Polarized distribution of Bcr-Abl in migrating myeloid cells and co-localization of Bcr-Abl and its target proteins


Bcr-Abl plays a critical role in the pathogenesis of Philadelphia chromosome-positive leukemia. Although a large number of substrates and interacting proteins of Bcr-Abl have been identified, it remains unclear whether Bcr-Abl assembles multi-protein complexes and if it does where these complexes are within cells. We have investigated the localization of Bcr-Abl in 32D myeloid cells attached to the extracellular matrix. We have found that Bcr-Abl displays a polarized distribution, co-localizing with a subset of filamentous actin at trailing portions of migrating 32D cells, and localizes on the cortical F-actin and on vesicle-like structures in resting 32D cells. Deletion of the actin binding domain of Bcr-Abl (Bcr-Abl-AD) dramatically enhances the localization of Bcr-Abl on the vesicle-like structures. These distinct localization patterns of Bcr-Abl and Bcr-Abl-AD enabled us to examine the localization of Bcr-Abl substrate and interacting proteins in relation to Bcr-Abl. We found that a subset of biochemically defined target proteins of Bcr-Abl redistributed and co-localized with Bcr-Abl on F-actin and on vesicle-like structures. The co-localization of signaling proteins with Bcr-Abl at its sites of localization supports the idea that Bcr-Abl forms a multi-protein signaling complex, while the polarized distribution and vesicle-like localization of Bcr-Abl may play a role in leukemogenesis.


Chronic myelogenous leukemia (CML) is a clonal myeloproliferative disorder resulting from the neoplastic transformation of a hematopoietic stem cell or a pluripotent progenitor cell (Cortes et al., 1996; Enright and McGlave, 1996; Goldman, 1997). In over 90% of the cases, CML is associated with the presence of the Philadelphia chromosome. The Philadelphia chromosome is a result of a reciprocal translocation between chromosomes 9 and 22 that fuses Bcr-encoded sequences upstream of a truncated c-abl, which encodes a non-receptor tyrosine kinase (Melo, 1996). This oncogene produces a fusion protein, Bcr-Abl, in which the Abl protein tyrosine kinase activity is increased. Depending on the precise breakpoint within the bcr and c-abl genes various Bcr-Abl fusion proteins, including p185, p210, and p230, can be generated, and are associated with different types of leukemia (Melo, 1996). Bcr-Abl is apparently important in both initiation and maintenance of the neoplastic transformation. It has been shown that Bcr-Abl transforms a variety of hematopoietic cell types in vitro, including established factor-dependent lymphoid and myeloid cell lines (Daley and Baltimore, 1988; Hariharan et al., 1988). Bcr-Abl can also transform primary bone marrow cells (Kelliher et al., 1993; McLaughlin et al., 1987, 1989; Young and Witte, 1988) and transform fibroblast cell lines (Lugo and Witte, 1989; Renshaw et al., 1995). In vivo, expression of Bcr-Abl in bone marrow cells of mice by retrovirus transduction induces a myeoloproliferative disorder in mice resembling CML (Daley et al., 1990; Elefanty et al., 1990; Kelliher et al., 1990; Zhang and Ren, 1998). Transgenic strains of mice expressing Bcr-Abl also develop a variety of hematopoietic neoplasms (Hariharan et al., 1989; Honda et al., 1995, 1998; Voncken et al., 1995).

CML cells are apparently deranged in apoptosis, proliferation, differentiation, and adhesion/migration (Sattler and Salgia, 1997). How Bcr-Abl activated signaling pathways deregulate these cellular processes in CML cells is a subject of intense investigation. Studies over the years have revealed that Bcr-Abl contains multiple functional domains and motifs that regulate and mediate its functions (Raitano et al., 1997). The Abl part of Bcr-Abl contains the Src-homology-3 (SH3), SH2, and tyrosine kinase domains in its N-terminal half, and nuclear localization signals, a DNA binding domain, an actin binding domain, and SH3 binding sites in its C-terminal region (Raitano et al., 1997; Wang, 1993). The Bcr part of all variants of Bcr-Abl contains a coiled-coil oligomerization domain, a serine/threonine kinase domain, and tyrosine phosphorylation sites, including the Grb2 SH2 domain binding site (at Y177) (Liu et al., 1993, 1996; Maru and Witte, 1991; McWhirter et al., 1993; Pendergast et al., 1993; Puil et al., 1994; Wu et al., 1998). Bcr-Abl/p210, in addition, contains a pleckstrin homology (PH) domain and a Dbl/CDC24 guanine-nucleotide exchange factor homology domain (Raitano et al., 1997). It is believed that the multiple domains of Bcr-Abl/p210 cooperatively to affect intracellular signaling pathways that ultimately lead to transformation of cells. Since the tyrosine kinase activity of Bcr-Abl has been shown to be essential for its oncogenic potential both in vitro and in vivo (Lugo et al., 1990; Zhang and Ren, 1998), a lot of effort has been made to identify proteins that are tyrosine phosphorylated in cells expressing Bcr-Abl. As a result, dozens of potential substrates of Bcr-Abl have been identified, including Bcr-Abl itself, Shc, Crk/Crkl, c-Cbl, PI3-kinase, p62-Dok, Ras-GAP, Paxillin, c-Bcr, and c-Abl (Sattler and Salgia, 1997). In addition, Bcr-Abl has been found to interact with a large number of proteins, including Crk/Crkl, Grb2, Shc, c-Cbl, p62-Dok, and PI3-kinase, and to activate multiple signaling pathways, including the Ras pathway, PI3-kinase/Akt pathway, Jak/STAT pathway, and adhesion/migration pathway (Gotoh and Broxmeyer, 1997; Raitano et al., 1997; Sattler and Salgia, 1997; Sawyers, 1997). However, the role and relative importance of each substrate/interacting protein of Bcr-Abl in activating the leukemogenic pathway remains to be elucidated.

Signal transduction involves recruiting and/or localizing signaling molecules as complexes or networks to correct compartments in the cell (Lester and Scott, 1997; Pawson and Scott, 1997). Most of our understanding of formation of Bcr-Abl-signaling complexes is based on results obtained biochemically through methods such as co-immunoprecipitation assay. However, these assays do not address directly the formation of protein complexes beyond each pair of proteins. Therefore, it remains largely unknown whether Bcr-Abl assembles multi-protein signaling complexes, and if it does which proteins are in the complex and where these complexes are within the cells. Subcellular co-localization study is a complementing method to gain insight into these questions. However, the overall cytoplasmic localization of Bcr-Abl in cells under the conditions reported previously was impossible to reveal co-localization of Bcr-Abl and its target proteins (McWhirter et al., 1993; McWhirter and Wang, 1993; 1997; Wetzler et al., 1993).

In this study we examined the subcellular localization of Bcr-Abl in 32D myeloid cells under the condition of adhesion and migration. We found that Bcr-Abl displayed a polarized distribution in migrating 32D cells and localized partly onto vesicle-like structures in non-migrating cells. The distinct localization pattern of Bcr-Abl enabled us to examine the localization of a number of Bcr-Abl substrates and interacting proteins in relation to Bcr-Abl. We found that a subset of the biochemically defined target proteins of Bcr-Abl, such as Grb2, Shc, and c-Cbl, co-localized with Bcr-Abl or Bcr-Abl mutant proteins on F-actin and/or vesicle-like structures. These results suggest that Bcr-Abl is localized on distinct intracellular compartments and that it assembles stable signaling complexes with specific signaling molecules on F-actin and vesicle-like structures.


Localization of Bcr-Abl in 32D myeloid cells

To investigate the cellular localization of Bcr-Abl and its interacting proteins in hematopoietic cells, we infected the 32D myeloid cell line with a retrovirus carrying the bcr-abl oncogene. Expression of Bcr-Abl can render 32D myeloid cells factor independent as previously described (Laneuville et al., 1991). We also observed that 32D cells expressing Bcr-Abl display an increased motility phenotype in the absence of IL-3. The migration of the Bcr-Abl expressing 32D cells was observed under the microscope (data not shown) and shown by their morphology and the F-actin distribution (Figure 1). Observation of the Bcr-Abl expressing 32D cells with a time-lapse video microscopy revealed two types of migrating cells. A large fraction of cells move fast. These cells appeared to continuously form filopodia and lamelipodia without forming a distinguishable trailing tail. This type of migrating cell did not adhere well, which is consistent with its fast movement and formation of only transient focal adhesions, and therefore, was easily washed off during the immunofluorescence procedure. The other type of cell moves slower, and usually has the typical polarized cellular morphology found in many migrating cells (Figure 1A-a, 1B-a) (Lauffenburger and Horwitz, 1996 and references therein). These results are consistent with the previous finding that Bcr-Abl expression increases the extension of filopodia and lamelipodia in progenitor cell line Ba/F3 and induces an increased spontaneous motility in both fibroblasts and Ba/F3 cells (Salgia et al., 1997). Stimulation of adhesion by Bcr-Abl in 32D myeloid cells has also been previously reported (Bazzoni et al., 1996). Since migration requires the formation of appropriate cell adhesions, the increased cell adhesion seen in Bcr-Abl expressing 32D cells may be related to the increased motility phenotype.

Figure 1

The localization of Bcr-Abl in migrating and stationary 32D cells. (A) The localization of Bcr-Abl (red, a′,b′,c′) in 32D cells was determined by indirect immunofluorescence as described in Materials and methods. A monoclonal anti-Abl antibody, Ab-3, was used as the primary antibody, and Texas-red-conjugated goat anti-mouse antibody the secondary. FITC conjugated phalloidin was added with secondary antibody to visualize F-actin (green, a,b,c). The localization of F-actin and Bcr-Abl are shown in a migrating 32D cell (a,a′), in a migrating 32D cell with a long pseudopodial extension (b,b′), and in a stationary 32D cell (c,c′). Arrowheads indicate sites of F-actin and Bcr-Abl co-localization, while the arrow points out the vesicle-like structures on which Bcr-Abl is localized. Notice that in c and c′ the arrow points out that there is no concentrated F-actin at the corresponding position of the Bcr-Abl localized vesicle-like structures. (B) F-actin and Bcr-Abl localization are shown in a migrating 32D cell which displays F-actin rich protrusions at the cell front (left side of the cell) and a lot of F-actin aggregates at the cell center (a,a′) and in 32D cells at different states of migration (b,b′). The localization of vinculin in a migrating Bcr-Abl expressing 32D cell is shown in c. Note the focal adhesions at the cell front. The localization of Bcr-Abl in a dividing 32D cell is shown in d

We took advantage of the enhanced adhesion of Bcr-Abl expressing 32D cells to perform localization studies by indirect immunofluorescence. This enabled us to avoid the use of cytocentrifugation, which has been used in most previous localization studies in hematopoietic cells. As a result, much more intracellular detail was preserved, especially the structure of the cytoskeleton. Preservation of these intracellular structures is very important, given the strong association of Bcr-Abl with the F-actin component of the cytoskeleton and the importance of this association for transformation.

We found that Bcr-Abl was generally localized in three different areas in 32D cells. First, there was a large amount of Bcr-Abl co-localizing with F-actin fibers and to a lesser extent with F-actin aggregates (Figure 1A-c, c′). Second, there was a large cytoplasmic pool of Bcr-Abl which did not seem to be associating with any recognizable intracellular structure (Figure 1A-c, c′). The third place where Bcr-Abl was localized were vesicle-like structures that lacked detectable F-actin (Figure 1A-c, c′). Surprisingly the fraction of Bcr-Abl localizing in each of these three areas, as well as the overall distribution of Bcr-Abl in cells, varied greatly in cells with different functional states. Figure 1A-a, 1B-a and 1B-c show migrating Bcr-Abl expressing 32D cells. These cells appear to be at different states of migration judging from their F-actin distribution. The cell in Figure 1B-a has a large lamelipodial extension and many F-actin rich filopodia. Most of the F-actin is found in the form of aggregates at the center of the cell. The cell in Figure 1A-a, on the other hand, had a large number of F-actin aggregates both at the cell front and in the center of the cell, with almost no filopodial extensions. In these migrating cells, Bcr-Abl displays an extremely polarized distribution with most of the protein found at trailing portions of the cells where they co-localize with F-actin (Figure 1A-a, a′, 1B-a, a′). The trailing portions of the cells in Figure 1A-a and 1B-a have much less F-actin than the leading edge and the center of the cell, yet Bcr-Abl is concentrated there, suggesting a preference of Bcr-Abl for the contractile F-actin fibers found in the rear of migrating cells. Most focal adhesions, defined by the localization of vinculin, were found at the front of migrating 32D cells and generally in the periphery of adhering cells expressing Bcr-Abl (Figure 1B-c). However no Bcr-Abl was detected on these focal adhesions.

Some migrating cells had a very long trailing tail (Figure 1A-b, b′). In these cells Bcr-Abl was found to be concentrated in the long extensions (Figure 1A-b, b′). Furthermore, the polarized distribution of Bcr-Abl was also apparent in cells which lacked a typical migrating morphology and F-actin distribution (Figure 1B-b,b′). We cannot distinguish whether these cells are the fast migrating cells or the slow migrating cells at their end stage of migration, nor can we distinguish the front and end of these cells. We are therefore unable to deduce whether Bcr-Abl is concentrated at the front or at the rear of these cells.

The polarized distribution of Bcr-Abl appeared also in dividing cells. But in this case Bcr-Abl was found predominantly at the cleavage furrow where it co-localizes with the contractile F-actin ring (Figure 1B-d). The polarized distribution of Bcr-Abl was, however, not apparent in resting rounded-up cells. In these cells, although some Bcr-Abl co-localized with F-actin fibers and aggregates, a large amount of Bcr-Abl was concentrated onto much larger and prominent vesicle-like structures which did not contain any detectable F-actin (Figure 1A-c,c′). Flattened cells which did not have a polarized morphology rarely had any of the vesicle-like localization of Bcr-Abl and they also lacked a polarized Bcr-Abl distribution (data not shown).

An actin-binding defective mutant of Bcr-Abl is predominantly localized on vesicle-like structures in 32D cells

The actin-binding domain of Bcr-Abl was found to contribute to transformation by Bcr-Abl (McWhirter and Wang, 1993). Bcr-Abl proteins with a deletion or mutation of the actin-binding domain have a reduced ability to transform Rat-1 fibroblasts and to abrogate the requirement for IL-3 in progenitor cell line Ba/F3. However the actin-binding domain mutants of Bcr-Abl nevertheless can still transform cells, suggesting that there are actin-binding independent mechanisms by which Bcr-Abl can transform cells. We examined the subcellular localization of the actin-binding domain-deletion mutant of Bcr-Abl (Bcr-Abl-AD) in 32D cells. Bcr-Abl-AD expressing 32D cells had a very similar phenotype with the Bcr-Abl expressing cells, having an increased ability to adhere and increased motility in the absence of IL-3. We found that Bcr-Abl-AD is primarily localized on vesicle-like structures (see characterization of the vesicle-like structure in the final section of Results) similar to the ones on which the wild-type Bcr-Abl was found (Figure 2A-c). The localization of Bcr-Abl-AD on the vesicle-like structures was the same in all cell types examined including NIH3T3 (Figure 2A-a), 3Y1 (data not shown), BOSC23 (Figure 5-a,b,c) and 32D cells (Figure 2A-c). A similar localization of Bcr-Abl-AD on F-actin-free/low doted structures was described previously in COS and Rat-1 cells (McWhirter and Wang, 1993). In most cells the vesicle-like structures were dispersed through the cytoplasm in all planes inside the cell. In some 32D cells all the vesicle-like structures were clustered in three or four groups, but this was not observed in NIH3T3 cells (data not shown). A Bcr-Abl mutant with an insertion mutation, equivalent to Bcr-Abl/isTth as previously described (McWhirter and Wang, 1993), that specifically disrupts the F-actin binding of Bcr-Abl, was also found to localize to the vesicle-like structures in NIH3T3 cells (data not shown).

Figure 2

Localization of Bcr-Abl-AD in NIH3T3 fibroblast and 32D cells. (A) NIH3T3 fibroblast and 32D cells infected with bcr-abl-ad containing retroviruses (a,a′,c,c′) and NIH3T3 cells infected the wild-type bcr-abl containing retroviruses (b,b′) were double stained with a polyclonal anti-Abl antibody/Texas-red-conjugated goat anti-mouse antibody (a,b,c) and FITC conjugated phalloidin (a′,b′) or a monoclonal anti-vinculin antibody/FITC-conjugated goat anti-mouse antibody (c′). a and a′ show two cells expressing Bcr-Abl-AD (center and right hand corner) and part of a NIH3T3 cell without Bcr-Abl-AD (left hand corner). Note the extremely bright stress fiber staining of the normal NIH3T3 cell compared to that of the Bcr-Abl-AD expressing cells. b,b′ show the co-localization of the wild-type Bcr-Abl and stress fiber in NIH3T3 cells. c and c′ show the localization of Bcr-Abl-AD (c) and vinculin (c′) in a migrating 32D cell. The arrows indicate the respective sites of the Bcr-Abl-AD localized vesicle-like structures in the picture showing the distribution of vinculin. The bright staining at the cell rear of c is another vesicular structure, which is out of focus. Note the exclusion of Bcr-Abl from the cell front and cell rear focal adhesions and the lack of co-localization of vinculin with Bcr-Abl on the vesicle-like structures. (B) Expression of actin in normal NIH3T3 cell and NIH3T3 cells transformed by Bcr-Abl or Bcr-Abl-AD

Figure 5

Translocation of signaling proteins onto Bcr-Abl-AD localized vesicle-like structures in Bcr-Abl-AD expressing 32D cells. Localization of signaling proteins, Grb-2 (a′), p85 subunit of PI3-kinase (b′), p130Cas (c′), c-Cbl (d′), and Shc (e′) on the vesicle-like structures in Bcr-Abl-AD expressing 32D cells were revealed by indirect immunofluorescence. The vesicle-like structures were revealed by immunofluorescence of Bcr-Abl-AD (b,c) or phosphotyrosine signals (a,d,e) in cases of double staining experiments with polyclonal antibodies, because anti-Abl monoclonal antibody Ab-3 does not bind the Bcr-Abl-AD mutant. A polyclonal anti-Abl antibody, K12, a monoclonal anti-phosphotyrosine antibody, 4G10, and the respective antibodies against the signaling proteins were used as primary antibodies. Texas-red-conjugated goat anit-rabbit antibody (a,d,e), Texas-red-conjugated goat anti-mouse antibody (b,c), FITC-conjugated goat anti-mouse antibody (a′,d′,e′), and FITC-conjugated goat anti-rabbit antibody (b′,d′) were used as the secondary antibodies. Note non-Bcr-Abl-AD-expressing cells on the left in (a,a′) and at the top left corner in (b,b′) lack the localization of Grb2 (a′) and p85 (b′) on vesicle-like structures

Unlike the wild-type Bcr-Abl, no polarization of the Bcr-Abl-AD mutant was observed in migrating 32D cells (Figure 2A-c). But like the wild-type Bcr-Abl, Bcr-Abl-AD did not display any detectable co-localization with focal adhesion proteins at focal adhesions of the cell (Figure 2A-c,c′). We have not found detectable differences in the actin cytoskeletons of 32D cells expressing Bcr-Abl or Bcr-Abl-AD. However, differences were found in NIH3T3 fibroblast cells expressing these proteins. Bcr-Abl expressing cells had reduced stress fibers as compared to normal NIH3T3 cells, and fibrobasts expressing Bcr-Abl-AD had even fewer stress fibers than Bcr-Abl expressing cells (Figure 2A-a′,b′), suggesting that both Bcr-Abl and Bcr-Abl-AD cause reduction of stress fibers but the binding of Bcr-Abl to stress fibers stabilizes them. A Western blot analysis revealed that the expression levels of actin were similar between Bcr-Abl and Bcr-Abl-AD expressing NIH3T3 cells (Figure 2B). Interestingly, both Bcr-Abl and Bcr-Abl-AD transformed NIH3T3 cells had reduced level of actin as compared to untransformed NIH3T3 cells (Figure 2B). The expression of focal adhesion proteins, vinculin and paxillin, on the other hand, was not significantly changed in these cells (data not shown).

The majority of phosphotyrosine immunofluoresent signals co-localize with Bcr-ABl and Bcr-Abl-AD

The distinct localization pattern of Bcr-Abl enabled us to examine the localization of Bcr-Abl's substrates and interacting proteins in relation to Bcr-Abl itself. We first examined the localization of total phosphotyrosine signals in normal and Bcr-Abl transformed cells by indirect immunofluorescence. For this study we used both 32D cells and NIH3T3 fibroblasts. In untransformed fibroblast cells phosphotyrosine signals were mainly found on focal adhesions (data not shown), in agreement with many published studies. However in Bcr-Abl transformed cells, the majority of phosphotyrosine immunofluorescent signals were found co-localizing on the stress fibers and on punctate structures with Bcr-Abl itself (Figure 3-a, b). Similar results were obtained in 32D cells where the majority of the phosphotyrosine signal co-localized with Bcr-Abl (data not shown). Since the pattern of localization of the actin deletion mutant Bcr-Abl-AD is more distinct, we examined the localization of tyrosine phosphorylated proteins in both NIH3T3 and 32D cells expressing Bcr-Abl-AD. In Figure 3c,c′ it is shown that the majority phosphotyrosine signals co-localize with Bcr-Abl-AD on vesicle-like structures in Bcr-Abl-AD transformed NIH3T3 cells. The phosphotyrosine signals should be in part due to tyrosine phosphorylated Bcr-Abl. However, since autophosphorylated Bcr-Abl accounts for only a fraction of the total tyrosine phosphorylated proteins in Bcr-Abl transformed cells, as judged by a Western blot analysis of total cell lysate of Bcr-Abl transformed cells with anti-phosphotyrosine antibodies (data not shown), the results shown here suggest that a number of tyrosine phosphorylated proteins in Bcr-Abl transformed cells may co-localize with Bcr-Abl.

Figure 3

Co-localization of the majority of tyrosine phosphorylated proteins with Bcr-Abl and Bcr-Abl-AD. The localization of Bcr-Abl and tyrosine phosphorylated proteins on F-actin stress fibers in NIH3T3 cells expressing Bcr-Abl (a,b) and co-localization of Bcr-Abl-AD and phosphotyrosine signals on vesicle-like structures in NIH3T3 cells expressing Bcr-Abl-AD (c,c′) were revealed by indirect immunofluorescence with a polyclonal anti-Abl antibody (a,c) and a monoclonal anti-phosphotyrosine antibody (b,c′). Note the localization of Bcr-Abl and phosphotyrosine signals on periodic bead-like structures of the stress fiber in NIH3T3 cells expressing Bcr-Abl (a,b)

A number of signaling proteins co-localize with Bcr-Abl

To check if specific target proteins of Bcr-Abl co-localize with Bcr-Abl, we examined the cellular localization of a number of signaling proteins in Bcr-Abl expressing cells. We have chosen to study proteins that have specific antibodies both suitable and available for immunofluorescent studies (see Materials and methods). The relative amount of a protein co-localizing with Bcr-Abl was determined by examining the extent of redistribution of each protein onto the distinct site(s) where Bcr-Abl localized as described above. We first examined adapter proteins Shc and Grb2. Shc has been shown to be phosphorylated in Bcr-Abl transformed cells and to associate with Bcr-Abl (reviewed by Sattler and Salgia, 1997). Grb2, though not a good substrate of Bcr-Abl, has been shown to bind Bcr-Abl. We found that both Shc and Grb-2 localize in the cytoplasm of normal NIH3T3 fibroblast cells (Figure 4A-a,d). Grb-2 was found to co-localize with caveolin in NIH3T3 cells (data not shown). In Bcr-Abl transformed cells, however, both Shc and Grb2 were redistributed and a large fraction of these proteins were found to co-localize on stress fibers and on punctate structures with Bcr-Abl (Figure 4A-b,e). The redistribution of Shc in NIH3T3 cells was more profound than the redistribution of Grb-2. The same results were obtained when 3Y1 rat fibroblasts were used (data not shown). We went on to test the localization of these two proteins in Bcr-Abl expressing 32D cells and found that Bcr-Abl expression caused a dramatic shift in the distribution of the cytoplasmic pools of both proteins. Both Grb-2 and Shc were enriched at the tail of migrating cells, co-localizing with Bcr-Abl (Figure 4B-a,a′,b,b′). c-Cbl was also found to dramatically translocate to the rear of the migrating 32D cells with Bcr-Abl (Figure 4B-c,c′). Such translocations were not detected for other signaling proteins like p85 subunit of PI3-kinase, p130Cas, and Crkl.

Figure 4

Co-localization of signaling proteins with Bcr-Abl and Bcr-Abl AD in NIH3T3 and 32D cells. (A) Cellular localization of Grb-2 (a,b,c) and Shc (d,e,f) in NIH3T3 (a,d), NIH3T3-Bcr-Abl (b,e) and NIH3T3-Bcr-Abl-AD (c,f) fibroblast cells were examined by indirect immunofluorescence with anti-Grb2 and anti-Shc antibodies, respectively. Reduced focal adhesions, revealed by the localization of vinculin, in NIH3T3-Bcr-Abl cells (h) as compared with normal NIH3T3 cells (g) are shown. (B) Co-localization of Bcr-Abl and Grb2 (a,a′) Bcr-Abl and Shc (b,b′), and Bcr-Abl and c-Cbl (c,c′) in migrating 32D cells expressing Bcr-Abl were determined by indirect immunofluorescence. Primary antibodies were anti-Abl (a, b and c)and anti-Grb2 (a′), anti-Shc (b′), and anti-c-Cbl (c′). Secondary antibodies were Texas-red-conjugated goat anti-mouse antibody (a,b,c) and Texas-red-conjugated goat anti-mouse antibody (a′,b′,c′). In b and b′ the bright sites on the right and top of the pictures are rounded up 32D cells

It is possible that a small fraction of proteins such as p85 do interact and co-localize with Bcr-Abl but their interactions were not sufficient to shift the cytoplasmic pools of these proteins to such an extent to be detected by immunofluorescence. Since the concentration of Bcr-Abl and Bcr-Abl-AD on the vesicle-like structures is high and the pattern of the localization is more distinct, it is possible to detect weakly recruited proteins on the vesicular structure with the Bcr-Abl proteins. We used the Bcr-Abl-AD mutant to examine whether signaling proteins co-localized with Bcr-Abl-AD on the vesicle-like structures, because the wild-type Bcr-Abl localized on vesicle-like structures which were much smaller and of lower intensity than the ones on which Bcr-Abl-Ad localized. Our results showed that in both NIH3T3 (Figure 4A-c,77f) and 32D cells (Figure 5a, a′,e,e′) Grb-2 and Shc were recruited to the vesicle-like structures and co-localized with Bcr-Abl-AD, as expected. We also found that c-Cbl, p85 subunit of PI3-kinase, p130Cas, c-Abl, and Crkl (in the order of decreasing extent) were all found on the vesicle-like structures (data shown for p130Cas, c-Cbl and PI3-kinase) co-localizing with Bcr-Abl-AD (Figure 5b,b′,c,c′,d,d′). On the other hand p190Rho-GAP was not detected on the vesicle-like structures (data not shown).

It has been previously reported that a number of focal adhesion proteins are tyrosine phosphorylated in Bcr-Abl transformed cells. We decided to investigate whether the vesicle-like structures contained any focal adhesion proteins, and whether there was any co-localization between Bcr-Abl and focal adhesion proteins. We found that the vesicle-like structures had no detectable levels of vinculin (Figure 2A-c,c′), paxillin, talin, or FAK (data not shown). This result is consistent with the finding that neither Bcr-Abl-AD nor the wild-type Bcr-Abl were detected on focal adhesions in NIH3T3 or 32D cells. Furthermore, expression of either wild-type Bcr-Abl or Bcr-Abl-AD in both NIH3T3 and 32D cells increased the amounts of the adhesion proteins localized in the cytoplasm as compared to the amounts localized on the focal adhesions (Figure 4A-g,h, data shown only for vinculin). As shown in Figure 4A-g and h, the intensity of the vinculin stain on focal adhesions and in cytoplasm was significantly different between the transformed and untransformed NIH3T3 cells. The cytoplasmic pools of the focal adhesion proteins were not found to be enriched at the rear of migrating Bcr-Abl expressing 32D cells, either (Figure 1B-c).

Characterization of the intracellular vesicle-like structures on which Bcr-Abl-Ad localizes

The fact that Bcr-Abl and its target proteins localized to the intracellular vesicle-like structures suggests that this structure may serve as a docking site for Bcr-Abl to assemble signaling complexes and transduce oncogenic signals. We further characterized the nature of these structures. First we examined whether the structures were F-actin containing complexes. We found that the vesicle-like structures that co-localized with Bcr-Abl-AD lacked detectable F-actin as stained by phalloidin (Figure 2A-a,a′), or immunofluorescence with a monoclonal antibody against actin (data not shown). Next we tested the effect of lantranculin-A, a cell permeable G-actin sequestering drug which causes F-actin depolymerization (Spector et al., 1989). Lantranculin-A treatment of Bcr-Abl-AD expressing NIH3T3 fibroblasts for 30 min dissolved all the stress fibers in these cells and caused the cells to completely round up, but had no effect on the vesicle-like structures or the localization of Bcr-Abl-AD on them (data not shown).

To facilitate the characterization of the vesicle-like structures in living cells we constructed fusion proteins of Bcr-Abl or Bcr-Abl-AD and GFP. Both the Bcr-Abl-GFP and the Bcr-Abl-Ad-GFP rendered 32D cells IL-3 independent. The localization of Bcr-Abl-AD-GFP was identical to that of Bcr-Abl-AD determined by indirect immunofluorescence. The localization of Bcr-Abl-GFP was generally similar to that of the wild-type Bcr-Abl determined by indirect immunofluorescence. Both Bcr-Abl-GFP fusion protein and wild-type Bcr-Abl were found to have a polarized distribution in migrating 32D cells and showed a co-localization with F-actin. But the Bcr-Abl-GFP fusion protein showed an increased localization on vesicle-like structures and decreased in polarized distribution in migrating 32D cells when compared to the wild-type Bcr-Abl. This partial change in localization of Bcr-Abl-GFP fusion protein may reflect an interference of the actin binding function by GFP. Despite all this the Bcr-Abl-GFP fusion protein enabled us to do single staining immunofluorescence experiments confirming the co-localization of Bcr-Abl with a number of signaling proteins as mentioned before and the absence of co-localization with the focal adhesion proteins (data not shown). The use of the Bcr-Abl-GFP fusion eliminated the possibility of cross reactivity between antibodies in multiple labeling experiments required before to demonstrate co-localization of two proteins.

We used the Bcr-Abl-AD-GFP for further characterization of the vesicle-like structures. We found that Bcr-Abl-AD-GFP containing vesicle-like structures can move inside the cells with a slower speed than other vesicles like lysosomes (Figure 6a,b,c). In general, smaller vesicle-like structures to which Bcr-Abl-AD-GFP localized moved much faster than larger ones although even the largest ones did move significantly. We went on to investigate whether these vesicle-like structures were endosomes or lysosomes. Co-staining experiments with EEA1, Rab7, rhodamine conjugated transferrin and lysotraker showed that they were neither endosomes nor lysosomes (data not shown). Staining for caveolin with which Grb-2 co-localizes in normal NIH3T3 cells, was also negative in these vesicle-like structures. In addition, we observed that Bcr-Abl-AD was localized on the surface of the vesicle-like structures, and that these structures are phase-dense under phase contrast microscopy (Figure 6d, d′,e,f), suggesting that Bcr-Abl-AD may localize on the surface of lipid containing vesicles.

Figure 6

Characterization of the Bcr-Abl-AD localized vesicle-like structures. (a,b,c). Images of a BOSC23 cell expressing Bcr-Abl-AD/GFP fusion protein acquired using a pixera digital camera. The images were taken at 2-min intervals. Note changes of the relative positions of the vesicle-like structures (movement). (d,d′) Phase contrast image (d) and immunofluorescence image of Bcr-Abl-Ad (d′) of a Bcr-Abl-AD expressing NIH3T3 fibroblast cell. The arrows point out the phase dense vesicle-like structures. (e) Immunofluorescence image of a Bcr-Abl-AD expressing NIH3T3 cell. The vesicular structure pointed to by the arrow is enlarged and shown in f

To investigate the effect of the kinase activity on the localization of Bcr-Abl-AD to the vesicle-like structures, we used a kinase defective Bcr-Abl-AD mutant (K1176R). The kinase deficient mutant lost its ability to localize on vesicle-like structures in all cell types examined including NIH3T3, BOSC23 and 32D cells (data not shown). These results are similar to previous results reporting a kinase dependent localization of Bcr-Abl-AD on phalloidin-stain negative dots (McWhirter and Wang, 1993). In contrast, a kinase defective Bcr-Abl mutant did not lose its ability to localize on stress fibers in NIH3T3 cells, although it did lose the ability to deregulate the actin cytoskeleton. These results demonstrated that the kinase activity is essential for the localization of Bcr-Abl-AD on the vesicular structure, but is not required for Bcr-Abl to bind to F-actin stress fibers.


In this study we have examined the localization of Bcr-Abl and of a number of Bcr-Abl substrates and interacting proteins in 32D myeloid cells under the condition of adhesion and migration. We have found that the localization of Bcr-Abl undergoes drastic changes depending on the functional state of the cell. In migrating cells, of which we can distinguish the front and rear, Bcr-Abl displayed a polarized distribution, localizing on the F-actin fibers in the trailing potions of the cells. In non-migrating cells, Bcr-Abl was equally distributed in the cytoplasm, localizing on both F-actin and F-actin-free/low vesicle-like structures. Deletion of the actin binding domain of Bcr-Able dramatically enhances the localization of Bcr-Abl on these vesicle-like structures. We have demonstrated that a number of important signaling proteins are recruited by Bcr-Abl onto the selected pool of actin filaments and the vesicle-like structures. However, although many of the focal adhesion proteins are tyrosine-phosphorylated in the 32D cells expressing Bcr-Abl, co-localization of Bcr-Abl and several focal adhesion proteins examined was not detected in our studies.

Our results demonstrated for the first time that the localization of Bcr-Abl can be polarized in cells, underlining its selective association with a subtype of F-actin. The F-actin fibers in the rear of migrating cells are generally myosin II containing fibers (Lauffenburger and Horwitz, 1996). Our results, therefore, suggest that Bcr-Abl preferentially binds to myosin II-associated F-actin fibers. This possibility is consistent with the observations that Bcr-Abl co-localizes with myosin II-associated F-actin fibers in other situations – in dividing 32D cells, Bcr-Abl was shown to be concentrated at the cleavage furrow (Figure 1B-d) where myosin II-associated F-actin fibers are typically localized, and in fibroblast cells Bcr-Abl showed a preference for the myosin II-associated contractile stress fibers (McWhirter and Wang, 1993). The selective association of Bcr-Abl to myosin II-associated F-actin fibers raises two further questions – what is the factor(s) determining the preferential F-actin binding by Bcr-Abl in cells and what is the significance of this preferential F-actin binding for Bcr-Abl function in leukemogenesis?

One possible mechanism by which preferential binding of Bcr-Abl to the myosin II-associated contractile F-actin could occur is through blockage of or competition for Bcr-Abl's binding sites on F-actin in the front and middle of the migrating 32D cells by other proteins (Figure 1B-a). The Abl actin binding domain has a weak affinity for F-actin, and it was shown that the S2-3 domain of Gelsolin was able to compete efficiently with Abl for binding to F-actin (Van Etten et al., 1993). The gelsolin family is believed to play an important role in regulating actin nucleation required for the extension of the cell front of migrating cells (Lauffenburger and Horwitz, 1996). Alternatively, the myosin II-containing F-actins could strongly recruit Bcr-Abl. In the latter scenario myosin II or other protein(s) on the F-actin fibers may bind Bcr-Abl in addition to the interaction of Bcr-Abl and F-actin. Consistent with this idea, we found that Bcr-Abl was concentrated on periodic bead-like structures decorating the stress fibers in NIH3T3 cells (Figure 3a). It will be interesting to study the mechanism of the preferential binding of Bcr-Abl to specific pool(s) of F-actin.

One potential function of the preferential binding of Bcr-Abl to the F-actins at the rear of migrating cells is to regulate the adhesion and/or migration of the cells. CML is characterized by a premature release of primitive progenitors and precursors to the peripheral blood, and continuous proliferation of these cells in tissues that are no longer hematopoietic in the adult human (Sattler and Salgia, 1997; Verfaillie et al., 1997). It is believed that this abnormal trafficking of CML progenitors is due, in part, to changes in adhesion molecule expression or function in CML progenitor cells. It has been reported that malignant progenitors in CML have a decreased ability to adhere to both stromal cells and fibronectin but increased adhesion to the basement membrane components laminin and collagen type IV (Verfaillie et al., 1992). There are also studies that have shown that Bcr-Abl can actually stimulate integrin function (Bazzoni et al., 1996). Another possible component of the mechanism underlying abnormal trafficking of CML progenitors is the increased motility of the malignant cells. Bcr-Abl has been shown to induce abnormalities of the cytoskeleton and increase motility of Ba/F3 lymphoid and NIH3T3 fibroblast cell lines (Salgia et al., 1997). Here we report that Bcr-Abl activates motility of the 32D myeloid cell line (Figure 1). Since cell migration needs appropriate cell adhesions, the changed adhesion and migration in Bcr-Abl transformed cells could be related. The polarized distribution of Bcr-Abl might be in part responsible for the deregulated motility of Bcr-Abl expressing cells. Although the fact that Bcr-Abl-AD still stimulates motility of 32D cells without displaying a polarized distribution indicates that the localization of Bcr-Abl to actin is not necessary for the spontaneous motility, detailed analysis of the migration of the Bcr-Abl-expressing versus Bcr-Abl-AD expressing 32D cells using time-lapse video microscopy are needed to determine the effect of the preferential actin binding of Bcr-Abl on cell migration.

Previous studies of the localization of Bcr-Abl have concentrated on the co-localization of Bcr-Abl with the F-actin cytoskeleton (McWhirter et al., 1993; McWhirter and Wang, 1993, 1997; Wetzler et al., 1993). Although certain mutants of Bcr-Abl have been found to localize onto F-actin-free/low doted structures (McWhirter and Wang, 1993, 1997), our results revealed for the first time that the wild-type Bcr-Abl localizes onto vesicle-like structures in hematopoietic cells, underlining the significance of such localizations in Bcr-Abl biology. Regardless of the nature of the vesicle-like structure, the localization of Bcr-Abl to it may be important for activation of signal transduction pathways by Bcr-Abl. This notion is supported by several lines of evidence. First, Bcr-Abl-AD, which is concentrated on the vesicle-like structures, still transforms cells, albeit with reduced efficiency. This indicates that Bcr-Abl-AD is still capable of transducing many of the signals required for cellular transformation from these vesicle-like structures. Second, the localization of Bcr-Abl-AD on the vesicle-like structures is dependent on the tyrosine kinase activity of Bcr-Abl, which correlates this vesicle-like structure localization of Bcr-Abl to the essential role of its tyrosine kinase activity in transformation. Third, Bcr-Abl is shown to recruit a number of signaling proteins to the vesicle-like structure. Last, it was reported that a mutant form of Bcr-Abl lacking amino acid residue 191 to 923 of the Bcr part of Bcr-Abl (Bcr/1-191-Abl) disrupted the stress fibers in fibroblast cells, localized to the cortical F-actin as well as phalloidin negative spots, and displayed an enhanced transforming activity (McWhirter and Wang, 1997). It is not clear whether the vesicle-like structures serve as an alternative signaling docking site, or if the localization of Bcr-Abl onto F-actin and the vesicular structure are both required for Bcr-Abl's full transforming activity. Further study of the mechanism(s) by which Bcr-Abl localizes to the vesicle-like structures could help to address these questions.

Although the importance of the localization of Bcr-Abl on the vesicle-like structures is uncertain, this distinct localization pattern of Bcr-Abl, particularly of Bcr-Abl-AD, provides a useful tool to reveal the co-localization of Bcr-Abl and other signaling proteins within the cell. It may also provide a more sensitive method to detect protein-protein interactions in cells and to compare the relative quantity/affinity of the interaction between Bcr-Abl and its target proteins. For example, we were unable to detect the interaction between Bcr-Abl or Bcr-Abl-AD and p130Cas in cells by immunoprecipitation assay (data not shown) but we did find that p130Cas co-localized with Bcr-Abl-AD on the vesicle-like structures.

There are a large number of proteins which have been shown to be phosphorylated by and/or to interact with Bcr-Abl (Sattler and Salgia, 1997). These proteins are potential compenents of Bcr-Abl activated signal transduction pathways. The distinct localization pattern of Bcr-Abl and Bcr-Abl-AD enabled us to examine the localization of these Bcr-Abl substrates and interacting proteins in relation to Bcr-Abl itself. Based on whether a protein can translocate and co-localize with Bcr-Abl on F-actin and/or vesicle-like structures sufficiently to be detected by immunofluorescence, we can divde the biochemically defined Bcr-Abl target proteins into three classes. One class of proteins, includes Grb2, Shc, and c-Cbl, which were all found to predominantly co-localize with both Bcr-Abl and Bcr-Abl-AD. These proteins are important adapter molecules that regulate and direct protein tyrosine kinase signaling pathways (Bonfini et al., 1996; Langdon, 1995; Schlessinger, 1994). The large quantitative co-localization of these proteins and Bcr-Abl may play critical roles in mediating Bcr-Abl's function in signal transduction. Indeed Bcr-Abl with a point mutation (Y177F) that disrupt the Grb2 SH2 binding had severe defects in inducing CML-like disease in mice (Pear, Ren, and Baltimore, unpublished data). Further identifications of the domain/sites in Bcr-Abl that are responsible for the co-localization of Bcr-Abl with Shc and c-Cbl and examination of the corresponding Bcr-Abl mutants using the mouse CML model will help to assess the importance of the co-localization of Bcr-Abl with Shc and c-Cbl. The second class of proteins, including p130Cas, PI3-kinase, c-Abl and Crkl, were found to co-localize only with Bcr-Abl-AD on the vesicle-like structures. It is not clear whether these proteins only form stable complexes with Bcr-Abl-AD on the vesicle-like structures but not on the F-actin filaments or if their small extent of recruitment by Bcr-Abl on F-actin was not detected. These proteins are included in the Bcr-Abl signaling complex, albeit with smaller quantity as compared to the first class of proteins. The third class of proteins, including p190Rho-GAP, paxillin, vinculin, talin, and FAK, were not found to co-localize with Bcr-Abl on F-actin or on the vesicle-like structures. The absence of co-localization of Bcr-Abl and adhesion proteins was consistent with the fact that Bcr-Abl appears to be excluded from the focal adhesions, defined by the localization of vinculin, in both fibroblast and 32D cells. Although our results could not rule out the possibility that some of the `third class' proteins do interact with Bcr-Abl in the cytoplasm where they overlapped with Bcr-Abl, the data did suggest that the majority of these proteins do not form stable complexes with Bcr-Abl on F-actin or on the vesicle-like structures. These proteins may be phosphorylated by Bcr-Abl without forming stable complexes with Bcr-Abl.

Anchoring protein kinases and their target proteins to certain locations in the cell and the assembly of the correct molecules into signaling complexes is crucial for signal transduction (Lester and Scott, 1997; Pawson and Scott, 1997). The results presented here suggest that Bcr-Abl is localized on distinct intracellular compartments and that it assembles stable signaling complexes which contain a subset of the biochemically defined Bcr-Abl target proteins. These results will help to understand the roles of different signaling proteins in the neoplastic transformation by Bcr – Abl.

Materials and methods

Cell culture

NIH3T3 mouse and 3Y1 Rat fibroblasts were grown in Dulbecco's modified Eagle's medium (DMEM) containing 10% calf serum, 100 units/ml penicillin and 100 μg/ml streptomycin. 32D cells were grown in DMEM containing 10% fetal calf serum, 100 units/ml penicillin, 100 μg/ml streptomycin, and 10% WEHI-3B conditional media as a source of IL-3. Cell lines expressing Bcr-Abl or Bcr-Abl-AD were generated by infecting NIH3T3, 3Y1 and 32D cells with retroviruses (described below) carrying cDNAs encoding the proteins as described (Pear et al., 1993). Bcr-Abl and Bcr-Abl-AD expressing 32D cells were selected by removal of IL-3 2 days after infection. Expression of the above proteins was monitored with Western blot analysis and immunofluorescence.

Plasmid construction and helper-free retrovirus production

Bcr-Abl/p210 (Daley et al., 1990) was cloned into the EcoRI site of both retroviral expression vectors MSCV (Hawley et al., 1995) and pBabe Puro (Morgenstern and Land, 1990). The GFP fusion proteins of Bcr-Abl and Bcr-Abl-AD was constructed by first PCR amplifying the 3′-region of bcr-abl including sequences encoding the actin binding domain with the common 5′ primer: 5′-TCA GGC TTC CGG TCT CCC CA-3′ and 3′ primer: 5′-GCG GAA TTC CTA GGC GGC CGC CCT CTG CAC TAT GTC ACT GAT-3′ (underlined sequences specifies NotI-Stop-codon-EcoRI sites) for Bcr-Abl-GFP and 3′ primer: 5′-GCG GAA TTC CTA GGC GGC CGC CTT TGG AGT GGC CGG GAG CAC-3′ for Bcr-Abl-AD-GFP. The PCR products were cut with SgrAI and EcoRI to generate 1.3 Kb and 0.8 Kb fragments, respectively, and cloned together with the EcoRI and SgrAI cut 5 Kb Bcr-Abl fragment into the EcoRI site of pBabe vector by carrying out a three way ligation. The resulted bcr-abl-NotI and bcr-abl-AD-NotI were then fused in frame with GFP through the NotI site by cutting pBabe bcr-abl and pBabe bcr-abl-AD with EcoRI and NotI and ligating with an MSCV vector containing GFP flanked by NotI and SalI. The modified GFP gene we used was a generous gift from Jacob and Chen in Baltimore's laboratory. The bcr-abl-AD mutant was generated similarly with bcr-abl-AD-NotI except using a 3′ primer without NotI restriction site (5′-CGG AAT TCC TAC TTT GGA GTG GCC GGG AGC-3′). The actin binding domain insertion mutation was generated by disruption of the conserved amino acid sequence, Asp-Ile-Val, at the COOH-end of Bcr-Abl by inserting three amino acids (Arg-Ser-Val) between Asp and Ile (equivalent of Bcr-Abl/isTth; (McWhirter and Wang, 1993). Kinase deficient mutants of Bcr-Abl and Bcr-Abl-AD were generated by replacing the kinase domain from a human c-abl kinase-negative mutant (Sawyers et al., 1994). Helper-free retroviruses were generated by transiently transfecting retroviral vectors into BOSC-23 cells as described (Pear et al., 1993).

Immunofluorescence and immunoblotting

Indirect immunofluorescence assays were carried out as described (Van Etten et al., 1989) with modifications. Cells were plated overnight on glass coverslips (for adhesion experiments the coverslips were first coated with 5 μg/ml fibronectin for 1 h, washed three times with ice cold phosphate buffered saline (PBS) containing 0.5 mM MgCl2 and 0.5 mM CaCl2 (PBS2+) and then fixed for 10 min in 4% paraformaldehyde solution in PBS. Fixation was followed by addition of 50 mM glycine solution in PBS and then the cells were permeabilized using 0.2% triton solution in PBS for 10 min. Permeabilized cells were blocked using 10% normal donkey serum (Jackson Immunoresearch, West Gove, PA, USA) for 30 min. Primary antibodies were added in 5% normal donkey serum solution in PBS and were incubated for 30 min. Cells were then washed for 4 h at room termperature in PBS2+. Following the wash the appropriate secondary antibody was added for 30 min. Dilutions of FITC- and LRSC-conjugated secondary antibodies (Jackson Immunoresearch, West Grove, PA, USA) were 1 : 500 while for Texas red-conjugated secondary antibodies (Molecular Probes, Eugene, OR, USA) were 1 : 1000. The cells were washed for another 3 h in PBS2+ and then were mounted on slides using fluromount (Fisher Scientific, Pittsburgh, PA, USA). Lysotracker, tetramethyl rodamine conjugated transferrin were added in the medium of BOSC23 cells prior to observation under the microscope as suggested by the manufacturer (Molecular Probes, Eugene, OR, USA). FITC- and lissamine rhodamine-conjugated phalloidins (Molecular Probes, Eugene, OR, USA) were added with secondary antibodies.

The slides were viewed using an Olympus IX 70 inverted microscope system equipped with fluorescence. Pictures were taken using Kodak Tmax 400 film. The negatives were digitally scanned with a Nikon scanner and printed on a codonicks NP1600 double sublimation printer. The final magnifications for each image is described as follows: (1A) a,a′ were 4200×; b,b′ were 2580× and c,c′ were 6000×. (1B) a,a′ were 5160×; b,b′ were 4998×; c was 5580× and d was 3600×. (2A) a,a′ were 5100×; b,b′ were 3600×; and c,c′ were 5160×. (3) a,b were 5400× and c,d were 4800×. (4A) a,b,f were 4200×; c was 4800×; d,e were 3600× and g,h were 2100×. (4B) a,a′ were 4800×; b,b′ were 600× and c,c′ were 5400×. (5) a,a′,c,c′,e,e′ were 2580×; b,b′ were 2160× and d,d′ were 1800×. (6) a,b,c were 1200×; d,d′ were 4800×; e was 6840× and f was 27 360×. All pictures were taken using 600× magnification and were enlarged electronically using Adobe Photoshop.

Western blot analysis of total cell lysates was performed as previously described (Ren et al., 1994).

Immune reagents

Anti-Abl monoclonal antibody Ab-3 and polyclonal K-12 were purchased from Oncogene Research Products, Cambridge, MA, USA and Santa Cruz Biotechnology, Santa Cruz, CA, USA, respectively. Anti-phosphotyrosine monoclonal antibody 4G10 was purchased from Upstate Biotechnology, Lake Placid, NY, USA. PY20 hybridoma conditional media was used as a source of the PY20 anti-phosphotyrosine monoclonal antibody. Anti-vinculin monoclonal antibody was purchased from Sigma, St. Louis, MO, USA. Rabbit polyclonal anti-Shc antibody was purchased from Transduction Laboratories, Lexington, KY, USA. Mouse monoclonal EEA1, paxillin, FAK, p190, and p85 subunit of PI3-kinase antibodies were purchased from Transduction Laboratories, Lexington, KY, USA. Grb-2 and c-Cbl polyclonal antibodies were purchased from Santa Cruz Biotechnology. The anti-p130Cas monoclonal antibody used for indirect immunofluorescence was a generous gift from Dr A Bouton (Bouton and Burnham, 1997) anti-Crkl monoclonal antibody from Dr R Salgia, anti-Rab-7 polyclonal antibody from Dr M Zerial and anti-caveolin from Dr M Lasanti. The monoclonal anti-actin and anti-talin antibodies were purchased from Sigma.

In immunofluorescence experiments we have individually tested and compared the staining pattern of each antibody to previously published patterns of each protein in order to determine their suitability for immunofluorescence studies. The antibodies we used against the focal adhesion proteins were monoclonal antibodies against FAK, vinculin, talin, and paxillin. All four antibodies stained focal adhesions in normal NIH3T3 and 3Y1 fibroblasts as expected (data not shown). The Grb-2 polyclonal antibody was used to do immunofluorescence studies in normal NIH3T3 cells and it gave a signal co-localizing with caveolin, which is in agreement with previous biochemical experiments demonstrating the enrichment of Grb-2 in caveolin-like low density membrane fractions (Wu et al., 1997). The specificity and suitability of the anti-Grb2 antibody for immunofluorescence studies was further confirmed with cells overexpressing Grb-2 (data not shown). The c-Cbl polyclonal antibody used to carry out immunofluorescence studies in normal NIH3T3 cells showed a pattern identical to the known localization of c-Cbl (Tanaka et al., 1995). The p130Cas monoclonal antibody had been previously characterized (Bouton and Burnham, 1997). The Shc polyclonal antibody showed a pattern identical to what was previously reported in normal fibroblast cells (Lotti et al., 1996). The anti-Crkl antibody was characterized by Dr Salgia (personal communication) and its specificity and suitability for immunofluorescence studies was confirmed with cells overexpressing Crkl (data not shown). The anti-p190Rho-GAP antibody has been previously used for immunofluorescence studies (Nakahara et al., 1998). The p85 subunit of PI3-kinase antibody gave a cytoplasmic/membrane staining pattern as expected and this pattern was altered in Bcr-Abl-AD cells.

Lantranculin treatment

Fibroblast cells were seeded onto fibronectin coated coverslips and grown overnight. Next morning they were treated with 0.5 mg/ml Lantranculin-A (Molecular Probes) for 10, 20, 30 and 60 min. Cells were then washed and prepared for immunofluorescence as described.

Time-lapse video microscopy

Cells growing on plastic tissue culture plates were examined by video microscopy using an Olympus IX 70 inverted microscope, the Pixera 120es camera system, a PC computer and the Pixera software.


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

  2. Bonfini L, Migliaccio E, Pelicci G, Lanfrancone L and Pelicci PG. . 1996 Trends Biochem. Sci. 21: 257–261.

  3. Bouton AH and Burnham MR. . 1997 Hybridoma 16: 403–411.

  4. Cortes JE, Talpaz M and Kantarjian H. . 1996 Am. J. Med. 100: 555–570.

  5. Daley G and Baltimore D. . 1988 Proc. Natl. Acad. Sci. USA 85: 9312–9316.

  6. Daley GQ, Van Etten RA and Baltimore D. . 1990 Science 247: 824–830.

  7. Elefanty AG, Hariharan IK and Cory S. . 1990 EMBO J. 9: 1069–1078.

  8. Enright H and McGlave PB. . 1996 Curr. Opin. Hematol. 3: 303–309.

  9. Goldman JM. . 1997 Curr. Opin. Hematol. 4: 277–285.

  10. Gotoh A and Broxmeyer HE. . 1997 Curr. Opin. Hematol. 4: 3–11.

  11. Hariharan IK, Adams JM and Cory S. . 1988 Oncogene Res. 3: 387–399.

  12. Hariharan IK, Harris AW, Crawford M, Abud H, Webb E, Cory S and Adams JM. . 1989 Mol. Cell. Biol. 9: 2798–2805.

  13. Hawley RG, Lieu FHL, Fong AZC and Hawley TS. . 1995 Gene Therapy 1: 136–138.

  14. Honda H, Fujii T, Takatoku M, Mano H, Witte ON, Yazaki Y and Hirai H. . 1995 Blood 85: 2853–2861.

  15. Honda H, Oda H, Suzuki T, Takahashi T, Witte ON, Ozawa K, Ishikawa T, Yazaki Y and Hirai H. . 1998 Blood 91: 2067–2075.

  16. Kelliher MA, McLaughlin J, Witte ON and Rosenberg N. . 1990 Proc. Natl. Acad. Sci. USA 87: 6649–6653.

  17. Kelliher MA, Weckstein DJ, Knott AG, Wortis HH and Rosenberg N. . 1993 Oncogene 8: 1249–1256.

  18. Laneuville P, Heisterkamp N and Groffen J. . 1991 Oncogene 6: 275–282.

  19. Langdon WY. . 1995 Aust. N.Z. J. Med. 25: 859–864.

  20. Lauffenburger DA and Horwitz AF. . 1996 Cell 84: 359–369.

  21. Lester LB and Scott JD. . 1997 Recent Prog. Horm. Res. 52: 409–429.

  22. Liu J, Campbell M, Guo JQ, Lu D, Xian YM, Andersson BS and Arlinghaus RB. . 1993 Oncogene 8: 101–109.

  23. Liu J, Wu Y, Ma GZ, Lu D, Haataja L, Heisterkamp N, Groffen J and Arlinghaus RB. . 1996 Mol. Cell. Biol. 16: 998–1005.

  24. Lotti LV, Lanfrancone L, Migliaccio E, Zompetta C, Pelicci G, Salcin AE, Falini B, Pelicci PG and Torrisi MR. . 1996 Mol. Cell. Biol. 16: 1946–1954.

  25. Lugo TG, Pendergast A-M, Muller AJ and Witte ON. . 1990 Science 247: 1079–1082.

  26. Lugo TG and Witte ON. . 1989 Mol. Cell. Biol. 9: 1263–1270.

  27. Maru Y and Witte ON. . 1991 Cell 67: 459–468.

  28. McLaughlin J, Chianese E and Witte ON. . 1987 Proc. Natl. Acad. Sci, USA 84: 6558–6562.

  29. McLaughlin J, Chianese E and Witte ON. . 1989 Mol. Cell. Biol. 9: 1866–1874.

  30. McWhirter JR, Gaalasso DL and Wang JYJ. . 1993 Mol. Cell. Biol. 13: 7587–7595.

  31. McWhirter JR and Wang JYJ. . 1993 EMBO J. 12: 1533–1546.

  32. McWhirter JR and Wang JYJ. . 1997 Oncogene 15: 1625–1634.

  33. Melo JV. . 1996 Blood 88: 2375–2384.

  34. Morgenstern JP and Land H. . 1990 Nucl. Acids Res. 18: 3587–3596.

  35. Nakahara H, Mueller SC, Nomizu M, Yamada Y, Yeh Y and Chen WT. . 1998 J. Biol. Chem. 273: 9–12.

  36. Pawson T and Scott JD. . 1997 Science 278: 2075–2080.

  37. Pear WS, Nolan GP, Scott ML and Baltimore D. . 1993 Proc. Natl. Acad. Sci. USA 90: 8392–8396.

  38. Pendergast AM, Quilliam LA, Cripe LD, Bassing CH, Dai Z, Li N, Batzer A, Rabun KM, Der CJ, Schlessinger J and Gishizky ML. . 1993 Cell 75: 175–185.

  39. Puil L, Liu J, Gish G, Mbamalu G, Bowtell D, Pelicci PG, Arlinghaus R and Pawson T. . 1994 EMBO J. 13: 764–773.

  40. Raitano AB, Whang YE and Sawyers CL. . 1997 Biochim. Biophys. Acta 1333: 201–216.

  41. Ren R, Ye Z and Baltimore D. . 1994 Genes Dev. 8: 783–795.

  42. Renshaw MW, McWhirter JR and Wang JYJ. . 1995 Mol. Cell. Biol. 15: 1286–1293.

  43. Salgia R, Li JL, Ewaniuk DS, Pear W, Pisick E, Burky SA, Ernst T, Sattler M, Chen LB and Griffin JD. . 1997 J. Clin. Invest. 100: 46–57.

  44. Sattler M and Salgia R. . 1997 Cytokine Growth Factor Rev. 8: 63–79.

  45. Sawyers C, McLaughlin J, Goga A, Havlik M and Witte O. . 1994 Cell 77: 121–131.

  46. Sawyers CL. . 1997 Baillieres Clin. Haematol. 10: 223–231.

  47. Schlessinger J. . 1994 Curr. Opin. Genet. Dev. 4: 25–30.

  48. Spector I, Shochet NR, Biasberger D and Kashman Y. . 1989 Cell Motil. Cytoskeleton 13: 127–144.

  49. Tanaka S, Neff L, Baron R and Levy JB. . 1995 J. Biol. Chem. 270: 14347–14351.

  50. Van Etten RA, Jackson P and Baltimore D. . 1989 Cell 58: 669–678.

  51. Van Etten RA, Jackson PK, Baltimore D, Sanders M, Matsudaira PT and Janmey PA. . 1993 J. Cell. Biol. 124: 325–340.

  52. Verfaillie CM, Hurley R, Lundell BI, Zhao C and Bhatia R. . 1997 Acta Haematol. 97: 40–52.

  53. Verfaillie CM, McCarthy JB and McGlabe PB. . 1992 J. Clin. Invest. 90: 1232–1239.

  54. Voncken JW, Kaartinen V, Pattengale PK, Germeraad WT, Groffen J and Heisterkamp N. . 1995 Blood 86: 4603–4611.

  55. Wang JYJ. . 1993 Curr. Opin. Genet. Dev. 3: 35–43.

  56. Wetzler M, Talpaz M, Van Etten R, Hirsh-Ginsberg C, Beran M and Kurzrock R. . 1993 J. Clin. Invest. 92: 1925–1939.

  57. Wu C, Butz S, Ying Y and Anderson RGW. . 1997 J. Biol. Chem. 272: 3554–3559.

  58. Wu Y, Lin J and Arlinghaus RB. . 1998 Oncogene 16: 141–146.

  59. Young JC and Witte ON. . 1988 Mol. Cell. Biol. 8: 4079–4087.

  60. Zhang X and Ren R. . 1998 Blood in press.

Download references


We thank Drs A Bouton, R Salgia, M Lasanti and M Zerial for providing anti-p130Cas, -Crkl, -caveolin, and -Rab7 antibodies, respectively, and X Zhang and A Gross for technical help and critical reading of the manuscript. This work was supported by National Cancer Institute grant CA68008 (to RR). RR is a recipient of American Cancer Society Junior Faculty Research Award.

Author information

Correspondence to Ruibao Ren.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Skourides, P., Perera, S. & Ren, R. Polarized distribution of Bcr-Abl in migrating myeloid cells and co-localization of Bcr-Abl and its target proteins. Oncogene 18, 1165–1176 (1999). https://doi.org/10.1038/sj.onc.1202407

Download citation


  • chronic myelogenous leukemia
  • Bcr-Abl
  • phosphorylation
  • Grb2
  • Shc
  • Cbl

Further reading