Proc. Natl. Acad. Sci. USA Vol. 88, pp. 5927-5931, July 1991 Biochemistry Evidence for regulation of the human ABL tyrosine kinase by a cellular inhibitor (AEL human protooncogene/BCR-ABL oncogene/signal transduction/leukemia) ANN MARIE PENDERGAST*, ALEXANDER J. MULLER*, MARIE H. HAVLIKt, ROBIN CLARKt, FRANK MCCORMICK0, AND OWEN N. WITrE*t§ *Department of Microbiology and Molecular Genetics and Molecular Biology Institute, and tHoward Hughes Medical Institute, University of California, Los Angeles, CA 90024; and tDepartment of Molecular Biology, Cetus Corporation, Emeryville, CA 94608 Communicated by H. Ronald Kaback, April 8, 1991

ABSTRACT Phosphotyrosine cannot be detected on nor- Intramolecular negative control by substrate-like sequences mal human ABL -tyrosine kinases, but activated onco- has been demonstrated for a number of protein kinases such genic forms of the human ABL protein are phosphorylated on as protein kinase C (19) and myosin light chain kinase (20). It tyrosine in vivo. Activation of ABL can occur by substitution of has been suggested (21) that sequences of the murine c- the ABL first exon with breakpoint duster region (BCR) protein encoded by the first exon may exert inhibitory effects sequences or by ofthe noncatalytic SH3 (src homology on its kinase activity by directly interfering with the active region 3) domain. An alternative mode for the activation ofthe site. Recently it was shown that removal of the SH3 domain ABL kinases is hyperexpression at >500-fold over endogenous but not first-exon-encoded sequences activates the kinase levels. This is not a consequence of transphosphorylation ofthe and transforming potential ofABL (14, 15). Alternatively, the hyperexpressed ABL molecules. ABL translated in ABL kinase may be negatively regulated by noncovalent viro lack phosphotyrosine, but tyrosine kinase activity is associations with intracellular factors. We present data that uncovered after immunoprecipitation and removal of lysate supports the hypothesis that the ABL kinase is partly regu- components. The rates of dephosphorylation of ABL and lated by an intermolecular noncovalent association with a BCR-ABL by phosphotyrosine-specific phos- cellular factor. phatases are approximately the same. These combined results indicate that inhibition ofABL activity is reversible and suggest MATERIALS AND METHODS that a cellular component interacts noncovalently with ABL to inhibit its autophosphorylation. Preparation of Plasmid Vectors and Viruses. Recombinant baculovirus vectors were constructed using the transfer The function of the receptor-type tyrosine kinases is to vectors pAcC6 or pAcC12 (22). The pAcBA2 vector con- transduce the signal provided by the binding of the corre- taining the human P210-encoding BCR-ABL oncogene was sponding growth factor. The nonreceptor tyrosine kinases, constructed as described (23). Baculovirus vectors contain- for which the c-src, c-fps, and c-abl tyrosine kinases are ing the human P185-encoding BCR-ABL (pAcBA3), ABL prototypes, are believed to function as signal-transducing type la (pAcAl), and ABL type lb (pAcA2) were prepared by molecules in cellular proliferation and differentiation pro- inserting the corresponding cDNAs into the EcoRI site ofthe cesses (1-4). There is a striking conservation of the SH2 and pAcC12 transfer vector. Construction ofthe cDNAs for P185 SH3 (src homology regions 2 and 3) noncatalytic domains of BCR-ABL protein (24) and for ABL types la and lb, BCR- these tyrosine kinases in other molecules involved in signal ABL deletion mutants, and ABL type lb(A15-138) deletion transduction pathways such as phospholipase C (PLC)-y and mutant (25) is described in detail in the references cited. The the ras GTPase-activating protein (GAP) (3). kinase-negative ABL type lb(290R) and P210 BCR- The c-abl protooncogene was first identified as the normal ABL(1172R) mutants were prepared by oligonucleotide- cellular homolog of the v-abl oncogene of Abelson murine directed mutagenesis in phage M13 (Amersham): the lysine at leukemia virus (5, 6). c-abl sequences participate in the position 290 of ABL type lb and position 1172 of P210 formation of three other oncogenes: the v-abl of HZ2- BCR-ABL was changed to arginine (R in single-letter code). feline sarcoma virus (7), the P210-encoding BCR-ABL fusion The carboxyl-terminal deletion mutant ABL lb(A731-1149) gene in human chronic myelogenous leukemia (CML) (8-10), was made by digestion of ABL type lb with Eco47III and and the P185-encoding BCR-ABL fusion gene in human acute addition of Xba I linkers. cDNAs for wild-type and mutant lymphoblastic leukemia (ALL) (11-13). The normal ABL ABL forms were also subcloned into the EcoRI sites of the proteins are tightly regulated in vivo, while the oncogenic COS (African green monkey kidney) cell expression plasmid forms escape intracellular regulation. No tyrosine phosphor- pcDL-SRa296 (26), the retroviral expression vector ylation ofnormal ABL proteins has been detected in vivo (10, pSRaMSVtkneo (25), and the pGEM-4 expression vector 14-16). Expression of the oncogenic ABL proteins results in (Promega). Recombinant baculoviruses (23) and retroviral tyrosine hyperphosphorylation of numerous cellular sub- stocks (25) were prepared as described. strates, including the ABL oncoproteins themselves (17). A Biochemical Techniques. Procedures for immunoblotting strong correlation exists between oncogenic transformation (25) metabolic labeling with [32P]orthophosphate (23), immu- and increase of ABL kinase activity (reviewed in ref. 18). nocomplex tyrosine kinase assays (29), phospho amino acid Regulation of ABL kinase activity could be achieved analysis (23), in vitro transcription and translation (30), and through covalent modifications and/or by noncovalent inter- dephosphorylation with phosphotysosine [Tyr(P)] specific actions, which may be intra- or intermolecular in nature. Abbreviations: Tyr(P), phosphotyrosine; HRP, horseradish peroxi- dase; BCR, breakpoint cluster region; GAP, GTPase-activating The publication costs of this article were defrayed in part by page charge protein; AcGAP, Autographa californica nuclear polyhydrosis vi- payment. This article must therefore be hereby marked "advertisement" rus; SH2 and SH3, src homology regions 2 and 3. in accordance with 18 U.S.C. §1734 solely to indicate this fact. §To whom reprint requests should be addressed.

5927 Downloaded by guest on September 25, 2021 5928 Biochemistry: Pendergast et al. Proc. Natl. Acad. Sci. USA 88 (1991)

phosphatases and general phosphatases (31, 32) are described Cell type Sf9 COS NIH 3T3 in detail in the references cited. Dilution of 1:1000 1:50 Neat lysates 1 2 r34 56 RESULTS

Hyperexpression ofHumanABL in Insect Cells Results in the A f* ^-200 Production of Proteins That Are Phosphorylated on Tyrosine in Vivo. The baculovirus expression system was used to * A - -116 hyperproduce human ABL proteins in insect cells. Similar levels of ABL and BCR-ABL proteins were obtained, which were full length and soluble (Fig. lA). In contrast to endog- enous c-abl proteins, human ABL proteins overproduced by baculovirus contained Tyr(P) in vivo at levels similar to those Exposure [ 6mmin of the BCR-ABL forms (Fig. 1B), as shown by immunoblot- time ting of immunoprecipitates of these proteins with Tyr(P)- specific antibodies (Fig. 1B) and as confirmed by phospho amino acid analysis of the 32P-labeled proteins after labeling B S " "~-200 of cells with [32P]orthophosphate (data not shown). Activation of ABL Kinases Is Dependent on the Level of Overexpression. The accumulation of Tyr(P) on ABL follow- ing hyperexpression suggests that the lack of Tyr(P) on endogenous c-abl may be due to inhibition of autophosphor- -116 ylation by factors that are present in limiting amounts in the cell and that could be titrated out by high ABL concentra- tions. To test this idea, two additional expression systems Exposure I-6 min -11 5 minI were used for producing ABL at different levels: (i) trans- time fection into monkey COS cells and (ii) retroviral infection into NIH 3T3 murine cells (Fig. 2). Whereas human BCR- FIG. 2. Overexpression of ABL cDNA in mammalian and insect ABL proteins were phosphorylated on tyrosine regardless of cells leads to activation in a dosage-dependent manner. Three the level of expression or cell type, normal human ABL expression systems were used to produce ABL type lb (lanes 1, 3, and 5) and P185 BCR-ABL (lanes 2, 4, and 6) at different levels: baculovirus infection of Sf9 insect cells (lanes 1 and 2) (23); trans- fection of monkey COS cells (lanes 3 and 4) (26); and retroviral infection of murine NIH 3T3 cells (lanes 5 and 6) (25). Cells were I infected or transfected and lysed after 2 days. Equal amounts of A protein from each lysate were determined by the method of Lowry. Insect and COS cell lysates were then diluted 1:1000 and 1:50, " 200 respectively, to obtain levels ofABL and BCR-ABL proteins similar to those observed in NIH 3T3 cells after retroviral infection. Proteins were separated by SDS/PAGE and electrophoretically transferred to nitrocellulose filters. The filters were incubated for 3 hr with either mouse monoclonal anti-ABL 21-63 (A) or mouse monoclonal anti- - 97 Tyr(P) (B) antibodies, washed, and then incubated for 1 hr with goat anti-mouse HRP. The filters were washed and developed with B enhanced chemiluminescence detection and to so reagents exposed - 200 autoradiography for 6 or 15 min as indicated. Sizes are shown in kDa. a1o proteins were tyrosine-phosphorylated in proportion to their 4j level of hyperexpression. Normal human ABL proteins in- troduced by acute retroviral infection into mouse NIH 3T3 - 97 cells were expressed at approximately 5- to 10-fold higher 1 2 3 45 levels than endogenous c-abl but lacked detectable Tyr(P) (Fig. 2, lane 5). In contrast, human ABL proteins overex- FIG. 1. Normal and altered ABL proteins are phosphorylated on pressed in COS monkey cells accumulated to -200-fold tyrosine in baculovirus-infected Sf9 insect cells. Sf9 (fall armyworm) higher levels than endogenous c-abl and contained high levels insect cells (3 x 106) were infected with recombinant baculoviruses. of Tyr(P) (Fig. 2, lane 3), similar to the BCR-ABL proteins Two days after infection, the cells were lysed in the presence of in the same cell lane and to the ABL phosphatase inhibitors (23), and the ABL proteins were immuno- expressed type (Fig. 2, 4) precipitated with anti-pEX-5. Samples were analyzed by SDS/ proteins hyperproduced in Sf9 insect cells (Fig. 2, lanes 1 and PAGE and the proteins were transferred to nitrocellulose filters. The 2). Other cellular proteins larger and smaller in size than the filters were incubated with mouse monoclonal anti-ABL antibody ABL proteins also become tyrosine-phosphorylated after 21-63 (27) (A) or with rabbit polyclonal anti-Tyr(P) antibodies (28) hyperexpression of ABL and BCR-ABL in Sf9 and COS (B). After incubation for 3 hr at room temperature, the filters were cells. The phosphorylation of cellular proteins was clearly washed and then incubated with either goat anti-mouse IgG horse- observed by immunoblotting (Western) analysis of undiluted radish peroxidase (HRP) conjugate (A) or 125I-labeled protein A (37 lysates with anti-Tyr(P) antibodies (data not shown). The kBq/ml) (B) for 1 hr at room temperature. Those filters incubated level of human ABL overexpression from baculovirus vec- with the goat anti-mouse HRP (A) were developed with the enhanced tors in Sf9 insect cells was about 5000-fold higher than that of chemiluminescence detection reagents (Amersham) and then ex- posed to x-ray film for 2 min. Filters incubated with 125I-labeled the endogenous c-abl (Fig. 2, lane 1). The amount of ABL protein A were exposed to x-ray film for 3 hr at -70°C. Lanes: 1-4, protein produced in the transfected COS cells is probably Sf9 insect cells infected with baculovirus vectors encoding ABL type >200-fold over the endogenous c-abl levels. Only 20-30% of la, ABL type lb, P185 BCR-ABL, and P210 BCR-ABL, respec- the COS cells express the ABL proteins after transfection as tively; 5, uninfected cells. Sizes are shown in kDa. determined by immunostaining (data not shown). Thus, the Downloaded by guest on September 25, 2021 Biochemistry: Pendergast et al. Proc. Natl. Acad. Sci. USA 88 (1991) 5929 ABL proteins are hyperexpressed at approximately 500-fold greater amounts over endogenous c-abl in each of the ABL- A c5, (. .GI expressing COS cells. PS We treated cells with sodium orthovanadate, a Tyr(P) phosphatase inhibitor. No Tyr(P) on ABL was recovered q 1 under these conditions. In contrast, P185 BCR-ABL, ex- aA BL II 1- aABL pressed at levels comparable to the endogenous c-abl, dis- - 200 played a 3-fold increase of Tyr(P) in vanadate-treated cells * -200 compared with untreated cells (data not shown). an fBmom aABLI I i We also measured the Tyr(P) in ABL made at lower levels - 116 -116 in Sf9 cells. This measurement was made to determine -200 whether recovery of Tyr(P) on ABL could be due to the caPTYR cxPrYR 2 3 4 5 6 absence of a ABL inhibitory activity in insect cells or, alternatively, could be a result of changes in the cell envi- - 200 _ -200 ronment triggered in response to baculovirus infection. A reduction in the level ofABL protein expressed by more than -116 a factor of 20 was obtained by simultaneous infection with a competing recombinant baculovirus such as GAP-carrying i2 3 4 5 6 7 8 1 2 34 56 Autographa californica nuclear polyhydrosis virus (AcGAP) cxexon lb +- +- + - + - (33) (Fig. 3A) or with wild-type baculovirus (data not shown). aABL This reduction in ABL protein produced yields ABL mole- C Terminus cules with low-to-undetectable Tyr(P) (Fig. 3B, lane 3). In KJNASE ACTIV'.:TY contrast, a similar reduction in the level of P185 BCR-ABL ^, 2 _ expression yields tyrosine-phosphorylated P185 BCR-ABL AEL lb + rL . i-. proteins (Fig. 3B, lane 5). These results favor the hypothesis ABL Ib(290R: ABL intracellular factors that are present of regulation of by P185, %3-426i at limiting concentrations. + ABL Proteins Are Not Transphosphorylated in Vivo after 8 Hyperexpression. An alternative explanation for the presence of P2101117RR Tyr(P) on ABL after hyperexpression is that ABL is transphos- I 3§ phorylated by other ABL molecules that are in close proximity because of concentration. Kinase-active FIG. 4. BCR-ABL but not ABL proteins are transphosphorylated their high intracellular in vivo after hyperexpression in COS monkey cells. COS cells were and -inactive forms ofABL were coexpressed in monkey COS transfected with 15 ,ug ofthe indicated DNAs. Cells were lysed 3 days cells (Fig. 4). A kinase-defective mutant of ABL lb was after transfection. Proteins were analyzed directly by SDS/PAGE constructed by substitution of arginine for Lys-290, which and immunoblotting (B). Alternatively, the ABL proteins were prevents ATP binding. No Tyr(P) was detected on the tyrosine immunoprecipitated from the lysates prior to gel electrophoresis (A). kinase-defective ABL lb(290R) mutant protein when coex- Immunoprecipitation was carried out with ABL exon lb-specific pressed with a deletion mutant, P185(A3-426), encoded by a antibody AM9-03 (aexon lb) (A, lanes 1, 3, 5, and 7) or with the construct lacking the first ABL exon. Sequence-specific anti- carboxyl terminus-specific anti-pEX-5 antibody (aABL C terminus) bodies were used to distinguish ABL lb(290R) and P185(A3- (A, lanes 2, 4, 6, and 8). Proteins were transferred electrophoretically molecularweights. The to nitrocellulose filters. The filters were then incubated with either 426), since the two proteins have similar mouse monoclonal anti-ABL antibody 21-63 (A and B Top) or with anti-ABL lb first-exon-specific antibody does not recognize mouse monoclonal anti-Tyr(P) (aPTYR) antibodies (A and B Mid- P185(A3-426) (Fig. 4A Upper, lane 5) but immunoprecipitates dle). After incubation for 3 hr at room temperature, the filters were then processed as described in Fig. 2. Sizes are shown in kDa. ABL P185 (Bottom) Structures of the corresponding protein products for the A transfected DNAs. ,- 200 the ABL lb(290R) kinase-defective protein (Fig. 4A Upper, lane 3). In cells that coexpress the two proteins, no Tyr(P) is detected on ABL lb(290R) after immunoprecipitation with the -97 anti-ABL lb first-exon-specific antibody (Fig. 4A Lower, lane 7). Similar results are obtained when the kinase-defective ABL B lb(290R) protein is coexpressed with a ABL lb carboxyl- -2cc terminal deletion mutant protein, ABL lb(A731-1149) (data not shown) or with the transforming P185 BCR-ABL (Fig. 4B Lower, lane 5). Interestingly, coexpression of a kinase-defective P210 -97 BCR-ABL mutant protein, P210(1172R), with the kinase- 1 2 3 45 active P185 BCR-ABL resulted in transphosphorylation of h ,AP - + P210 (Fig. 4B Lower, lane 6). BCR sequences may permit association ofBCR-ABL molecules such that transphosphor- FIG. 3. Reduction in the level ofABL expression in baculovirus- ylation may occur. These results are in agreement with the infected Sf9 cells results in the recovery of ABL proteins lacking previous finding that highly enriched P210 BCR-ABL mi- Tyr(P). Sf9 cells (3 x 106) were singly or doubly infected with grates with a molecular mass of800 kDa on a Sephacryl S-400 recombinant baculoviruses expressing ABL type la (lanes 2 and 3), gel filtration column (23), suggesting the existence of a P210 P185 BCR-ABL (lanes 4 and 5), and GAP (33) (lanes 1, 3, and 5). Two BCR-ABL oligomeric structure. days after infection, the cells were lysed in the presence of phos- phatase inhibitors (23), and the ABL proteins were immunoprecip- Negative Regulator of ABL Kinases Is Present in Reticulo- itated with anti-pEX-5. Samples were analyzed by SDS/PAGE and cyte Lysates. Approximately equal levels of wild-type and Western blotting with anti-ABL (A) or anti-Tyr(P) (B) antibodies as mutant forms of ABL and BCR-ABL proteins could be described in Fig. 2. Sizes are shown in kDa. rapidly synthesized in the reticulocyte lysate. The intrinsic Downloaded by guest on September 25, 2021 5930 Biochemistry: Pendergast et al. Proc. NatI. Acad. Sci. USA 88 (1991) tyrosine kinase activity of the various proteins could then be of Tyr(P) were found on P185 and P210 BCR-ABL proteins compared by measuring the Tyr(P) levels on each ofthe ABL (Fig. SB, lanes 3 and 4). Two BCR-ABL deletion mutants that forms (Fig. 5). No Tyr(P) was detected on ABL type la, ABL retained various portions of BCR sequence contained lower type lb, or a BCR-ABL deletion mutant that lacks -all BCR but detectable levels of Tyr(P) (Fig. SB, lanes 5 and 6). sequences (Fig. SB, lanes 1, 2, and 7). In contrast, high levels Deletion of the SH3'domain in ABL type lb (Fig. SB, lane 8) and the BCR-minus mutant (Fig. 5B, lane 9) yielded proteins A B that contained high Tyr(P) levels. The lack of Tyr(P) in ABL 2 3 4 5 6 7 8 9 10 2 3 4 5 6 7 8 9 10 la cDNA, ABL lb cDNA, and the BCR-minus mutant in reticulocyte lysates was not due to translation of enzymati- cally inactive forms of these proteins because all three X enzymes could be shown to be active tyrosine kinases after immunQprecipitation and removal of lysate components fol- is lowed by an in vitro autophosphorylation assay (Fig. SD, lanes' 1, 2, and 7). 3 ABL and BCR-ABL Proteins Are Dephosphorylated by Tyr(P)-Specitic Phosphatases at Similar Rates. The lack of Tyr(P) in the normal'ABL proteins could be the result ofrapid dephosphorylation -by Tyr(P)-specific phosphatases. We C D have compared the rates of dephosphorylation for tyrosi-ne- phosphorylated forms of ABL and BCR-ABL (Fig. 6). ABL 2 3 4 5 6 7 8 910 1 2 3 4 5 6 7 8 9 10 and BCR-ABL proteins were produced in baculovirus- in the _ - 200 infected insect cells, isolated by immunoprecipitation S presence' of Tyr(P)-specific phosphatase inhibitors, and in- we 1wa cubated with various lysates or purified Tyr(P)-specific phos- phatases. Tyr(P) on ABL and P185 BCR-ABL was removed -97 at similar'rates in vitro by Tyr(P)-specific phosphatases present in rabbit reticulocyte (Fig. 6 B and D, lanes 1-5) and fibroblast cell lysates (data not shown). Similar results were obtained with a purified T-cell Tyr(P)-specific phosphatase A 12345 678910 SH - 200 32 1 1 ABL la ABL--0010"0 -116 2 ABL lb >l 1 I'T"].F1 B 3. Pi855 ~ ~ - 200

PA * 0 4. P210S ______ABL -"""be M, - 116 5. P185(A3-242) _- C BCR a -200 6. P185(3176-426) -ABL-WII_

7. PI 85(A3-426) II -116 8 P185(A3-519) D -200 BCR-ABL- 600mWe 9 ABL lb (A'5-138- NoWNBum.

10 NORNA -116 SOURCE OF FIG. 5. Negative regulator ofABL kinases is present in reticulocyte LYSATE/Pase RETICULOCYTE lysates. Wild-type and altered forms ofthe ABL gene were cloned in the LYSATE PTPase pGEM4 vector and transcribed from the SP6 promoter with SP6 _I +++ + polymerase. The mRNAs were translated for 1 hrin a reticulocyte lysate INCUBATION in the presence of [35S]-methionine. After translation, halfofeach lysate TIME (MIN ) 15 0 2 5 15 10 0 2 5 IC was mixed with an equal volume of double-strength sample buffer and boiled for 5 min. The proteins were separated on SDS/8% polyacrylam- FIG. 6. ABL and BCR-ABL have similar rates of dephosphory- ide gels and transferred electrophoretically to nitrocellulose filters. The lation for Tyr(P). ABL type lb and P185 BCR-ABL proteins were filters were incubated with either mouse monoclonal anti-ABL (A) or isolated from 3-day-infected baculovirus-infected Sf9 cells. Infected mouse monoclonal anti-Tyr(P) (B) antibodies and processed as indicated cells were lysed in the presence of the Tyr(P)-specific phosphatase in Fig. 1. The remainder of each lysate was immunoprecipitated with (PTPase) inhibitors. The ABL proteins were immunoprecipitated anti-ABL antibody (anti-pEX-5) and halfofeach immunoprecipitate was with anti-pEX-5 and incubated in the absence or presence of rabbit analyzed on SDS/8% polyacrylamide gels and subjected to fluorography reticulocyte lysate (lanes 1-5) and a partially purified T-cell PTPase to detect the 3S-labeled proteins (C). The second halfofeach immuno- (lanes 6-10) in the presence of 50 mM EDTA. The immunoprecip- precipitate was processed for in vitro autophosphorylation in the pres- itates were washed, and the proteins were eluted with double- ence of [-32P]ATP and MnCl2 after washing lysate components as strength sample buffer. The proteins were analyzed on SDS/8% described (29). The reactions were terminated after 30 min at 30WC and polyacrylamide gels and transferred to nitrocellulose filters. The analyzed on SDS/8% polyacrylamide gels. The 32P-labeled proteins were filters were incubated with mouse monoclonal anti-ABL (A and C) detected by autoradiography with four sheets of aluminum foil to block or mouse monoclonal anti-Tyr(P) (B and D) antibodies for 3 hr at the 35S-label signal (D). Sizes are shown in kDa. Theidentityoflanes 1-10 room temperature. The filters were washed and incubated with is described in the schematic diagram at the bottom ofthe figure. Various '2-I-labeled goat anti-mouse IgG [30 1tCi (1 gCi = 37 kBq)/blot; 0.125 amounts of BCR sequences are represented by the dotted boxes. The jug/ml]. After 1 hr ofincubation at room temperature, the filters were SH1, SH2, and SH3 domains are indicated. washed and exposed to x-ray film. Sizes are shown in kDa. Downloaded by guest on September 25, 2021 Biochemistry: Pendergast et al. Proc. NatL. Acad. Sci. USA 88 (1991) 5931 (34) (Fig. 6 B and D, lanes 6-10) and potato acid phosphatase assistance in preparation of the manuscript, and to Mike Cohan for (data not shown). The levels of ABL and BCR-ABL proteins photography. We thank Dr. Hidesaburo Hanafusa for the generous remained essentially constant as determined by immunoblot- gift of rabbit polyclonal anti-Tyr(P) antibodies and Drs. Ed Fischer ting (Fig. 6A and C). These results suggest that lack ofTyr(P) and Nick Tonks for the purified T-cell Tyr(P)-specific phosphatase. This work was supported in part by grants from the National on inhibition of the rate of tyrosine ABL results from Institute to O.N.W. and U.S. Public Health Servicee'National Re- phosphorylation and not from differential rates of dephos- search Service Award4GM07185 from the National Institutes of phorylation for the ABL kinases. Health to A.J.M. A.M.P. is a Leukemia Society special Fellow. is an Investigator of the Howard Hughes Medical Institute. DISCUSSION O.N.W. Protein tyrosine kinases are regulated by covalent modifica- 1. Hunter, T. & CoopeT, J. A. (1985) Annu. Rev. Biochem. 54, 897-930. tions and/or noncovalent associations with cellular factors. 2. Bishop, J. M. (1987) Science 235, 305-311. Covalentmodifications suchasphosphorylation/dephosphor- 3. Paws~n, T. (1988) Oncogene 3, 491-495. 4. floffmann, F. M. (1989) Curr. Topics Microbiol. Immunol. 147, 1-29. ylation of tyrosine, serine, and threonine residues has been 5. Witte, 0. N., Rosenberg, N. E. & Baltimore, D. (1979) Nature (London) shown to modulate the activity of both growth factor receptor 238, 826-831. tyrosine kinases (35) and nonreceptor tyrosine kinases such 6. Goff, S. P., Gilboa, E., Witte, 0. N. & Baltimore, D. (1980) Cell 22, as pp60csrc (36). Noncovalent associations of growth-factor 777-785. with their receptors leads to activation of the corre- 7. Besmer, P., Hardy, W. D., Jr., Zuckerman, E. E., Bergold, P., Leder- ligands man, L. & Snyder, H. W., Jr. (1983) Nature (London) 303, 825-828. sponding receptor-associated tyrosine kinase. Ligand- 8. Shtivelmant E., Lifshitz, B., Gale, R. P., Roe, B. A. & Canaani, E. induced activation of the kinase domain appears to be me- (1986) Cell 47, 277-284: diated by receptor oligomerization (reviewed in ref. 35). 9. Mes-Masson, A. M., McLaughlin, J., Daley, G. Q., Paskind, M. & p561ck, a member of the src family of cytoplasmic tyrosine Witte, 0. N. (1986} Proc. Natl. Acad. Sci. USA 83, 9768-9772. (1984) Cell 37, 1035- is with the cell-surface 10. Konopka, J. B., Watanabe, S. M. & Witte, 0. N. kinases, associated noncovalently 1042. glycoproteins CD4 and CD8, which are expressed on func- 11. Fainstein, E., Marcelle, C., Rosner, A., Canaani, E., Gale, R. P., tionally distinct subpopulations' of'T cells. Cros's-linking of Dreazen, O., Smith, S. D. & Croce, C. M. (1987) Nature (London) 330, CD4' and- CD8 with antibody induces a rapid increase in the 386-388. tyrosine phosphorylation of p561ck (37, 38).' 12. Hermans, A., Heisterkamp, N., von Lindem, M., van Baal, S., Meijer, Plas, D., Wiedemann, L. M., Groffen, J., Bootsma, D. & in a or in D., van der ABL synthesized reticulocyte lysate produced Grosveld, G. (1987) Cell 51, 33-40. mammalian cells at low levels lacks detectable Tyr(P), but 13. Clark, S. S., McLaughlin, J., Crist, W. N., Champlin, R. & Witte, 0. N. after immunoprecipitation and removal of lysate compo- (1987) Science 235, 85-88. nents, ABL exhibits tyrosine kinase activity. These results 14. Jackson, P. & Baltimore, D. (1989) EMBO 1. 8, 449-456. imply that the major inhibition of ABL'activity involves 15. Franz, W. M., Berger, P. & Wang, J. Y. J. (1989) EMBO J. 8, 137-147. 16. Ponticelli, A. S., Whitlock, C. A., Rosenberg, N. & Witte, 0. N. (1982) noncovalent rather than covalent alterations. Furthermore, Cell 29, 953-960. because the rate of Tyr(P) dephosphorylation is the same for 17. Lugo, T. G., Pendergast, A. M., Muller, A. J. & Witte, 0. N. (1990) ABL and BCR-ABL in vitro and treatment of cells with Science 247, 1079-1082. Tyr(P)-specific phosphatase inhibitors does not result in the 18. Pendergast, A. M. & Witte, 0. N. (1987) in Balliere's Clinical Haema- recovery of tyrosine-phosphorylated ABL, we suggest that tology, ed. Goldman, J. M. (Saunders, London), Vol. 1, No. 4, pp. 1001-1020. the putative ABL regulatory factor inhibits the ABL kinases 19. House, C. & Kemp, B. (1987) Science 238, 1726-1728. by affecting the rate of tyrosine phosphorylation and not the 20. Pearson, R. B., Wettenhall, R. E. H., Means, A. R., Hartshorne, D. J. rate of dephosphorylation. & Kemp, B. E. (1988) Science 241, 970-973. Like ABL, the catalytic activities of other nonreceptor 21. Wang, J. Y. J. (1988) Oncogene Res. 3, 293-298. 22. Quilliam, L. A., Der, C. J., Clark, R., O'Rourke, E. C., Zhang, K., tyrosine kinases are under strong negative regulatory control McCormick, F. & Bokoch, G. M. (1990) Mol. Cell. Biol. 10, 2901-2908. in vivo. Data from several laboratories support the hypothesis 23. Pendergast, A. M., Clark, R., Kawasaki, E. S., McCormick, F. P. & that phosphorylation of Tyr-527 suppresses the kinase and Witte, 0. N. (1989) Oncogene 4, 759-766. transforming potential of pp6csrc in vivo (reviewed in ref. 24. McLaughlin, J., Chianese, E. & Witte, 0. N. (1989) Mol. Cell. Biol. 9, 1866-1874. 36). Other src-related kinases such as pp56lck are also neg- 25. Muller, A. J., Young, J. C., Pendergast, A. M., Pondel, M., Landau, atively regulated by phosphorylation of a carboxyl-terminal N. R., Littman, D. R. & Witte, 0. N. (1991) Mol. Cell. Biol. 11, tyrosine (36). ABL differs significantly from the src family of 1785-1792. protein tyrosine kinases in that it does not contain a tyrosine 26. Takebe, Y., Seiki, M., Fujisawa, J.-I., Hoy, P., Yokata, K., Arai, K.-I., Yoshida, M. & Arai, N. (1988) Mol. Cell. Biol. 8, 466-472. equivalent to the c-src Tyr-527 (8). The principal regulatory 27. Schiff-Maker, L., Bums, M. C., Konopka, J. B., Clark, S., Witte, 0. N. region of ABL is found in the amino-terminal region of the & Rosenberg, N. (1986) J. Virol. 57, 1182-1186. protein and corresponds to the SH3 domain (14, 15). Dele- 28. Hamaguchi, M., Grandori, C. & Hanafusa, H. (1988) Mol. Cell. Biol. 8, tions and of the SH3 domain in c-src can also 3035-3042. 29. Konopka, J. B. & Witte, 0. N. (1985) Mol. Cell. Biol. S. 3116-3121. activate the c-src kinase (39, 40). 30. Muller, A. J. & Witte, 0. N. (1989) Mol. Cell. Biol. 9, 5234-5238. Activation of ABL through deletion of the SH3 domain 31. Tonks, N. K., Diltz, C. D. & Fischer, E. H. (1988) J. Biol. Chem. 263, suggests that this region may be involved in mediating 6731-6737. negative regulation of the tyrosine kinase. This region does 32. Cooper, J. A. & King, C. S. (1986) Mol. Cell. Biol. 6, 4467-4477. G. A., since away of 33. Trahey, M., Wong, G., Halenbeck, R., Rubinfeld, B., Martin, not appear to be directly inhibitory, washing Ladner, M., Long, C. M., Crosier, W. J., Watt, K., Koths, K. & cellular components after immunoprecipitation results in McCormick, F. (1988) Science 242, 1697-1700. activation of ABL. Upstream fusion of BCR first exon 34. Cool, D. E., Tonks, N. K., Charbonneau, H., Walsh, K. A., Fischer, sequences to ABL specifically mimics the effect on the E. H. & Krebs, E. G. (1989) Proc. Natl. Acad. Sci. USA 86, 5257-5261. domain (25) and may interfere with 35. Ullrich, A. & Schlessinger, J. (1990) Cell 61, 203-212. kinase ofdeleting the SH3 36. Cooper, J. A. (1990) in Peptides and Protein Phosphorylation, eds. negative regulation mediated through this domain. The find- Kemp, B. & Alewood, P. F. (CRC, Boca Raton, FL), pp. 85-113. ing that BCR/ABL but not ABL molecules can transphos- 37. Veillette, A., Bookman, M. A., Horak, E. M., Samelson, L. E. & Bolen, phorylate in vivo suggests that oligomerization through BCR J. B. (1989) Nature (London) 338, Z57-259. play a role in activation. 38. Luo, K. & Sefton, B. M. (1990) Mol. Cell. Biol. 10, 5305-5313. segments might 39. Kato, J.-Y., Takeya, T., Grandori, C., Iba, H., Levy, J. B. & Hanafusa, J. (1986) Mol. Cell. Biol. 6, 4155-4160. We are grateful to Arnie Berk and Bill Wickner for critical review 40. Potts, W. M., Reynolds, A. B., Lansing, T. J. & Parsons, J. T. (1988) of the manuscript, to Carol Crookshank and Naomi Nobles for Oncogene Res. 3, 343-355. Downloaded by guest on September 25, 2021