The Src Family Kinase Hck Interacts with Bcr-Abl by a Kinase-independent Mechanism and Phosphorylates the Grb2-binding Site of Bcr*

bcr-abl, the oncogene causing chronic myeloid leukemia, encodes a fusion protein with constitutively active tyrosine kinase and transforming capacity in hematopoietic cells. Various intracellular signaling intermediates become activated and/or associate by/with Bcr-Abl, including the Src family kinase Hck. To elucidate some of the structural requirements and functional consequences of the association of Bcr-Abl with Hck, their interaction was investigated in transiently transfected COS7 cells. Neither the complex formation of Hck kinase with Bcr-Abl nor the activation of Hck by Bcr-Abl was dependent on the Abl kinase activity. Both inactivating point mutations of Hck and dephosphorylation of Hck enhanced its complex formation with Bcr-Abl, indicating that their physical interaction was negatively regulated by Hck (auto)phosphorylation. Finally, experiments with a series of kinase negative Bcr-Abl mutants showed that Hck phosphorylated Bcr-Abl and induced the binding of Grb2 to Tyr177 of Bcr-Abl. Taken together, our results suggest that Bcr-Abl preferentially binds inactive forms of Hck by an Abl kinase-independent mechanism. This physical interaction stimulates the Hck tyrosine kinase, which may then phosphorylate the Grb2-binding site in Bcr-Abl.

Bcr-Abl (p210 bcr-abl ), the transforming agent in chronic myeloid leukemia, is the gene product of the bcr-abl hybrid gene, which results from the Philadelphia translocation t(9;22) by fusing parts of the c-abl gene, normally located on chromosome 9, to the bcr gene on chromosome 22 (1,2). Previous studies have demonstrated that Bcr-Abl is a constitutively active tyrosine kinase (3) that has transforming capacity in fibroblasts and hematopoietic cells (4,5) Bcr-Abl-induced transformation seems to require the activation of the Ras signaling pathway (6,7), involving at least two different signaling intermediates, Grb2 and Shc (8 -10). In addition, the association of Bcr-Abl with the SH3-SH2 adaptor protein CRKL (11), the activation of the Jak-STAT pathway (12,13) and of the PI3-Kinase pathway (14 -16), the phosphorylation of a variety of cytoskeletal proteins (17)(18)(19) and the interaction with cytokine and growth factor receptors (20,21) may also play pivotal roles in the pathogenesis of chronic myeloid leukemia. However, the precise mechanisms of transformation by p210 bcr-abl are unknown, and some characteristics of chronic myeloid leukemia, like induction of blast crisis after chronic phase or prolonged viability of chronic myeloid leukemia cells under serum starvation, are still unexplained.
Some critical domains of Bcr-Abl that are necessary for transformation and induction of leukemia have been identified. The coiled-coil oligomerization domain, localized at the N terminus of Bcr-Abl, seems to induce tetramerization of Bcr-Abl, which is in turn necessary for the constitutive activation of the tyrosine kinase of Bcr-Abl, as well as for the complex formation with other Src homology (SH) 1 2-containing proteins (22,23). Further important residues or domains within Bcr are the tyrosine at position 177, which is a binding site for the Ras adaptor protein Grb2 (24), and a SH2-binding motif, alternatively named A-Box and B-Box (25,26), that binds SH2 domains by a phosphotyrosine-independent mechanism. The Nterminal portion of Abl mainly consists of molecular modules with homology to corresponding domains of the tyrosine kinase c-Src, therefore called SH domains 3 and 2 (27). The putative function of these domains is to direct the subcellular localization of Bcr-Abl to compartments where it interacts with specific proteins via specific binding motifs (28). The optimal binding motif for SH3 domains is polyproline (PXXP) (29), whereas SH2 domains predominantly bind to phosphorylated tyrosine residues in a specific amino acid context (30). In addition, the SH3 and SH2 domains of Bcr-Abl seem to regulate the tyrosine kinase activity as well as the transforming capacity of Abl proteins in vivo (26,31,32). As in Src family kinases, a kinase domain (SH1 domain) is located next to the SH2 domain. The C-terminal part includes proline-rich motifs that are the molecular anchor for the adaptor protein CRKL (33,34), a nuclear localization sequence, which is "inactivated" in Abl fusion genes (35), a DNA-binding domain (36), and an actin binding site allowing interaction with the cytoskeleton (37).
We have recently described the activation and association of two members of the Src family of tyrosine kinases, p53/56 lyn and p59 hck , with Bcr-Abl (38). Src kinases are composed of a N-terminal unique domain, a PXXP-binding SH3 domain, a phosphotyrosine-binding SH2 domain, a tyrosine kinase domain, and a C-terminal tail, which is closely involved in negative regulation of the kinase activity (39). One of the common features of Src family kinases seems to be their mechanism of autoregulation. Two cooperative mechanisms negatively regulate the activity of Src family kinases (40): the interaction of the SH3 domain with a polyproline type II helix located between the SH2 domain and the kinase domain and an interaction of the tyrosine phosphorylated C-terminal tail (Tyr 501 in Hck) with the SH2 domain. On the contrary, phosphorylation of a conserved autophosphorylation site within the activation loop of the kinase domain (Tyr 390 in Hck) positively regulates the kinase activity (41). Autophosphorylated Src retains significant activity even if phosphorylated at the negative regulatory tyrosine in the C-terminal tail (42).
This study presents experiments on structural and functional requirements for the interaction of Bcr-Abl with Src kinases. Using several mutants of Bcr-Abl and Hck we were able to demonstrate that the mechanism of interaction of Bcr-Abl with Src-Kinases is independent of Src and Abl kinase activity and is not mediated by any of the known Bcr-Ablbinding motifs. Moreover, the experiments showed that Tyr 177 is phosphorylated by Hck, because coexpression of Hck induced the binding of Grb2 to k.n. Bcr-Abl; this effect was abrogated by a Y177F point mutation.

MATERIALS AND METHODS
Reagents and Antibodies-Reagents for cell lysis were purchased from Sigma Chemicals (Deisenhofen, Germany). SDS-polyacrylamide gel electrophoresis was performed with chemicals provided by Bio-Rad (Mü nchen, Germany) with the exception of acrylamide/bisacrylamide, which was purchased from Boehringer Bioproducts (Ingelheim, Germany). The polyclonal antibodies (Abs) against Hck (N-30), Lyn (44), Bcr (N-20 and 7C6), Abl (K-12), and Grb2 (c-23) and the anti-phosphotyrosine Ab PY20 as well as the corresponding blocking petides were obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). The monoclonal anti-Abl Ab Ab3 was purchased from Oncogene Sciences (Uniondale, NJ). For immunoblotting, all primary Abs were used at 1:1000 dilutions, with the exception of PY20 and anti-Abl Ab3, which were used in a 1:500 dilution. Secondary Abs were either purchased from Bio-Rad (coupled with alkaline phosphatase) or Amersham (coupled with horseradish peroxidase; ECL detection system) and used in dilutions from 1:2000 to 1:5000.
Cloning of bcr-abl Point and Deletion Mutants-The wild type cDNA clone of bcr-abl was provided by Dr. George Daley (Massachusetts General Hospital, Boston, MA) (43). The point mutations Y177F, R1053L, and Y1294F were introduced by site-directed mutagenesis using the Bio-Rad mutaphage kit. For the introduction of the Y177F mutation, a bcr-abl subfragment containing the bcr sequences up to the AccI restriction site was used. This fragment was subsequently recloned into full-length bcr-abl using StuI/NsiI digests. The R1053L and Y1294F mutations were introduced into a KpnI/EcoRI fragment. Thereafter, a KpnI/KpnI fragment from pUCbcr-abl (one KpnI site form pUC, the other at position 3082 in bcr-abl) containing bcr and the abl 5Ј sequences up to the KpnI site was added to reconstitute full-length bcr-abl. All mutated fragments were completely sequenced to confirm the mutations and to exclude additional mutations.
Double and triple mutants of these residues were accomplished by substitution of the DraIII/DraIII fragment of the R1053L mutant with a corresponding fragment containing the Y1294F mutation and/or by substitution of the HindIII/HindIII fragment (first HindIII site from pUC19 and second HindIII site at position 2609 in bcr-abl) of the R1053L and Y1294F single mutants and of the R1053L/Y1294F double mutant with a corresponding fragment containing the Y177F mutation. The A-/B-box deletion (⌬A-/B-box) was introduced by digesting a BsiEI/ SacI subfragment of Bcr-Abl cloned into pUC19 with EcoNI and BglII. Following religation the fragment was sequenced to ensure that the reading frame had been saved. Finally, the mutation was repackaged into full-length bcr-abl using BsaAI/NheI digests. The double mutants bcr-abl k.n./Y177F and bcr-abl k.n./⌬A-/B-box were obtained using the same strategy as used for the generation of the Y177F/R1053L or Y177F/Y1294F double mutants.
Deletion of amino acids 1-223 of bcr-abl was accomplished by PCR mutagenesis using the following primers: 5Ј-ggcgaattcatgggggatgcatccaggcccccttac-3Ј and 5Ј-ccggaattctcattttgaactctgcttaaatccagt-3Ј. The PCR fragment was subcloned into pUC containing an EcoRI/HindIII fragment of wild type bcr-abl using EcoRI/NheI digests. This fragment was converted to full-length by ligating the deleted subfragment with a HindIII fragment containing the missing bcr and abl sequences. For deletion of the noncatalytical C-terminal portion of Bcr-Abl (amino acids 1426 -2031), a similar strategy was used. The primer sequences were 5Ј-cacgccagtcaacagtctggag-3Ј and 5Ј-gggcaggaattctcactgcagcaaggtactcacaga-3Ј. The PCR fragment was subcloned into the KpnI/EcoRI sites of pUC19. This fragment was converted to full-length by ligation of the subfragment with a KpnI fragment containing the residual bcr-abl sequences. All PCR reactions were run with Vent polymerase (New England Biolabs, Beverly, MA). Fragments were sequenced to ensure that no mutations had been introduced during the PCR reactions. All bcr-abl mutants were cloned into pcDNA3 (Invitrogen, Leek, Netherlands) for expression in COS7 cells.
Cloning of hck Point Mutations-The wild type (wt) cDNA clone of hck was purchased from ATCC (Rockville, Maryland). Tyr 501 was mutated to phenylalanin by PCR modification using the following primers: 5Ј-gaatgtgaattcatggggtgcatgaagtccaag-3Ј and 5Ј-tggatagaattctcazggctgctgttgaaactggctctc-3Ј. The resulting full-length hck fragment was subcloned into pUC19 and sequenced subsequently. The mutations K269R and Y390F were accomplished using a two-fragment PCR strategy; wt hck was cloned into a pUC vector that had been modified by deleting one PvuI site. The resulting vector, pUC⌬NdeI/XbaI was used as a template for two PCR reactions producing PCR fragments overlapping at the remaining PvuI site within in the ampicillin resistance (AmpR) gene of pUC and meeting within the hck insert near the triplet to be mutated. Thus, the primers binding in hck allowed us to obtain PCR fragments that could be blunt end-ligated without introducing a deletion; one of these primers contained a mutagenic triplet. Finally, the PCR fragments were PvuI-digested and ligated. For mutation of Lys 269 the following primers were used: 5Ј-gtggcagtgcggacgatgaagccaggg-3Ј and 5Ј-cttggtgtgcttgttgtaggtggc-3Ј. Sequences for primers used for the mutation of Tyr 390 were 5Ј-gacaacgagtttacggctcgggaaggg-3Ј and 5Ј-gtcaatgacccgggccaggccaaagtc-3Ј. The sequences for the primers binding in pUC were 5Ј-ccagccagccggaagggccgagcg-3Ј and 5Ј-ctcttactgtcatgccatccg-3Ј. All hck constructs as well as a wt lyn cDNA were cloned into the EcoRI site of the expression vector pApuro (vector and lyn cDNA provided by Dr. Seth Corey, Children's Hospital of Pittsburgh, PA).
COS7 cells were routinely grown in Dulbecco's modified Eagle's medium containing 4.5 g/liter glucose. For transient transfection, cells of one confluent 175-cm 2 flask were diluted 1:3 and replated into 175-cm 2 tissue culture flasks. 18 -24 h thereafter, cells grown to Ͼ95% confluency were transiently transfected by lipofection using DOTAP (Boehringer, Mannheim, Germany) according to the guidelines of the manufacturer. Briefly, 50 g of bcr-abl cDNA and/or 25 g of hck cDNA were diluted to concentrations of 0.1 g/l and preincubated for 15 min with a 6-fold excess (in g) of DOTAP. For transfection, Dulbecco's modified Eagle's medium containing 1.0 g/liter glucose, 10% fetal calf serum, and antibiotics was used. 24 h after transfection, cells were washed twice in ice-cold phosphate-buffered saline (Life Technologies, Inc., Eggersheim, Germany) and serum-deprived by incubation in Dulbecco's modified Eagle's medium containing 1.0 g/liter glucose and 0.5% fetal calf serum. Transfected cells were normally harvested 48 h after transfection by trypsinization. To protect cells from forming unresuspendible aggregates, 10 g/ml aprotinin was added to the cells immediately after trypsinization.
Cell Lysis-32D cells were lysed in lysis buffer containing 1% Brij97 as described previously (38). For lysis, COS7 cells were washed twice in ice-cold phosphate-buffered saline to remove remaining serum. Thereafter, cells were lysed in lysis buffer containing 1% Nonidet P-40, 20 mM Tris (pH 8.0), 50 mM NaCl, and 10 mM EDTA as well as 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 10 g/ml leupeptin, and in most cases 2 mM sodium orthovanadate. In general, pelleted cells from one 175-cm 2 tissue culture flask (about 5 ϫ 10 7 cells) were resuspended in 500 l of lysis buffer solution and incubated on ice for 25 min. Thereafter unsoluble material was removed by centrifugation at 15,000 ϫ g. Afterward lysates were checked for protein concentrations using a Bio-Rad protein assay.
Immunoprecipitation and Immune Complex Kinase Assay-For immunoprecipitation (IP), 150 l of COS7 cell lysate was diluted by the addition of 450 l of incubation buffer containing 20 mM Tris (pH 8.0), 50 mM NaCl, and 10 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 10 g/ml leupeptin, and 2 mM sodium orthovanadate to inhibit phosphatase activity where desired. Lyn, Hck, and Bcr-Abl were precipitated by adding 5 g of the appropriate Abs, i.e. anti-Lyn 44 for precipitation of Lyn, anti-Hck N-30 for precipitation of Hck, and either anti-Bcr 7C6 or anti-Abl K-12 for precipitation of Bcr-Abl. IP reactions were incubated overnight at 4°C on a rotating plate. After 18 h of incubation, 125 l of Sepharose A beads (Pharmacia Biotech Inc., Freiburg, Germany) diluted 1:1 in IP buffer (0, 1% Nonidet P-40, 20 mM Tris (pH 8.0), 50 mM NaCl, and 10 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 10 g/ml leupeptin, and 2 mM sodium orthovanadate) were added to each sample. Following an additional 2 h of incubation at 4°C, the precipitates were washed three times with IP buffer and subsequently boiled in 2ϫ sample buffer before loading on SDS gels. Peptide blocking experiments were performed as described previously (38).
For immune complex kinase assays of Src kinases precipitated from 32D cells, cell lysis, and the IP protocol were slightly modified; IP incubation periods were reduced to 3 h, and three times of washing with IP buffer were followed by washing the precipitates one time with kinase buffer (50 mM Tris (pH 7.4), 10 mM MnCl 2 ). Kinase reaction and analysis of autophosphorylation was performed as described (38).
Gel Electrophoresis and Immunoblotting-Gel electrophoresis and immunoblotting were performed using standard methods. Proteins were either transfered to polyvinylidene difluoride membranes (Millipore, Eschborn, Germany) or nitrocellulose (Schleicher & Schü ll, Dassel, Germany). Immunoblots with PY20 were developed by using alkaline phosphatase-conjugated secondary Abs at a dilution of 1:2000 in Tris-buffered saline containing 5% bovine serum albumin when polyvinylidene difluoride membranes were probed. For detection of phosphorylated proteins transferred to nitrocellulose or other membranes, secondary horseradish peroxidase-conjugated Abs were used. The ECL detection system was used according to the guidelines of the manufacturer (Amersham, Braunschweig, Germany).
Protein Dephosphorylation-Dephosphorylation was achieved by omitting orthovanadate, a potent phosphatase inhibitor, from the lysis buffer, followed by preincubation of cleared lysate at 4°C for 24 h prior to IP.

Complex Formation of Hck Kinase with Bcr-Abl in COS7
Cells-We have recently described the association and activation of two kinases of the Src family, p53/56 lyn and p59 hck , with or by Bcr-Abl (38). The mechanism and function of this inter-action is unknown. This led us to investigate the structural and functional requirements for the complex formation of Bcr-Abl with Hck kinase, a Src family member preferentially expressed in hematopoietic cells. To establish an expression system that would allow the rapid screening of Bcr-Abl and Hck mutants, wt full-length cDNAs of bcr-abl and hck were cloned into appropriate mammalian expression vectors (pcDNA3 or pApuro) and prepared for transient transfection into COS7 cells. cDNAs were introduced into these cells using lipofection (see "Materials and Methods"). To demonstrate that the expressed kinases were active in vivo, the blot was stripped and reblotted with an anti-phosphotyrosine Ab, PY20. Fig. 1B shows that both kinases were highly (auto)phosphorylated and that expression of these kinases resulted in an increased overall phosphotyrosine content in cellular proteins. Because both kinases, Bcr-Abl and Hck (lanes 2 and 3), seemed maximally activated, no synergism was detectable by antiphosphotyrosine blotting when Bcr-Abl and Hck were coexpressed.
We then wished to demonstrate that Bcr-Abl and Src kinases were found in a complex in COS7 cells, similar to our previous findings in myeloid cells (38). For this purpose, lysates of transiently transfected COS7 cells coexpressing Bcr-Abl and Hck were subjected to immunoprecipitation with the polyclonal anti-Abl Ab K-12. Subsequent immunoblotting with the anti-Hck Ab N-30 demonstrated that Hck (Fig. 1C, lane 1) formed a complex with Bcr-Abl. Coprecipitation was completely blocked by the addition of corresponding blocking peptide, indicating that the coprecipitation of Hck with Bcr-Abl was not caused by unspecific binding (lane 2). Similar results were obtained when Bcr-Abl was precipitated using the polyclonal anti-Bcr Ab 7C6 (not shown). Moreover, we could also coprecipitate Bcr-Abl in anti-Hck IPs (not shown). Finally, similar results were obtained in cotransfection and coprecipitation experiments with Lyn and Bcr-Abl in COS7 cells (not shown).

Mutations of Several Known Binding Motifs of Bcr-Abl Do Not Disrupt Complex Formation with Hck
Kinase-Several domains have been described as functionally relevant protein interaction modules of Bcr-Abl ( Fig. 2A) (8, 45, 46). To investigate which of these regions were necessary for the interaction of Bcr-Abl with Src kinases, several mutations were introduced into bcr-abl cDNAs ( Fig. 2A). Surprisingly, none of these mutations introduced into bcr-abl alone or in combination was able to disrupt or diminish the formation of Bcr-Abl-Hck complexes in COS7 cells (Fig. 2D). Fig. 2B shows a typical example of such an experiment. The hck gene was cotransfected into COS7 cells either in combination with a control vector (lane 1) or with various single, double, or triple bcr-abl mutants containing amino acid substitutions recently shown to be critical for Ras activation by Bcr-Abl (8) essential for SH2-directed binding to phosphorylated tyrosines; and Tyr 1294 , a major autophosphorylation site located in the Abl kinase domain. Lysates of transiently transfected cells were used for anti-Abl IPs. Subsequent immunoblot analysis demonstrated that considerable amounts of Hck were coprecipitated with wt Bcr-Abl (Fig. 2B, middle panel, lane 2). However, similar amounts of Hck were found in complexes with all single, double, and triple mutants (lanes 3-9). Similar results were obtained when Lyn was coexpressed with these mutants (not shown). Anti-Grb2 immunoblots showed that the association of Bcr-Abl with the Ras adaptor protein Grb2 was constantly disrupted by the Y177F mutation (Fig. 2B, bottom  panel, lanes 3, 6, 7, and 9). Anti-Abl immunoblots demonstrated that equal amounts of Bcr-Abl were precipitated from Bcr-Abl-expressing cells (Fig. 2B, top panel). Finally, control blots of whole cell lysates showed that comparable amounts of Hck and Grb2 were expressed in the different cells used for the experiment (Fig. 2C). Longer exposure of films during chemoluminiscence detection also revealed that small amounts of Hck could be coprecipitated with endogenous c-Abl (faintly visible in Fig. 2B, lane 1). In addition, the deletion of the coiled-coil oligomerization domain (amino acids 1-223), the complete deletion of an SH2-binding motif in bcr (A-/B-Box), and the truncation of the noncatalytical C terminus, which included binding sites for the SH2-SH3 adaptor protein CRKL and an actinbinding region, did not interfere with Bcr-Abl-Src binding (Fig.  2D). In conclusion, these results suggest that the interaction of Bcr-Abl and Src kinases is not mediated by any of the known protein interaction modules of Bcr-Abl. Moreover, the coprecipitation of Hck kinase with Bcr-Abl ⌬1-223 lacking the oligomerization site necessary for the constitutive activation of the Bcr-Abl tyrosine kinase (Fig. 2D), as well as the complex formation of Hck kinase with endogenous c-Abl that normally has only low kinase activity in COS7 cells, suggested that the tyrosine kinase activity of Bcr-Abl might not be indispensable for this interaction.
Dephosphorylation Induces Complex Formation between Bcr-Abl and Hck-To investigate whether Bcr-Abl-Hck complex formation was due to binding of the SH2 domain of Hck to phosphorylated tyrosine residues in Bcr-Abl (similar to the binding mechanism of Grb2 to Tyr 177 ) or vice versa, immunoprecipitations of Bcr-Abl from COS7 cell lysates coexpressing Bcr-Abl and Hck were performed under conditions allowing dephosphorylation of cellular proteins. Dephosphorylation of cell lysates was achieved by omitting orthovanadate, a phosphatase inhibitor, from the lysis buffer, followed by preincubation of the cleared lysate at 4°C for 24 h prior to IP. After immunoprecipitation and SDS-polyacrylamide gel electrophoresis, the precipitates were assayed by immunoblotting for co-precipitating Hck and Grb2. Dephosphorylation of the lysate resulted in nearly complete abrogation of the binding of Grb2 to Bcr-Abl (Fig. 3A, right panels, anti-Grb2 blot, lane 2). In marked contrast to Grb2, Hck was still found in a complex with Bcr-Abl, and the amount of coprecipitated Hck was even enhanced by dephosphorylation (Fig. 3A, right panels, anti-Hck  blot, lanes 1 and 2). Similar results were obtained with dephosphorylation of Hck by the addition of potato acid phosphatase (not shown). Anti-Abl blotting of the precipitates indicated that similar amounts of Bcr-Abl were precipitated from both lysates (Fig. 3A, upper panels). In addition Bcr-Abl, Hck, and Grb2 blotting showed that about equal amounts of Bcr-Abl, Hck, and Grb2 were detectable in both lysates. Control experiments (Figs. 3, B and C) were performed to demonstrate that dephosphorylation of the signaling proteins involved was nearly complete. Bcr-Abl (Fig. 3B) and Hck (Fig. 3C) were precipitated from the lysates used in the experiment of Fig. 3A. Bcr-Abl and Hck blots (left panels) demonstrated that similar amounts of each protein were precipitated from the lysates using the indicated Abs. Stripping and reblotting with anti-phosphotyrosine Abs revealed that the content of phosphorylated tyrosine residues in Bcr-Abl and in Hck was substantially reduced in the dephosphorylated lysates. Similar observations were made in the bcr-abl positive human K562 cell line, where dephosphorylation enhanced the binding of Bcr-Abl to the Src family kinase Lyn (not shown). In conclusion, the results suggested that the complex formation of Src kinases and Bcr-Abl was strongly enhanced by dephosphorylation. Moreover, the interaction of Bcr-Abl and Src kinases was probably not mediated by binding of a SH2 domain to phosphotyrosine residues.
Enhanced Binding of Kinase Inactive Mutants of Hck to Bcr-Abl-The regulation of the activity of Src kinases critically depends on the phosphorylation of at least two distinct tyrosine residues (41). In Hck, C-terminal phosphorylation of Tyr 501 induces a "closed" kinase inactive conformation, whereas phosphorylation of Tyr 390 is necessary for full activation of the catalytic domain. Therefore, we hypothesized that the enhanced binding of Bcr-Abl to dephosphorylated Hck was due to dephosphorylation of one of these two important tyrosine residues.
To test these hypotheses, several activating (Y510F) and inactivating (K269R and Y390F) point mutations were introduced into Hck (Fig. 4A): K269R to delete the ATP-binding site of the kinase; Y390F to delete the positive regulatory autophosphorylation site of the catalytic domain; and Y501F to delete the C-terminal negative regulatory phosphorylation site. Finally, the Y501F mutation was combined with the K269R mutation. All four mutants were coexpressed with wt Bcr-Abl in COS7 cells (Fig. 4B). Bcr-Abl was precipitated from the lysates using anti-Bcr antisera (7C6). Anti-Bcr blotting showed that similar amounts of Bcr-Abl were precipitated (Fig. 4B, upper  panel, lanes 1-6). Again, wild type Hck was found to co-precipitate with Bcr-Abl (Fig. 4B, lower panel, lanes 1 and 4). Coprecipitation was not enhanced with the Hck Tyr 501 mutant, suggesting that "dephosphorylation" of this negative regulatory tyrosine residue was not important for the binding to Bcr-Abl. In marked contrast, co-precipitation of all three inactivating Hck mutants (K269R, Y390F, and K269R/Y501F) with Bcr-Abl was significantly increased (Fig. 4B, lanes 2, 3, and 6). Control blots of cell lysates with Hck antisera showed that similar amounts of each mutant were expressed (Fig. 4C). These results strongly suggested that Bcr-Abl preferentially bound inactive forms of Hck kinase.
The Abl Kinase Is Not Required for Bcr-Abl-Hck Complex Formation in COS7 Cells-We asked next whether the Abl kinase was required for the interaction of Bcr-Abl with Hck. Therefore, we expressed wt Bcr-Abl or a k.n. mutant of Bcr-Abl (K1172R) with either wt or k.n. (K269R) Hck in COS7 cells. Immunoprecipitations with anti-Abl Abs showed that similar levels of wt Hck were found in complex with wt and k.n. Bcr-Abl (Fig. 5A, lower panel, lanes 1 and 3). As observed above (Fig. 4B), co-immunoprecipitation of both wt and k.n. Bcr-Abl was stronger with k.n. Hck than with wt Hck (Fig. 5A, lower  panel). Peptide controls for all precipitations demonstrated the specifity of the precipitating antibodies used (Fig. 5A, lanes 2,  4, 6, and 8). Similar amounts of Hck were expressed in all cell lines (Fig. 5B). Taken together, the results indicated that the complex formation of Hck with Bcr-Abl was enhanced by inactivating the Hck kinase but not the Abl kinase. Furthermore, the kinase activity of Bcr-Abl was not necessary for the complex formation with Hck kinase.

Bcr-Abl Induces Activation of Src Kinases Lyn and Hck in 32D Cells by a Kinase-independent Mechanism-Experiments
on the activation of Src kinases Hck and Lyn by Bcr-Abl were difficult in COS7 cells, because both kinases showed a relatively high constitutive activation level in these cells. To address this question, we therefore had to use the murine, interleukin-3-dependent hematopoietic cell line, 32D, which shows a lower activation of Src kinases Hck and Lyn than COS7 cells. In these experiments Src kinases Lyn and Hck were immunoprecipitated from 32D cells transfected with wt or k.n. bcr-abl. (K1172R). Fig. 6A shows a representative immune complex kinase assay for the Lyn kinase. Compared with control cells, both wt and k.n. Bcr-Abl induced a substantial increase of autophosphorylation of Lyn (Fig. 6A). Blots of aliquots of the IP reactions and of lysates indicated that equal amounts of precipitated Lyn were used for the assay (Fig. 6B) and that comparable amounts of Lyn were expressed in the different 32D sublines (Fig. 6C). Immune complex kinase assays with precipitated Bcr-Abl from the same cell lysates demonstrated that the k.n. Bcr-Abl had a dramatically reduced kinase activity (Fig.  6D). Again, anti-Abl immunoblotting of IP aliquots used for the assay showed that equal amounts were precipitated (Fig. 6E). Comparable results were obtained with Hck, although Hck expression was slightly lower than Lyn expression in 32D cells (not shown). In conclusion, these results further support an Abl kinase-independent mechanism for the activation of Src kinases (Hck and Lyn).
Hck Induced Association of Bcr-Abl with Grb2-Preliminary results indicated that substantial amounts of Grb2 were found in complex with Bcr-Abl, even in 32D cells expressing k.n. Bcr-Abl (K1172R), despite the fact that the docking site for Grb2, Tyr 177 in Bcr, is thought to be (auto)phosphorylated by the Abl kinase. Because we had observed that k.n. Bcr-Abl activated Src kinases (Fig. 6), we hypothesized that Src kinases might phosphorylate Tyr 177 in k.n. Bcr-Abl, thus creating a binding site for Grb2. To demonstrate that Hck induced phosphorylation of k.n. Bcr-Abl, the blot from a previous experiment (Fig. 5) was reprobed using the anti-phosphotyrosine Ab PY20. Fig. 7A shows that the content of phosphorylated tyrosine was dramatically reduced in k.n. Bcr-Abl (K1172R) as compared with wt Bcr-Abl (compare lanes 1 and 3 with lanes 5 and 7). However, k.n. Bcr-Abl showed a slightly increased phosphotyrosine staining, when wt Hck (but not k.n. Hck) was coexpressed, suggesting that Hck was able to induce some phosphorylation of k.n. Bcr-Abl. The same blot then was reanalyzed by anti-Grb2 blotting. As shown in the lower panel of Fig. 7A, coexpression of wt Hck induced a severalfold increase of the amount of Grb2 bound to k.n. Bcr-Abl. A larger amount of Grb2 co-precipitated with wt Bcr-Abl, indicating that Hck did not completely substitute for the Abl tyrosine kinase activity. In control blots, equal amounts of Hck and Grb2 were detected in the COS7 cell lysates used (Fig. 7B).
The SH2-binding Motif of Bcr Is Not Required for the Effects of Hck Kinase on Tyr 177 Phosphorylation and Grb2 Binding-To study the effects of Hck on Tyr 177 phosphorylation in k.n. Bcr-Abl in more detail, the K1172R mutation was combined with the Y177F point mutation or the deletion of the SH2-binding motif (A-/B-Box). With these mutants, we wished to determine whether any of these regions was important for Tyr 177 phosphorylation by Hck kinase (Fig. 8A). Previous studies indicated that isolated SH2 domains of Src kinases are able to bind to Bcr via this SH2-binding motif (A-/B-Box) in vitro (25). We therefore speculated that the binding of the Hck SH2 domain to the A-/B-Box motif could bring the Hck kinase domain into proximity with Tyr 177 , thus allowing its transphosphorylation. Fig. 8B shows an immunoblot of COS7 cells transiently transfected with these mutants. Anti-Bcr blotting of precipitated k.n. Bcr-Abl showed that similar amounts of Bcr-Abl were purified from each lysate (Fig. 8B, top panel). Only little Grb2 was detected by anti-Grb2 blotting when wt Hck was not co-expressed (Fig. 8B, bottom panel, lane 1). The expression of wt Hck induced a severalfold increase of the amount of Grb2 coprecipitating with Bcr-Abl (Fig. 8B, bottom panel,  lane 2). The association of Grb2 with Bcr-Abl was completely abolished in the Y177F mutant (Fig. 8B, bottom panel, lane 3). In marked contrast, the deletion of the A-/B-Box motif had no influence on the amount of Grb2 bound to Bcr-Abl, indicating that this region was not essential for the phosphorylation of Tyr 177 by Hck (Fig. 8B, bottom panel, lane 4). In contrast to Grb2, Hck was found to coprecipitate with all three Bcr-Abl constructs (Fig. 8B, middle panel), suggesting that Tyr 177 acted as a substrate but not as binding site for Hck kinase. Finally, the blot was reprobed with anti-phosphotyrosine Ab to investigate the influence of the different Bcr-Abl mutants on the tyrosine phosphorylation by Hck (Fig. 8C). Again, a slight  increase in tyrosine phosphorylation of k.n. Bcr-Abl was detectable when Hck was coexpressed (compare lanes 1 and 2). However, we did not find any decrease in Hck-induced phosphorylation in the two double mutants, suggesting that Tyr 177 may not be the only tyrosine residue to be phosphorylated by Hck. Fig. 8D demonstrates that equal amounts of Bcr-Abl and Grb2 were expressed in all four cell lines and that no Hck was expressed in COS7 cells not transfected with the hck cDNA. DISCUSSION We have recently reported the interaction of Bcr-Abl with two Src family kinases, p53/56 lyn and p59 hck in myeloid cells (38). In this manuscript, structural requirements and functional consequences of the complex formation of Bcr-Abl with Hck and Lyn kinase were investigated. Our results suggest that Hck (and Lyn) kinase might cooperate in phosphorylating a binding site for Grb2, Tyr 177 , in Bcr. Although we have not yet been able to identify the domain(s) necessary for the interaction of Bcr-Abl with Src kinases, experiments with dephosphorylated proteins (Fig. 3) and inactivating mutants of Hck (Figs. 4 -6) gave some clues concerning potential mechanisms for this association. Src kinases are thought to be regulated by phosphorylation at distinct tyrosine residues (47). Phosphorylation of Tyr 390 is necessary for full activation of Hck, whereas phosphorylation of Tyr 501 is thought to negatively regulate the kinase activity by intramolecularly stabilizing an inactive conformation (39,40). In this regard, the demonstration that dephosphorylation of cellular proteins prior to IP increased Bcr-Abl-Hck complex formation seemed particularly interesting. This observation led us to ask whether inactive Hck kinase could interact with Bcr-Abl. Therefore, we performed additional experiments using two mutations of the Hck kinase, Y390F, inactivating a critical regulatory autophosphorylation site in the kinase domain, and K269R, inactivating the catalytic ATP-binding site. Both mutants showed enhanced binding to Bcr-Abl when compared with wt Hck, confirming that Bcr-Abl preferentially bound to inactive forms of Hck. In addition, both binding and activation of Src kinases Hck and Lyn by Bcr-Abl was shown to be Abl kinase-independent, suggesting that the complex formation with Bcr-Abl alone was sufficient for activation of Src kinases. A similar mechanism has previously been reported for the activation of Lck following the complex formation with Syk in T cells (48).
The comparison of the crystallographic analysis of inactive and active forms of Src kinases has revealed that major structural differences are solely found within the kinase domain of different Src kinases (40,49,50). This suggests that the affinity of Src kinases for Bcr-Abl may be regulated by activation statedependent conformational differences in this domain and that at least one interaction site in Src kinases may be the kinase domain itself, similar to the mechanism proposed for binding of Src kinases to Polyomavirus middle T antigen (51). Binding of Bcr-Abl to the kinase domain of Src kinases may then induce an alteration of the orientation of certain amino acid residues necessary for catalytic activity, finally leading to increased activity and enhanced autophosphorylation of Src kinases. A similar mechanism has previously been proposed for the induction of kinase activity of Cdk2 by Cyclin A (52,53).
One might ask whether Src kinases are also activated by endogenous c-Abl, because the Abl tyrosine kinase activity does not seem to be necessary for Src kinase activation. Regarding this, it is important to recall that Bcr does not only alter the kinase activity of Abl but also its subcellular distribution. Although c-Abl is normally found in both the nucleus and the cytoplasm, the fusion to Bcr seems to favor its cytosolic localization (54), thus bringing it in closer proximity to Src kinases. The transfer of the (Bcr)Abl kinase to the cytoplasm is sup-posed to induce new signaling events, including the activation of (membrane-associated) Src kinases.
In addition to the evidence for a physical interaction of Bcr-Abl and Src kinases, our results indicate that these kinases may also functionally cooperate with each other. Co-expression of Hck with a k.n. mutant of Bcr-Abl resulted in increased tyrosine phosphorylation of Bcr-Abl. Further analysis of this Hck-induced phosphorylation of Bcr-Abl demonstrated that one tyrosine to be phosphorylated by Hck was Tyr 177 , resulting in increased binding of Grb2 to Abl kinase negative Bcr-Abl. A similar effect has been reported for the coexpression of Bcr and Fps/Fes tyrosine kinases in Sf9 insect cells by Maru et al. (55). In some contrast to the findings of Maru and colleagues, our kinase negative double mutants containing either a point mutation of Tyr 177 or a deletion of amino acids 190 -412 (comprising a total of 8 tyrosine residues) barely showed any decrease in tyrosine phosphorylation, strongly suggesting that multiple tyrosine residues within Bcr-Abl might be substrates for Src kinases. Although it is unclear whether Src kinases are necessary for the phosphorylation of tyrosine residues in wt Bcr-Abl in vivo, the phosphorylation of k.n. Bcr-Abl by activated Src kinases may explain some so far unexplained findings in 32D cells expressing high levels of k.n. Bcr-Abl. In these cells, we found substantial amounts of Grb2 in complex with k.n. Bcr-Abl, an observation that contrasts the current opinion that Tyr 177 is an autophosphorylation site of the Abl tyrosine kinase. Because Src kinases were activated in these cells by an Abl kinase-independent mechanism, they were likely to substitute for the Abl kinase activity in phosphorylating Tyr 177 . Whether Src kinases cooperate with Bcr-Abl in activating the Ras signaling pathway remains to be determined.
Taken together, the results imply that Bcr-Abl preferentially binds inactive Hck molecules. One might hypothesize that this interacton leads to the activation of Hck by altering the threedimensional structure of Hck. Once activated, Hck may loose its affinity for its "activation site" in Bcr-Abl and become oriented toward its substrates, one of which is located in Bcr-Abl itself. Further experiments will now have to elucidate the precise mechanism of activation of Src kinases by Bcr-Abl and the pathophysiological role of this interaction in Philadelphia chromosome positive leukemias.