression were also evaluated in Ba/F3 cells after transient transfection of a construct in which CRKL and green fluorescent protein (GFP) were expressed as separate proteins through a single transcript so that transfected cells could be easily detected using fluorescence. In three separate experiments, after enriching transfected cells by fluorescence-activated cell sorting (FACS) to >95% purity, Ba/F3 cells which expressed wild-type CRKL and GFP had increased short-term adhesion to fibronectin when compared to equally brightly fluorescent Ba/F3 cells transfected with the same vector expressing CRKLΔSH3(N) or GFP alone (Figure 2c).

Since adhesion to extracellular matrix proteins such as fibronectin is mediated by integrins, we tested the effects of adding an anti-β1 integrin antibody to the adhesion assays. As expected, this antibody blocked binding of Ba/F-mock as well as Ba/F-CRKL clones by >90% (Figure 2d). The possibility that CRKL overexpression augmented β1-integrin surface expression was therefore considered, but FACS analysis with an anti-β1-integrin monoclonal antibody showed no detectable effect (Figure 3a,b). Finally, we considered the possibility that CRKL overexpression induced β integrins to enter an activated state. We therefore looked for spontaneous induction of the activation epitope identified by the 9EG7 monoclonal antibody. However, as shown in Figure 3c, overexpression of CRKL did not affect either spontaneous or ligand-induced expression of this epitope, indicating that the effects of CRKL on adhesion are not likely mediated by directly increasing integrin affinity for ligand, but modifying post-integrin receptor events.

Figure 3

Overexpression of CRKL does not induce expression of total β1 integrins nor the 9EG7 epitope on β1 integrins. (a, b) Total β1 integrin surface expression of two representative clones of Ba/F3 (a) or 32Dcl3 (b) was measured using an anti-β1 integrin antibody, HMβ1-1 using FACS. The verical axis shows mean fluorescence intensity. (c) Ba/F-mock (), Ba/F-CRKL (clones 1 and 2; ▪) and Ba/F-CRKLΔSH3N (clones 1 and 2; ) cells were analysed for expression of the 9EG7 epitope after exposure to a peptide ligand (GRGDSP; 0-1000 μM for 30 min at 37°C). Binding of the 9EG7 monoclonal antibody was dectected by FACS. Vertical axis; mean fluorescence intensity, horizontal axis; concentration of GRGDSP peptide added. The diagonal bars within the graph indicate discontinuity of the log scale to show the zero value. The data shown is a representative experiment of a total of two. To confirm that the 9EG7 antibody recognizes a conformational change of `activated' β1 integrin, expression of the HMβ1-1 epitope and the 9EG7 epitope was compared before and after the treatment with manganese (MnCl2, 3 μM, 30 min) which is known to strongly induce a conformational change of β1 integrin. The 9EG7 epitope increased after manganese treatment [mean fluorescence intensity increased from 0.77 – 1.53 (log scale, arbitrary units)], while the HMβ1-1 epitope did not change (mean fluorescence intensity: 3.01 vs. 3.11)

CRKL overexpression alters cell morphology

In the experiments described above, we noted that occasional cells overexpressing wild-type CRKL, but not CRKL deletion mutants or mock transfected cells, had a specific change in morphology characterized by the presence of long, fine, cytoplasmic protrusions. This morphological change was present in less than 1% of CRKL-transfected cells at any single point in time. Using time-lapse video microscopy on fibronectin-coated surfaces, these structures appeared to form at the trailing edge of the cell (Figure 4). These protrusions contained F-actin when stained by rhodamine-phalloidin (data not shown). The number of CRKL-overexpressing cells with these long trailing protrusions was markedly reduced when cells were cultured on plastic, rather than fibronectin, suggesting that integrin crosslinking in some way contributes to this change. In an experiment in which cells were irradiated to examine the effects of CRKL overexpression on viability after ionizing radiation, we noted that this morphological change was strikingly enhanced. In more detailed studies (Figure 5), gamma radiation increased the number of cells with this altered morphology, modestly for untransfected, mock-transfected, or CRKL deletion mutant overexpressing cells, and strikingly for all sublines of Ba/F3 overexpressing wild-type CRKL (Figure 5a). 32Dcl3 derived sublines were also evaluated for CRKL-related changes in morphology, and similar results were obtained (Figure 5b). To ensure that this phenomenon was not nonspecifically selected for during the derivation of these stable cell lines, the effects of transient expression of GFP – CRKL or GFP were evaluated in Ba/F3 cells and 32Dcl3 cells. Twenty-four hours after transfection, cells were radiated and plated on a fibronectin-coated surface, and 8 h later, morphology was observed. In three separate experiments, the transfection efficiency (% GFP positive cells) of Ba/F3 cells (25 – 65%) was higher than 32D cells (<5%), but in both cases, cells with long cytoplasmic protrusions were observed more often among GFP – CRKL transfected Ba/F3 and 32Dcl3 cells as compared with GFP-alone transfected Ba/F3 and 32Dcl3 cells, respectively (Figure 5c). To exclude the possibility that the GFP – CRKL fusion protein does not reflect the function of CRKL alone, these studies were repeated with a new vector, pE-CRKL-IRES-EGFP, in which CRKL and GFP are separately expressed, and again similar findings were obtained. These results suggest that overexpression of CRKL, either transiently or stably, alters cell morphology, and that under certain conditions of cell stress, such as gamma-irradiation, this morphological change is further enhanced. It is possible that the synergy between radiation and CRKL overexpression is due to the enhancement of cell-cell and cell-matrix adhesion known to be caused by gamma radiation (Onoda et al., 1992; Jacobson et al., 1996). We found that radiation of either Ba/F3 or 32Dcl3 cells (10 Gy) increases adhesion to fibronectin within 6 h without increase in the surface β1 integrin expression (data not shown).

Figure 4

Overexpression of CRKL induces abnormal cell motility on fibronectin. Ba/F-CRKL cells were plated onto fibronectin-coated plate, and observed using time-lapse video microscopy. The pictures shown were captured from the video showing the same field at 5 min intervals. The cell indicated by dark arrows was moving to the right side. The images show the formation of a long cellular extension forming from the trailing edge of the cell body. The smaller arrow shows a trailing cellular extension that is continuing to elongate as the cell migrates further to the right

Figure 5

Overexpression of CRKL alters cell morphology. (a) Ba/F-mock, Ba/F-CRKL, Ba/F-CRKLΔSH2, and Ba/F-CRKLΔSH3N cells before and after radiation treatment (10 Gy) were plated onto fibronectin-coated plates. Twenty hours after plating, photographs were taken with phase-contrast microscopy. Cells which formed cytoplasmic protrusions (longer than twice the length of the cell body) were counted among 1×103 cells in triplicate plates. (b) The same analysis was performed on 32D-mock, 32D-CRKL, and 32D-CRKLΔSH3N cells. (c) Morphological effects of overexpression of GFP-CRKL expressed transiently. Ba/F3 and 32Dcl3 cells were transfected GFP-CRKL fusion protein or GFP alone (as a control) expression plasmids. GFP-expressing cells were enriched to >95% fluorescent cells by FACS. Twenty-four hours after transfection, cells were radiated (10 Gy), and plated in triplicate onto fibronectin coated plates (1×103 cells/plate). In the bar graph, results from three independent transfections of Ba/F3 cells are shown as mean±s.d. of triplicate plates


In this study, we have examined the biological effects of overexpressing full length or SH2/SH3 deletion mutants of CRKL in two IL-3-dependent hematopoietic cell lines, Ba/F3 and 32Dcl3. In both lines, IL-3 is required for viability, induces proliferation, and also transiently enhances adhesion to fibronectin (Bazzoni et al., 1996). A series of Ba/F3 and 32Dcl3 cell lines stably expressing wild-type CRKL, CRKLΔSH2 or CRKLΔSH3(N) were generated, and in several independently derived sublines, overexpression of these proteins had no demonstrable effect on viability or proliferation in the presence or absence of IL-3. However, adhesion to fibronectin was enhanced by wild-type CRKL, but not by CRKL mutants lacking either the N-terminal SH3 domain or the SH2 domain. The effects of CRKL on adhesion to fibronectin were confirmed by observing similar effects in both Ba/F3 and 32Dcl3 cells after transient expression of CRKL and CRKL deletion mutants. These results suggest that CRKL is involved in a signaling pathway that regulates cellular adhesion through integrins, and further suggest that CRKL may not be a critical component of pathways leading to proliferation or control of cell viability in these cells. Recently, similar effects of CRKL overexpression on cell adhesion to fibronectin were reported using another hematopoietic cell line, FL5.12 (Senechal et al., 1998).

The mechanisms whereby CRKL might affect cell adhesion do not appear to involve altering integrin affinity. CRKL overexpression does not increase surface expression of β1 integrins and does not lead to increased expression of the activation epitope identified by the 9EG7 monoclonal antibody. Thus, CRKL could affect diffusion of integrins to sites of contact, strengthen the connection between an integrin molecule and the cytoskeleton, or alter events which lead to release of integrins from ligand. CRKL is primarily a cytoplasmic protein, although we have previously shown that CRKL can interact through its SH2 domain with several proteins located in part or in whole in the cytoskeleton, including the focal adhesion protein paxillin (Salgia et al., 1995b), and two related proteins known to be involved in integrin signaling, p130CAS and HEF1 (Salgia et al., 1996a; Sattler et al., 1997b). Paxillin is a 68 kDa phosphoprotein located exclusively in focal adhesions which is phosphorylated on serine and tyrosine residues in response to integrin cross-linking in many cell types (Bellis et al., 1997). Paxillin also interacts directly with at least two other focal adhesion proteins known to be involved in integrin signaling, focal adhesion kinase (FAK) and vinculin (Salgia et al., 1995a). p130CAS was initially identified as a prominent tyrosine phosphorylated substrate of the oncoproteins v-src and v-crk (Kanner et al., 1991; Birge et al., 1992). It has also been shown that p130CAS is tyrosine phosphorylated in response to adhesion and crosslinking of β1 integrins (Nojima et al., 1995; Manie et al., 1997). This molecule has been shown to associate with the focal adhesion proteins p125FAK and tensin in vivo and in vitro (Lo et al., 1994; Polte and Hanks, 1995). p130CAS contains an N-terminal SH3 domain which has been shown to bind to the proline rich portion in the C-terminus of p125FAK. In the middle of the molecule is the `substrate domain' with nine tyr-asp-X-pro motifs optimum for binding the c-CRK-SH2 domain (Sakai et al., 1994). Mayer and colleagues have shown that the c-Abl-SH2 domain binds to the substrate domain of p130CAS in vitro (Mayer et al., 1995). Lo et al. (1994) have shown that the SH2-domain of tensin is also capable of binding p130CAS in v-src transformed cells. Interestingly, CRK and p130CAS have recently been linked to the control of migration in an epithelial cell line (Klemke et al., 1998), suggesting that both CRK and CRKL may contribute to the regulation of several cytoskeletal functions. Overall, while the function of CRKL remains unknown, it is clear that CRKL physically associates with a number of cytoskeletal proteins which are believed to play some role in regulating adhesion. The results presented here further suggest that the amount of CRKL present in a cell can dramatically affect a pathway which enhances adhesion to fibronectin, possibly by affecting integrin function.

In addition to proteins such as paxillin, CAS, and Hef1 which interact with CRKL through its SH2 domain, there are a number of proteins known to bind to CRKL through its N-terminal SH3 domain, including C3G, c-ABL, and p85P13K. The interaction of CRKL with c-ABL may be of particular interest in integrin signaling. Lewis et al. (1996) have recently shown that cell adhesion regulates the kinase activity and subcellular localization of c-ABL. Binding of fibroblasts to fibronectin resulted in a transient recruitment of a subset of c-ABL to focal contacts coincident with the export of c-ABL from the nucleus to the cytoplasm. CRKL is known to bind to c-ABL through its SH3 domain. Interestingly, we found that overexpression of a mutant CRKL with a deletion of the N terminal SH3 domain inhibited adhesion to a moderate degree, at least in Ba/F3 cells. These results suggest that the interaction of CRKL with an SH3-binding protein, such as c-ABL, could be important in regulating adhesion to fibronectin. These results are of interest because of the prominent effects of BCR/ABL on adhesion of cells to fibronectin and other extracellular matrix proteins, and the apparent involvment of CRKL in BCR/ABL signaling.

C3G is also of potential interest in explaining the activities of CRKL overexpression. Ichiba and colleagues have previously demonstrated that expression of either CRKL or CRK in Cos1 cells significantly increased the guanine-nucleotide exchange activity of C3G for its target, Rap1 (Ichiba et al., 1997). Both the SH2 and SH3 domains of CRK were required for this activity, but CRK did not stimulate C3G activity in vitro, suggesting that this activation does not involve a simple allostearic mechanism. Interestingly, the requirement for the SH2 domain could be compensated for by the addition of a farnesylation signal to CRK, suggesting that translocation of C3G to the membrane was an important part of the mechanism. We do not have any data which suggest that this particular pathway is important for adhesion signaling, but it would be of interest to determine if overexpressing CRKL activates C3G, and thereby Rap1. Rap1, also known as smg p21 or Krev-1, was originally identified by its ability to reverse the morphologic transformation of of v-Ki-ras-transformed fibroblasts (Kitayama et al., 1989). Rap1 is likely to antagonize p21 ras function by competing for binding to ras effector molecules, such as Raf-1. Interestingly, Rap1 has been implicated in regulating cytoskeletal function and cell morphology in Dictyostelium (Rebstein et al., 1997). Also, BUD1, a homolog of rap1 in Saccharomyces cerevisiae, forms a complex with BUD5 (a homolog of GDI), BEMI, CDC24, and CDC42 and has been linked to cytoskeletal structure in yeast (White et al., 1993), and specifically with the determination of cell polarity (Park et al., 1997). Thus, activation of Rap1 could be expected to alter cytoskeletal structure in mammalian cells, and it will be worthwhile to determine if CRKL overexpression exerts its effects on adhesion through C3G activation and subsequent activation of Rap1.

The effects of CRKL overexpression on cell morphology were of interest. A small fraction of cells overexpressing CRKL, but not CRKL deletion mutants, displayed long cytoplasmic protrusions that contained actin. This morphological change occurs within 30 min on fibronectin, and is consistent, although only a small fraction of cells display this morphological change at any single point in time. Time lapse video microscopy suggested that these structures formed from the trailing edge of motile cells, and appeared to be caused by a failure to release from the extracellular matrix at an appropriate time. The observations made in stable expressing cell lines were confirmed after transient transfection of CRKL, CRKLΔSH2 or CRKLΔSH3(N) in both Ba/F3 and 32D cells. This complex change in cytoskeletal morphology was dramatically enhanced by at least one stress, radiation, in both Ba/F3 and 32D cells. Recently, it has been reported that a related protein, v-Crk, induced stress fibers, focal adhesion, lamellipodia and flattened phenotype in PC12 cells (Altun-Gultekin et al., 1998). These results suggest that overexpression of CRKL may affect cytoskeletal morphology as well as cell adhesion on a fibronectin coated surface.

Overall, the results presented suggest that although CRKL may be involved in a number of signaling pathways, in hematopoietic cells, a major biological function is likely related to regulation of adhesion, possibly involving integrin function. The identification of possible dominant negative mutant of CRKL (ΔSH3(N)) should provide a valuable tool to dissect the critical pathway involved here in more detail. Finally, the results have implications for the understanding of the myeloproliferative disorder CML, where we and others have previously shown that CRKL is a prominent substrate of BCR/ABL tyrosine kinase. Since these cells have constitutive increases in adhesion to fibronectin, it will be of great interest to determine if that effect is mediated by CRKL.

Materials and methods


A human CRKL cDNA was obtained from Dr John Groffen (Children's Hospital, Los Angeles, CA, USA). A hemagglutinin tag sequence was added before the stop codon by PCR, and the resulting cDNA was cloned into the mammalian expression vector pEBBneo. An SH2-domain deletion mutant (CRKLΔSH2) and an N-terminal SH3-domain deletion mutant (CRKLΔSH3(N)), corresponding to in-frame deletions of amino acids 14 – 64 and amino acids 131 – 179, respectively, were made by PCR. A CRKL-green fluorescent protein (GFP) fusion construct was constructed by inserting the open reading frame of CRKL into expression vector pEGFP-C (Clontech, Palo Alto, CA, USA) to create pEGFP-C-CRKL. Also, the CRKL and CRKL-ΔSH3(N) cDNA's were cloned into pE-IRES-EGFP plasmid, which was generated from pIRES-EGFP (Clontech) replacing the promoter with EF1a (Mizushima and Nagata, 1990), to create pE-CRKL-IRES-EGFP or pE-CRKLΔSH3(N)-IRES-EGFP.

Cell lines and cell culture

The murine hematopoietic cell lines, Ba/F3 and 32Dcl3, were cultured in RPMI1640 containing 10% fetal calf serum (FCS) and 15% WEHI-3B conditioned medium (WEHI-CM) as a source of IL-3. Ba/F3 cells which stably overexpress wild-type CRKL, CRKLΔSH2 and CRKLΔSH3(N) (Ba/F-CRKL, Ba/F-CRKLΔSH2, and Ba/F-CRKLΔSH3(N)) were generated by transfection of pEBBneo-CRKL, pEBBneo-CRKLΔSH2, or pEBBneo-CRKLΔSH3(N), respectively, by electroporation. Transfected cells were divided in 24-well culture plates and selected with G418 (1 mg/ml), and 3 – 5 independent subclones of each were obtained after confirming the expression of the transgene by Western blotting. Also, a polyclonal control cell line (Ba/F-mock) was generated by transfecting Ba/F3 cells with the pEBBneo vector and selecting in G418 as above. 32Dcl3 cells which stably overexpress wild-type CRKL and CRKL-ΔSH3(N) (32D-CRKL and 32D-CRKLΔSH3(N)) as well as 32D-mock were generated by the same procedure. All these cells were maintained in RPMI 1640 containing 10% FCS and 15% WEHI-CM.

Transient expression

Cells were transfected with pEGFP-C or pEGFP-C-CRKL, as well as pE-IRES-EGFP, pE-CRKL-IRES-EGFP or pE-CRKLΔSH3(N)-IRES-EGFP by electroporation using standard conditions (Matulonis et al., 1993). Twenty-four hours after transfection, cells were treated with 10 Gy ionizing radiation using a Gammacell 1000 (Atomic Energy of Canada) and plated in 35-mm plates coated with fibronectin (5 μg/cm2). Eight hours after plating, cells were observed under phase-contrast and fluorescence microscopy using an Olympus inverted system microscope equipped with standard FITC filters. To enrich GFP expressing cells, cells were sorted by an EPICS Elite-EPS cytometer (Coulter Corp., Hialeah, FL, USA) using a FITC filter. FACS-purified cells were tested for adhesion to fibronectin and analysed for morphological changes as described above.

Immunoblotting and immunoprecipitation

Cells were lysed in buffer containing 50 mM Tris (pH 8.0), 150 mM NaCl, 10% glycerol, 1% NP-40 (w/v), 10 mM EDTA, 100 mM NaF, 1 mM phenylmethylsulfonyl fluoride, 20 mg/ml aprotinin, 1 mM sodium orthovanadate, and 40 mg/ml leupeptin at 108 cells/ml. For immunoprecipitation, cell lysates were incubated with anti-CRKL antibody (monoclonal antibody #5 – 6) (Uemura et al., 1997) and protein A sepharose for 3 h at 4°C. Protein samples were separated under reducing conditions by SDS-polyacrylamide gel electrophoresis and transferred to Immobilon PVDF membranes by electroblotting. The membranes were blocked with 5% non-fat dry milk in TBS (10 mM Tris, pH 8.0, 150 mM NaCl) and probed with primary antibodies overnight at 4°C. The primary antibodies used were anti-CRKL (5 – 6), anti-ABL (Ab-3; Oncogene Science, Manhasset, NY, USA), anti-C3G polyclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA), and anti-p85PI3K polyclonal antibody (Upstate Biotechnology, Lake Placid, NY, USA). After washing, membranes were further probed with HRP-coupled secondary antibodies for 1 h, washed and subjected to the ECL chemiluminescence system (Amersham).

Adhesion assays

Cell adhesion to fibronectin was performed as previously described (Bazzoni et al., 1996). Briefly, 96-well plates were coated with fibronectin (5 μg/cm2: GIBCO – BRL, Gaithersburg, MD, USA), and blocked with 0.1% heat-inactivated BSA. Cells were labeled with the fluorescent dye BCECF-AM (Molecular Probes, Eugene, OR, USA), and were added to each well (5×104 cells/100 ml of RPMI 1640 containing 0.1% heat-inactivated BSA/well). Then, plates were incubated for 30 min at 37°C followed by washing with warm RPMI to remove unbound cells. Plates were subjected to a fluorescence analyzer Cytofluor 2300 (Millipore, Bedford, MA, USA) before and after washing. After adjustment by input value (before washing) and subtraction of background cell binding to BSA-coated wells, estimates of cells bound/mm2 were calculated. In some experiments, an anti-β1 integrin antibody, Ha2/5 (Pharmingen, San Diego, CA, USA) was added to adhesion assays to determine the role of β1 integrins in the adhesion to fibronectin.

Flow cytometry

Total β1 integrin surface expression was measured using an anti-β1 integrin antibody, HMβ1-1 (Pharmingen) and fluorescence activated flow cytometry (FACS, Coulter Elite, Coulter Corporation, Miami, FL, USA). Expression of the conformational epitope detected by monoclonal antibody 9EG7 on β1 integrin has previously been shown to be associated with ligand binding or activation (Lenter et al., 1993). The 9EG7 epitope was measured as described previously (Bazzoni et al., 1996). Briefly, starved cells were washed twice with Tris-buffered saline, and incubated for 30 min at 37°C with or without a peptide ligand for β1 integrins (GRGDSP; GIBCO – BRL). Monoclonal antibody 9EG7 (Pharmingen) was then added and fluorescence detected by FACS.

Phase contrast and time-lapse video microscopy

Cells were plated onto 35-mm plates coated with fibronectin (5 μg/cm2). In some experiments, cells were first treated with 10 Gy ionizing radiation using a Gammacell 1000 (Atomic Energy of Canada). Twenty hours after plating, cells were observed under phase-contrast microscopy using a Olympus inverted system microscope IX70 to look for changes in morphology. Similarly, in some experiments, cells were examined by video microscopy using a Olympus inverted system microscope IX70, Omega temperature control device for controlling the temperature and CO2, Optronics Engineering DEI-750 3CCD digital video camera, and Sony SVT-S3100 time-lapse S-VHS video recorder. For imaging presentation, video images were captured and printed with the Sony Color Video Printer UP-5600MD.


  1. Altun-Gultekin ZF, Chandriani S, Bougeret C, Ishizaki T, Narumia S, deGraaf P, Van Bergen en Henegouwen P, Hanafusa H, Wagner JA and Birge RB. . 1998 Mol. Cell. Biol. 18: 3044–3058.

  2. Andoniou CE, Thien CB and Langdon WY. . 1996 Oncogene 12: 1981–1989.

  3. Barber DL, Mason JM, Fukazawa T, Reedquist KA, Druker BJ, Band H and D'Andrea AD. . 1997 Blood 89: 3166–3174.

  4. Bazzoni G, Carlesso N, Griffin JD and Hemler ME. . 1996 J. Clin. Invest. 98: 521–528.

  5. Bellis SL, Perrotta JA, Curtis MS and Turner CE. . 1997 Biochem. J. 15: 375–381.

  6. Birge RB, Fajardo JE, Mayer BJ and Hanafusa H. . 1992 J. Biol. Chem. 267: 10588–10595.

  7. Birge RB, Fajardo JE, Reichman C, Shoelson SE, Songyang Z, Cantley LC and Hanafusa H. . 1993 Mol. Cell. Biol. 13: 4648–4656.

  8. Chin H, Saito T, Arai A, Yamamoto K, Kamiyama R, Miyasaka N and Miura O. . 1997 Biochem. Biophys. Res. Commun. 239: 412–417.

  9. de Jong R, ten Hoeve J, Heisterkamp N and Groffen J. . 1995 J. Biol. Chem. 270: 21468–21471.

  10. de Jong R, ten Hoeve J, Heisterkamp N and Groffen J. . 1997 Oncogene 14: 507–513.

  11. Feller SM, Knudsen B and Hanafusa H. . 1994 EMBO J. 13: 2341–2351.

  12. Feller SM, Knudsen B and Hanafusa H. . 1995 Oncogene 10: 1465–1473.

  13. Hasegawa H, Kiyokawa E, Tanaka S, Nagashima K, Gotoh N, Shibuya M, Kurata T and Matsuda M. . 1996 Mol. Cell. Biol. 16: 1770–1776.

  14. Ichiba T, Kuraishi Y, Sakai O, Nagata S, Groffen J, Kurata T, Hattori S and Matsuda M. . 1997 J. Biol. Chem. 272: 22215–22220.

  15. Jacobson K, Kravitz J, Kincade PW and Osmond DG. . 1996 Blood 87: 73–82.

  16. Kanner SB, Reynolds AB, Wang HC, Vines RR and Parsons JT. . 1991 EMBO J. 10: 1689–1698.

  17. Kitayama H, Sugimoto Y, Matsuzaki T, Ikawa Y and Noda M. . 1989 Cell 56: 77–84.

  18. Klemke RL, Leng J, Molander R, Brooks PC, Vuori K and Cheresh DA. . 1998 J. Cell Biol. 140: 961–972.

  19. Lenter M, Uhlig H, Hamann A, Jeno P, Imhof B and Vestweber D. . 1993 Proc. Natl. Acad. Sci. USA 90: 9051–9055.

  20. Lewis JM, Baskaran R, Taagepera S, Schwartz MA and Wang JY. . 1996 Proc. Natl. Acad. Sci. USA 93: 15174–15179.

  21. Lo SH, Weisberg E and Chen LB. . 1994 Bioessays 16: 817–823.

  22. Manie SN, Beck AR, Astier A, Law SF, Canty T, Hirai H, Druker BJ, Avraham H, Haghayeghi N, Sattler M, Salgia R, Griffin JD, Golemis EA and Freedman AS. . 1997 J. Biol. Chem. 272: 4230–4236.

  23. Matsuda M, Tanaka S, Nagata S, Kojima A, Kurata T and Shibuya M. . 1992 Mol. Cell. Biol. 12: 3482–3489.

  24. Matulonis U, Salgia R, Okuda K, Druker B and Griffin JD. . 1993 Exp. Hematol. 21: 1460–1466.

  25. Mayer BJ, Hamaguchi M and Hanafusa H. . 1988 Nature 332: 272–275.

  26. Mayer BJ, Hirai H and Sakai R. . 1995 Curr. Biol. 5: 296–305.

  27. Mizushima S and Nagata S. . 1990 Nucl. Acids Res. 18: 5322.

  28. Nichols GL, Raines MA, Vera JC, Lacomis L, Tempst P and Golde DW. . 1994 Blood 84: 2912–2918.

  29. Nojima Y, Morino N, Mimura T, Hamasaki K, Furuya H, Sakai R, Sato T, Tachibana K, Morimoto C and Yazaki Y, et al. 1995 J. Biol. Chem. 270: 15398–15402.

  30. Oda T, Heaney C, Hagopian JR, Okuda K, Griffin JD and Druker BJ. . 1994 J. Biol. Chem. 269: 22925–22928.

  31. Onoda JM, Piechocki MP and Honn KV. . 1992 Radiation Res. 130: 281–288.

  32. Park H-O, Bi E, Pringle JR and Herskowitz I. . 1997 Proc. Natl. Acad. Sci. USA 94: 4463–4468.

  33. Polte TR and Hanks SK. . 1995 Proc. Natl. Acad. Sci. USA 92: 10678–10682.

  34. Rebstein PJ, Cardelli J, Weeks G and Spiegelman GB. . 1997 Exp. Cell Res. 231: 276–283.

  35. Ren R, Zheng-Sheng Y and Baltimore D. . 1994 Genes Dev. 8: 783–795.

  36. Rosen MK, Yamazaki T, Gish GD, Kay CM, Pawson T and Kay LE. . 1995 Nature 374: 477–479.

  37. Sakai R, Iwamatsu A, Hirano N, Ogawa S, Tanaka T, Mano H, Yazaki Y and Hirai H. . 1994 EMBO J. 13: 3748–3756.

  38. Salgia R, Li J-L, Lo SH, Brunkhorst B, Kansas GS, Sobhany ES, Sun Y, Pisick E, Hallek M, Ernst T, Tantravahi R, Chen LB and Griffin JD. . 1995a J. Biol. Chem. 270: 5039–5047.

  39. Salgia R, Pisick E, Sattler M, Li JL, Uemura N, Wong WK, Burky SA, Hirai H, Chen LB and Griffin JD. . 1996a J. Biol. Chem. 271: 25198–25203.

  40. Salgia R, Sattler M, Pisick E, Li JL and Griffin JD. . 1996b Exp. Hematol. 24: 310–313.

  41. Salgia R, Uemura N, Okuda K, Li JL, Pisick E, Sattler M, de Jong R, Druker B, Heisterkamp N, Chen LB and Griffin JD. . 1995b J. Biol. Chem. 270: 29145–29150.

  42. Sattler M, Salgia R, Okuda K, Uemura N, Durstin MA, Pisick E, Xu G, Li JL, Prasad KV and Griffin JD. . 1996 Oncogene 12: 839–846.

  43. Sattler M, Salgia R, Shrikhande G, Verma S, Pisick E, Prasad KV and Griffin JD. . 1997a J. Biol. Chem. 272: 10248–10253.

  44. Sattler M, Salgia R, Shrikhande G, Verma S, Uemura N, Law SF, Golemis EA and Griffin JD. . 1997b J. Biol. Chem. 272: 14320–14326.

  45. Sawasdikosol S, Ravichandran KS, Lee KK, Chang JH and Burakoff SJ. . 1995 J. Biol. Chem. 270: 2893–2896.

  46. Schumacher C, Knudsen BS, Ohuchi T, Di Fiore PP, Glassman RH and Hanafusa H. . 1995 J. Biol. Chem. 270: 15341–15347.

  47. Senechal K, Halpern J and Sawyers CL. . 1996 J. Biol. Chem. 271: 23255–23261.

  48. Senechal K, Heaney C, Druker B and Sawyers CL. . 1998 Mol. Cell. Biol. 18: 5082–5090.

  49. Songyang Z, Shoelson SE, Chaudhuri M, Gish G, Pawson T, Haser WG, King F, Roberts T, Ratnofsky S, Lechleider RJ, Neel BG, Birge RB, Fajardo JE, Chou MM, Hanafusa H, Schaffhausen B and Cantley LC. . 1993 Cell 72: 767–778.

  50. Tanaka S, Morishita T, Hashimoto Y, Hattori S, Nakamura S, Shibuya M, Matuoka K, Takenawa T, Kurata T and Nagashima K, et al. 1994 Proc. Natl. Acad. Sci. USA 91: 3443–3447.

  51. ten Hoeve J, Arlinghaus RB, Guo JQ, Heisterkamp N and Groffen J. . 1994a Blood 84: 1731–1736.

  52. ten Hoeve J, Kaartinen V, Fioretos T, Haataja L, Voncken JW, Heisterkamp N and Groffen J. . 1994b Cancer Res. 54: 2563–2567.

  53. ten Hoeve J, Morris C, Heisterkamp N and Groffen J. . 1993 Oncogene 8: 2469–2474.

  54. Uemura N, Salgia R, Li JL, Pisick E, Sattler M and Griffin JD. . 1997 Leukemia 11: 376–385.

  55. White GC, Crawford N and Fischer TH. . 1993 Adv. Exp. Med. Biol. 344: 187–194.

Download references


This work was supported by NIH grants CA36167 and DK560654 (JDG).

Author information



Corresponding author

Correspondence to James D Griffin.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Uemura, N., Salgia, R., Ewaniuk, D. et al. Involvement of the adapter protein CRKL in integrin-mediated adhesion. Oncogene 18, 3343–3353 (1999).

Download citation


  • adhesion
  • integrins
  • signal transduction
  • adapter proteins
  • CRKL

Further reading