|Home | About | Journals | Submit | Contact Us | Français|
Here we report c-Abl kinase inhibition mediated by a phosphotyrosine located in trans in the c-Abl substrate, Abi1. The mechanism, which is pertinent to the nonmyristoylated c-Abl kinase, involves high affinity concurrent binding of the phosphotyrosine pY213 to the Abl SH2 domain and binding of a proximal PXXP motif to the Abl SH3 domain. Abi1 regulation of c-Abl in vivo appears to play a critical role, as demonstrated by inhibition of pY412 phosphorylation of the nonmyristoylated Abl by coexpression of Abi1. Pervanadate-induced c-Abl kinase activity was also reduced upon expression of the wild type Abi1 but not by expression of the Y213 to F213 mutant Abi1 in LNCaP cells, which are naturally deficient in the regulatory pY213. Our findings suggest a novel mechanism by which Abl kinase is regulated in cells.
The ubiquitous nonreceptor tyrosine kinase, c-Abl kinase, plays an essential role in cell signal transduction, balancing events leading to apoptosis or to increased cell proliferation [1, 2]. The critical role of c-Abl kinase in cell proliferation is illustrated by the dramatic manifestation of chronic myelogenous leukemia (CML) due to expression of BCR-Abl, a kinase-activated mutant form of c-Abl tyrosine kinase [3, 4]. Rational approaches to curtail BCR-Abl kinase activity led to the development of STI-571 (imanitib mesylate, or Gleevec)  as the successful treatment of CML . However, the appearance of Gleevec-resistance mutations [7, 8], that becomes an issue in advanced stage CML  led to re-thinking of the mechanism of BCR-Abl regulation. This has brought attention back to c-Abl, since BCR-Abl and c-Abl share most regulatory domains [2, 10].
Autoinhibition has emerged as the mechanism of regulation of c-Src and c-Abl [11, 12]. These kinases share high structural homology conferred by the presence of highly conserved structural domains: SH3, SH2, and the catalytic domain. Crystal structures of c-Src [13, 14] and c-Abl [15, 16] revealed that SH3 and SH2 domains bind to the catalytic domain (CD) inducing an autoinhibitory conformation, which provides the basic mechanism of regulation of these kinases. This basic regulation is preserved in BCR-Abl, which retains the c-Abl SH2 and SH3 domains.
c-Src and c-Abl differ from each other in two mechanisms that inhibit activities of these kinases. In c-Src, inhibition is achieved by intramolecular interaction of the SH2 domain with the phosphorylated tyrosine 527 located in the C-terminal region of the same molecule . In c-Abl there is no internal phosphotyrosine-SH2 domain interaction, precluding this inhibitory mechanism. Additional inhibitory constraints are imposed on c-Abl both by the myristoylated cap which binds directly to the C-terminal lobe of the kinase domain, and by the cap region phosphoserine 69 which binds to the SH2 domain . These interactions further serve to lock the SH3-SH2 “clamp” onto the catalytic domain. The myristoyl group [15, 16], or small compounds mimicking its action , stabilize the position of the C-terminal helix of the catalytic domain, αI, resulting in the inhibited conformation of the kinase. The “molecular lock” imposed by the myristoylated cap, however, does not exist in the nonmyristoylated form of c-Abl, isoform 1a, which contains only a partial cap region, or in BCR-Abl, in which the cap region is replaced by BCR. Thus, the kinase activities of BCR-Abl, and c-Abl-1a, are not regulated by the myristoylated cap although the phosphoserine 69 is preserved in c-Abl-1a, where it may contribute to the autoinhibitory mechanism . Intramolecular interactions of the cap region may also regulate accessibility, and therefore, may regulate binding of Abl SH3 or SH2 ligands including phosphotyrosine-containing peptides from growth factors, which may play a role in the myristoylated c-Abl kinase activation . Considering the complexity of Abl regulation, activation of the kinase activity is likely to involve multiple steps leading to uncoupling of SH3 and SH2 domains from the catalytic domain thereby “freeing” the kinase from inhibition. Although not yet demonstrated with Abi proteins, peptides that combine both Abl SH3 and SH2 binding sites into a single consolidated ligand show enhanced binding affinities for the dual SH3-SH2 domain, pointing to the possibility that these domains may act as one functional unit in c-Abl .
Various proteins, including Abi1 and Abi2, that bind to c-Abl kinase have been proposed to be c-Abl co-inhibitors [12, 21, 22]. Abi1 and Abi2 have been thought to play a role in the regulation of cell growth [21-23] but the molecular mechanism is not clear. Abi1 and Abi2 were proposed to regulate c-Abl kinase activity by interaction with C-terminal PXXP sequences [21, 22], and through interaction with the c-Abl SH3 domain . No SH3- or SH2-based mechanism of c-Abl kinase regulation, however, has been demonstrated for Abi proteins. The LNCaP prostate tumor cell line contains a mutation in the Abi1 gene that results in deletion of exon 6 . Exon 6 of Abi1 is within the SH3 domain-binding region  pointing to the possibility that this region might be critical for c-Abl kinase regulation.
We have further investigated regulation of c-Abl kinase by Abi1 and here report the discovery of a novel allosteric mechanism of inhibition of nonmyristoylated c-Abl kinase mediated by peptides derived from the Abl SH3- and SH2-binding regions of Abi1. Mutations in either binding site can abrogate the capacity to inhibit Abl kinase activity. These data are consistent with observed enhanced binding affinity of the consolidated Abi1 ligand for the dual Abl SH3-SH2 domain over single, i.e. SH3 or SH2, domain Abi1-derived ligands. We propose that Abi1 plays critical role in regulating Abl kinase activity in cells.
See Figure 1 for diagrams of peptides. All peptides were synthesized commercially. Anti-pY213 polyclonal antibody was produced to peptide pY213, and affinity purified using the phosphopeptide-specific column followed by absorption on the nonphosphopeptide (Y213) column. Polyclonal and monoclonal HA antibodies were from Covance (Berkeley, CA) and Roche Diagnostic Corporation (Indianapolis, IN). Antibodies to c-Abl were 8E9 (BD Biosciences, San Jose, CA), K12 (Santa Cruz Biotechnology, Santa Cruz, CA), and pY412 (Biosource International, Camarillo CA). Antibodies to Crk were from BD Biosciences, San Jose, CA (mouse monoclonal), Santa Cruz Biotechnology, Santa Cruz, CA (rabbit polyclonal), and Cell Signaling Technology (phospho-Crk pY221). Polyclonal antibody, Ab-2, to Abi1 was described previously . Monoclonal antibody, 7B6, to Abi1 was produced to recombinant Abi1. The epitope bound by this antibody is identical to that bound by Mab 4E2 . Antibody to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was from Imgenex Corporation (San Diego, CA). GST antibody was from Zymed (San Francisco, CA). GFP antibody was from Invitrogen (Carlsbad, CA). Generic antibody to phosphotyrosine, PY99, was from Santa Cruz Biotechnology, Santa Cruz, CA.
His-tagged, partially capped, active (nonmyristoylated) c-Abl, E46 through C-terminus (isoform 1b numbering) was produced in baculovirus from plasmid (a kind gift of Tony Koleske, Yale University, New Haven, CT) and purified as described  following treatment of insect cells with 30 μM STI-571 (Novartis Pharma AG, Basel, Switzerland) for 48 hrs prior to cell lysis. The expressed protein was affinity purified on nickel-nitriloacetic acid agarose, washed to remove inhibitor, and subsequently purified by ion-exchange chromatography using a Mono S column (Amersham Biosciences, Piscataway, NJ). GST fusions of c-Abl SH3 and SH2 domains and the SH2 variant containing an R171K mutation were obtained from Bruce Mayer (University of Connecticut Health Center, Farmington, CT). For use in fluorescence quenching experiments the dual domain SH3-SH2 polypeptide of c-Abl was expressed from plasmid pTXB1 (New England Biolabs, New England, CT) in E. coli BL21 cells. The recombinant fusion protein was purified through chitin affinitive binding (New England Biolabs, New England, CT). After DTT cleavage the SH3-SH2 domain was further purified by SP Sepharose (GE Healthcare, Piscataway, NJ) cation exchange.
Wild type or mutant Abi1 (GenBank Accession No. NM 005470 and U87166) isoform 2, residue numbering according to  were expressed from plasmids. The mutant Abi1-F213 contains a Y213F replacement. At residues 181-185 the mutant Abi1-Pro replaces the sequence AESEA with PPSPP, which results in the loss of a PXXP SH3 binding motif. All Abi1 cDNAs were subcloned into the pEGFP-N2 plasmid (Clontech, Mountain View, CA) following removal of GFP-encoding sequences and introduction of an HA tag at the C-terminus. Untagged wild type isoform 2 of Abi1 was also used for transfections. In vitro translation of the N-terminus of Abi1 was performed as described . The C-terminal GFP fusion of the nonmyristoylated c-Abl (isoform 1a) was obtained from Bruce Mayer.
Measurement of kinase activity was essentially as described in , using biotinylated model substrate peptide GGEAIYAAPFKK, [27, 28] and 32P-γ-ATP. SAM2 streptavidin-coated membrane (Promega Corporation, Madison WI) was used to capture the substrate. Kinase assays were carried out in kinase buffer (50 mM Tris-HCl pH 7.5, 10 mM MgCl2, 1 mM EGTA, 2 mM dithiothreitol, 0.01 % Brij 35, 100 μM ATP) along with 2nM Abl kinase, substrate peptides, and Abi1 ligand peptides as indicated. Reactions were carried out for 5 min. at 30°C. To evaluate c-Abl kinase activity in LNCaP cell lines, cells were treated with 0.1 mM sodium pervanadate for 10 min. prior to cell lysis; and the kinase was immunoprecipitated from lysed cells. c-Abl kinase activity was evaluated by measuring 1) phosphorylation levels of activation loop tyrosine 412, 2) total tyrosine phosphorylation, and 3) phosphorylation of endogenous Abl substrate Crk.
Mass spectrometric analyses of GST-Abi1 peptides were performed using an Applied Biosystems (Foster City, CA) Voyager DE MALDI mass spectrometer. Spectra were calibrated against an external or internal standard as needed.
LNCaP and Cos7 cells (ATCC, Rockville, MD) were maintained according to ATCC protocols. Co-transfections of Abi1 with c-Abl in Cos7 cells were performed with the isoform 1a of c-Abl (nonmyristoylated) and isoform 2 of Abi1 using Lipofectamine Plus Reagent (Invitrogen, Carlsbad, CA). At 22 h post-transfection, cells were processed for immunoprecipitation as described  following treatment with 10 μM Gleevec for 30 min. LNCaP cell lines stably expressing either wild type, clone Abi1(+), HA-tagged Abi1 isoform 2, or HA-tagged mutants of Abi1 isoform 2 were obtained using G418 selection (Invitrogen, Carlsbad, CA).
c-Abl tyrosine kinase was activated by treatment of LNCaP cells for 10 min with 0.1 mM sodium pervanadate (freshly prepared from 100 mM activated sodium orthovanadate and 100 mM H2O2),  prior to lysis. Immunoprecipitation was performed as described . Western blotting and overlay binding assay to quantify Abl SH3 domain binding were performed as described . All blots were developed using Supersignal West Pico Chemiluminescence Substrate (Pierce Biotechnology, Rockford, IL). Images were acquired using a Kodak GL 440 Imaging System and quantified using Kodak 1D Image Analysis Software (Version 3.6.4).
Surface plasmon resonance was performed using a Biacore 3000 instrument (BIAcore Inc., Piscataway, NJ). Biotinylated 14-residue peptides, pY213 or Y213, were coupled to the surface of a streptavidin-coated (SA) biosensor chip (BIAcore Inc., Piscataway, NJ). Binding reactions were done in HBS-EP buffer (10 mM HEPES, pH 7.4, 150 mM NaCl, 3 mM EDTA, 0.05% (v/v) surfactant P20). The surface was regenerated before each new injection using 50 mM NaOH and 1M NaCl. The Biacore instrument was programmed to perform a series of binding assays with increasing concentrations of GST Abl SH2 or GST Abl SH3-SH2 polypeptides over the same regenerated surface. Derived sensograms (plots of changes in response unit on the surface as a function of time) were analyzed using the software BIAeval 3.0. Affinity constants were estimated by curve fitting using a 1:1 binding model.
Protein and peptide binding affinities were measured by intrinsic fluorescence quenching using a Fluorolog-3 fluorimeter (Horiba Jobin Yvon Inc., Edison, NJ), with excitation at 287 nm and emission detection at 345 nm as described previously . Fluorescence intensity change was monitored as SH3-SH2 peptides (2– 4 μM in Tris buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 1 mM EDTA)) were titrated by sequential addition of appropriate peptide stock solution. Data were analyzed using OriginPro 7.5 software (OriginLab Corp., Northampton, MA) to fit a modified Stern-Volmer equation .
Postulating that phosphotyrosines would be components of the active motif, we searched for tyrosine phosphorylation sites of Abi1. We determined that candidate tyrosine residues, Y198 and Y213, are located within the proline-rich region of Abi1 previously demonstrated to bind the Abl SH3 domain . Using in vitro-translated polypeptides encoding the N-terminal half of the protein (Figure 1A) we determined that Y213 is the preferred site in the N-terminus of Abi1 of phosphorylation by Abl kinase in vitro (Figure 1B). The phosphorylation site at Y213 was confirmed by mass spectrometry of tryptic peptides following kinase reactions containing the recombinant Abi1 and active c-Abl (Table 1).
Using filter overlay assays we have previously localized binding of the c-Abl SH3 domain to residues 144-260 of Abi1 . This region of Abi1 contains a PXXP SH3 binding motif located upstream of tyrosine 213. Deletion of residues 173 through 187 of Abi1 containing the PXXP sequence, 181PPSPP185, almost completely abolished binding to the c-Abl SH3 domain (Figure 2A).
The proximity of the Abl SH3 binding site to tyrosine 213 led us to hypothesize that phosphorylated tyrosine 213 would interact with the Abl SH2 domain. This putative interaction was analyzed using a biotinylated 14-residue peptide containing phosphorylated tyrosine 213 (pY213) and a GST-Abl SH2 domain fusion protein (Figures 2B-D). First, we established that the Abl SH2 domain interacted with the phosphopeptide pY213 but not with the non-phosphorylated peptide Y213 (Figure 2B). Then, using varying concentrations of Abl SH2 we determined that pY213 binds to the SH2 domain with high affinity (dissociation constant, KD=3.32 × 10-8 M, dissociation rate, Koff= 3.76 × 10-3 M-1 s-1) (Figure 2C). Binding of pY213 with similar high affinity to the dual GST Abl SH3-SH2 domain was observed (dissociation constant, KD= 3.55 × 10-8M; dissociation rate, Koff= 2.27 × 10-3 M-1 s-1). However, control experiments showed that the Abl SH2 R171K mutant did not interact with pY213 (Figure 2D) consistent with the finding that the R171 mutation renders SH2 domains incapable of interaction with phosphopeptides .
High binding affinity measurements of the GST Abl-SH2 domain to pY213 using surface plasmon resonance (SPR) suggested the possibility of a dimerization effect of the GST fusion tag [32, 33]. Therefore we employed an intrinsic fluorescence quenching method to determine binding affinities of Abi1 peptides to Abl SH3 and SH2 domains as previously demonstrated for optimized Abl ligands . The binding affinities of Pro-Y198 (containing an SH3 domain-binding sequence), pY213 (containing an SH2 domain-binding sequence), and Pro-pY213 (containing the consolidated dual domain SH3-SH2-binding sequence) to the c-Abl SH3-SH2 dual domain were 2.3 ± 0.4 × 102 μM, 2.4 ± 0.4 × 10 μM, and 9.9 ± 0.4 ×10-1 μM, respectively. These data, obtained via intrinsic fluorescence quenching, suggest that both pY213 and Pro-Y198 are relatively weak binders to their target domains (Figure 2E). However, when two ligand binding sequences are conjoined as in Abi1, the binding of this natural consolidated ligand, i.e., the peptide Pro-pY213, is of much higher affinity.
Based on the crystal structure of c-Abl kinase, the SH2 domain-phosphopeptide interaction has the potential to regulate c-Abl kinase activity . Therefore we performed experiments to test whether the sequences of Abi1 containing tyrosine 213 affect c-Abl kinase activity. In these experiments we used active nonmyristoylated, uncapped, c-Abl , the model substrate peptide [27, 34] and forms of the 14-residue peptide containing phosphotyrosine, pY213, or tyrosine to phenylalanine replacement at position 213, F213 (for peptide sequences see Figure 1A). Addition of the pY213 phosphopeptide resulted in a 38% reduction of Vmax but no significant effect on Km of the substrate peptide, consistent with a noncompetitive mechanism of inhibition (Figures 3A-B, and Table 2). No effect on the kinase activity was observed with the peptide F213.
To determine the role of SH3 domain binding in the regulation of c-Abl kinase activity we performed kinase assays using as inhibitors various peptides derived from residues 169-217 of Abi1 (Figure 1A). As shown in Figure 3C, peptide pY213 inhibited kinase activity in a concentration-dependent manner. Peptide F213, lacking the regulatory tyrosine, did not inhibit kinase activity. Peptide Pro-pY213, which spans the entire region, exhibited a more complex effect including slight enhancement of kinase activity from 0.5 to 1 mM followed by profound inhibition of activity at higher concentrations. Further experiments demonstrated that Pro-pY213 is capable of up-regulating or down-regulating c-Abl activity in a concentration-dependent manner, and that the sequences 181PPSPP185, Y198, and pY213 of Abi1 are critical for this regulation (Figure 3D).
To determine if Abi1 regulates c-Abl tyrosine kinase activity in cells we used the LNCaP prostate cancer cell line, which, as a consequence of a heterozygous deletion of Abi1 exon 6, lacks the region containing Y213 on one allele. The mutated protein is not detected in LNCaP cells, and wild type Abi1 is expressed at a lower than expected level as compared to Abi1 expression in the primary prostate cell line, PrEC (data not shown). This result is consistent with haploinsufficiency and suggests the possibility that c-Abl kinase activity is dysregulated in LNCaP cells.
To test this possibility we first established that the endogenous c-Abl tyrosine kinase could be activated by pervanadate treatment  of LNCaP cells (Figure 4A). Activation of c-Abl kinase was consistent with phosphorylation of the regulatory tyrosine pY412  and with total tyrosine phosphorylation of c-Abl . Therefore in this experimental setting we could examine whether expression of Abi1 inhibits Abl kinase activation. We stably transfected LNCaP cells with untagged wild type isoform 2 of Abi1 (Abi1(+)) or Ha-tagged (Wt.Ha), or with Ha-tagged Abi1 mutants Y213F (F213.Ha), and 181PPSPP185 to 181AESEA185 (Pro.Ha). As shown in the immunoblots (Figure 4B left panel) and the histograms (Figure 4B, right panel) expression of intact Abi1 inhibited c-Abl kinase activity, whereas expression of Abi1 variants carrying a mutated SH2 domain-binding motif, Y213F, or a mutated SH3 domain-binding motif, 181AESEA185, did not.
Apparently, the introduction of the HA-tag at the C-terminus of Abi1 reduces the ability of the wild type Abi1 to inhibit Abl kinase activity (compare Abi1(+) to Wt.Ha cell lines, Figure 4B). This would be consistent with a possible negative effect of the HA tag on the interaction between the SH3 domain at the Abi1 C-terminus and the PRL region of Abl as demonstrated in numerous studies [21, 22].
The data from LNCaP cells suggested that Abi1 is capable of regulating c-Abl kinase activity. To determine if the observed mechanism of regulation is pertinent to the nonmyristoylated kinase as indicated by in vitro kinase data we co-expressed Abi1 and the nonmyristoylated Abl isoform 1a in Cos 7 cells. As shown in Figure 4C Abi1 reduced levels of pY412 phosphorylation of the nonmyristoylated Abl, albeit to a lower extent than treatment with STI-571. This treatment also inhibited phosphorylation of Abi1 pY213, and reduced the physical interaction with Abl (Figure 4C). A pY213-dependent association of Abi1 with c-Abl was also observed in LNCaP cells (Figure 4D).
Based on these results we postulate that phosphorylation of Y213 of Abi1 by c-Abl kinase is followed by binding of Abi1 to the Abl SH2 domain with subsequent inhibition of c-Abl kinase activity. If verified, this would be the first demonstration of inhibition of c-Abl kinase by a phosphopeptide located in trans in another protein, in this case, Abi1.
We propose that Abi1 phosphopeptides inhibit c-Abl kinase through an allosteric mechanism. This mechanism involves binding of the phosphorylated Y213 to the Abl SH2 domain. An observed decrease of the Vmax, with no change of the Km is consistent with a noncompetitive mechanism of inhibition of kinase activity by the phosphopeptide containing pY213. However, the effect of pY213 on Abl kinase activity is relatively weak (decrease of Vmax of about 38% at peptide concentration of 1 mM). This is in contrast to high binding affinity data obtained from surface plasmon resonance studies using the GST tagged Abl SH2 domain. The binding data obtained from intrinsic fluorescence quenching experiments, obtained with the untagged protein, or from overlay binding assays  (K. Machida, B.J. Mayer and L. Kotula, unpublished results), indicate much lower binding affinity of pY213, i.e. in the micromolar range. These results suggest a strong effect of the GST tag, most likely due to its dimerization effect, on SH2 binding affinities obtained using SPR as previously suggested . Our binding data for pY213-SH2, obtained with untagged protein, is consistent with the relatively weak effect on Abl kinase activity in vitro.
The in vitro kinase data demonstrated here pertains to the nonmyristoylated, partially capped, partially activated form of c-Abl kinase. However, there is no crystal structural information on the active kinase in the context of the SH3 and SH2 domains. Recent SAXS  studies indicate that the active kinase is likely to exist in the elongated form. Thus it is possible that pY213 decreases Abl kinase activity by increasing the rigidity of the kinase domain through the interactions of the SH2 domain and the N-lobe of the kinase domain, hence the noncompetitive mechanism of inhibition. The noncompetitive nature of the inhibition is demonstrated by Lineweaver-Burke plot despite a relatively high concentration of the peptide used in kinase assay. Supporting the hypothesis that pY213 regulates c-Abl tyrosine kinase activity through interactions with the c-Abl SH2 domain, is the fact that pY213 regulates physical association of Abi1. This is demonstrated here by binding assays showing interaction of pY213 with the Abl SH2 domain (Figure 2) as well as by immunoprecipitation results indicating that Abi1-pY213 interacts with the active Abl kinase in LNCaP cells (Figure 4C and 4D). pY213 phosphorylation, and consequently the strength of Abi1-Abl kinase interaction, is STI-571-dependent as indicated by co-transfection experiments. Interestingly, treatment of K562 cells with STI-571 reduced pY213 levels as compared to untreated cells  suggesting the possibility that Abi1 is a substrate of, and functions downstream of, BCR-Abl.
The lack of a crystal structure of the nonmyristoylated kinase makes it difficult to interpret the effects of Pro-pY213 peptide on kinase activity. We base our interpretation on the following facts: 1) Pro-pY213 represents the region from Abi1 that regulates c-Abl kinase activity in vitro. At higher concentrations the peptide inhibits Abl kinase activity. 2) Mutated or truncated peptides either inhibit to a lesser extent than Pro-pY213 or do not inhibit at all. These data permit identification of three critical elements affecting activity: 181PPSPP185 (representing the core Abl SH3 domain binding PXXP motif); pY213 (representing the critical SH2 domain binding phosphotyrosine); and Y198, which works together with the 181PPSPP185motif as demonstrated by Pro-Y198 peptide. 3) Multiple conformations of the autoinhibited (inactive) and active Abl kinase are possible based on recent findings on Abl  that include SAXS studies as well as Src kinase . These studies indicate significant rearrangements of the SH3-SH2 dual domain around the catalytic domain of Src-like family kinases. Therefore, we hypothesize that different effects of the pY213 and Pro-pY213 and Abl kinase activity might be a result of different peptide affinities to transient Abl conformations. For example, the presence of residual autoinhibitory interactions of the SH3 and SH2 domain assembly with the catalytic domain of the kinase  in the nonmyristoylated Abl kinase might explain the lack of effect of Pro-pY213 on the kinase activity at lower peptide concentrations. 4) The fact that the binding affinity of the dual domain-binding Pro-pY213 with Abl SH2 and SH3 domains is much greater than the sum of the binding affinities of single SH3 and SH2 ligands suggests that there is a requirement for concurrent interaction of SH2 and SH3 domains with a consolidated ligand for regulation of Abl kinase activity.
KI obtained from the kinetic analysis of Abi1 peptides (not shown) are higher and do not correlate with corresponding KD obtained from binding assays. One possible explanation may include the fact that the binding assays were performed with Abl domain(s) purified from recombinant bacteria, which would be nonphosphorylated, whereas kinase assays were performed with partially active, tyrosine phosphorylated kinase obtained from baculovirus. In this regard, a low level of pY412 and PY-99 immunoreactivity was confirmed in baculovirus-purified kinase (not shown). Importantly, these types of kinase preparations are extremely prone to activation due to autophosphorylation leading to observed differences in basal kinase activity (compare Figure 3C and 3D). Tyrosine phosphorylation of Abl  due to autophosphorylation at or near Abi1 peptide binding regions may significantly influence their binding affinities.
The Pro-pY213 region of Abi1 represents an important element that regulates Abl kinase activity in vivo as demonstrated in LNCaP cells. Abi1 Y213F or 181AESEA185 mutants did not inhibit Abl kinase activation, which indicates that concurrent binding of Abi1 to both SH3 and SH2 domain of Abl is critical for regulation. This is consistent with in vitro binding data demonstrating significant enhancement of the binding affinity of the consolidated Abi1 ligand over single-site ligands. It is possible that the 181AESEA185 mutant, despite having a higher affinity binding site for Abl, is incapable of Abl inhibition in LNCaP cells because of lower expression of total Abi1 in comparison to the clone that expresses wild type Abi1-Ha. LNCaP cells express both isoforms of Abl i.e. myristoylated and nonmyristoylated as determined by mRNA analysis (data not shown). Thus, the effects of the recombinant Abi1 on both isoforms of kinase cannot be excluded in these cells.
The hypothesis that Abi1 acts on the nonmyristoylated isoform of Abl is suggested by inhibition of the kinase in co-transfection experiments in Cos7 cells. Apparently, the nonmyristoylated kinase is constitutively active upon transfection into cells, while Abl kinase must be activated with pervanadate in LNCaP cells in order to demonstrate regulation by Abi1. As pervanadate is considered a general tyrosine phosphatase inhibitor, the action of Abi1 on Abl may be through an allosteric effect, as we postulate, or through a “shielding” effect on the catalytic domain by Abi1 SH3 domain interacting with the proline-rich region of c-Abl . Thus, steric hindrance caused by tagging of Abi1 at the C-terminal may decrease its inhibitory effect on Abl kinase as demonstrated here. We cannot exclude the possibility that Abi1 is also a competitive inhibitor in vivo in addition to its allosteric inhibition. Importantly, Abi1 must be phosphorylated at pY213, presumably by Abl, for the proposed regulation to occur. Consequently, in intact cells, Abi1 is both a candidate substrate and the candidate regulator of Abl kinase activity.
The proposed mechanism of regulation of c-Abl by Abi1 includes the possibility that Abi1 plays a role in the initial activation of c-Abl as proposed . This hypothesis would most likely apply to the myristoylated, autoinhibited kinase. Structural studies of c-Abl [15, 16] indicate that the phosphotyrosine binding site is partially occluded in the crystal structure of the myristoylated c-Abl fragment containing the SH3-SH2-catalytic domain assembly. Thus, Pro-pY213 could potentially activate the myristoylated kinase through the SH2 domain interaction as proposed [15, 16]. It is also possible that Abi1 may downregulate c-Abl that has been activated by phosphopeptides . This might occur, for example, by competing off the activating phosphopeptide by Abi1-pY213.
Conservation of the regulatory sequences suggests that other members of Abi/Hssh3bp1 family of proteins also regulate Abl. The region containing the regulatory tyrosine 213 is highly conserved between Abi1 and Abi2 from Xenopus through human, and is present in Drosophila Abi (Figure 5). The conserved sequences also include the PXXP motif, 181PPSPP185, which binds to the c-Abl SH3 domain, and tyrosine 198. All isoforms of Abi1 [22, 24, 39], or Abi2  contain the regulatory sequence indicating the conservation of c-Abl regulation in all Abi isoforms. The conserved region of Abi1 apparently plays a role in the regulation of c-Abl kinase activity in cells; here we addressed the role of the regulatory sequences in the context of Abi1 isoform 2. The fact that multiple isoforms are expressed from the Abi1/Hssh3bp1 gene  suggests the possibility of differential effects on Abl kinase activity as well as multiple downstream effects on actin cytoskeleton and Wave complex regulation [40, 41].
In summary, we identified a candidate molecular mechanism of regulation of nonmyristoylated Abl kinase. Nonmyristoylated, mutated forms of Abl, such as BCR-Abl, are implicated in chronic myelogenous leukemia and in some forms of acute lymphocytic leukemia [3, 4]. Although the use of STI-571 has brought great promise for the treatment of these diseases, some patients have become resistant to the drug following long-term treatment. STI-571-resistance mutations are found in Abl SH3 and SH2 domains  suggesting the possibility that in these cases BCR-Abl escapes possible residual regulation by other interactions including that with Abi1. The studies described here may improve our understanding of the mechanisms of Abl regulation, and may directly impact studies of BCR-Abl.
We thank B Mayer and T Koleske for plasmids and for helpful discussions; J Farmar for help in protein analysis; S Heck for help with flow cytometry; R Kascsak and J Chen for assistance with Abi1/Hssh3bp1 monoclonal antibodies; J Xu for technical assistance; and E Luna, C Redman, and M Narla for helpful discussions. Supported by grants from Department of Defense Prostate Cancer Research Program (DAMD17-01-1-0096)(LK), NINDS (NS44968)(LK), NIGM (GM47021) (DC). S Hossain was supported by the FM Kirby Foundation, Inc., Morristown, NJ. The Voyager DE was purchased through funding from the Horace W. Goldsmith Foundation, (New York, NY) and the Hyde & Watson Foundation (Chatham Township, NJ). The Biacore 3000 was purchased through funding from the Abby R. Mauze Charitable Trust (New York, NY).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.