The Carboxyl Terminus of v-Abl Protein Can Augment SH2 Domain Function

Abstract

Abelson murine leukemia virus (Ab-MLV) transforms NIH 3T3 and pre-B cells via expression of the v-Abl tyrosine kinase. Although the enzymatic activity of this molecule is absolutely required for transformation, other regions of the protein are also important for this response. Among these are the SH2 domain, involved in phosphotyrosine-dependent protein-protein interactions, and the long carboxyl terminus, which plays an important role in transformation of hematopoietic cells. Important signals are sent from each of these regions, and transformation is most likely orchestrated by the concerted action of these different parts of the protein. To explore this idea, we compared the ability of the v-Src SH2 domain to substitute for that of v-Abl in the full-length P120 v-Abl protein and in P70 v-Abl, a protein that lacks the carboxyl terminus characteristic of Abl family members. Ab-MLV strains expressing P70/S2 failed to transform NIH 3T3 cells and demonstrated a greatly reduced capacity to mediate signaling events associated with the Ras-dependent mitogen-activated protein (MAP) kinase pathway. In contrast, Ab-MLV strains expressing P120/S2 were indistinguishable from P120 with respect to these features. Analyses of additional mutants demonstrated that the last 162 amino acids of the carboxyl terminus were sufficient to restore transformation. These data demonstrate that an SH2 domain with v-Abl substrate specificity is required for NIH 3T3 transformation in the absence of the carboxyl terminus and suggest that cooperativity between the extreme carboxyl terminus and the SH2 domain facilitates the transmission of transforming signals via the MAP kinase pathway.

Abelson murine leukemia virus (Ab-MLV) is a highly oncogenic retrovirus that transforms NIH 3T3 cells and pre-B cells in vitro and induces pre-B-cell lymphomas in vivo (60). The virus arose via recombination between Moloney murine leukemia virus (Mo-MLV) and the c-abl proto-oncogene gene and encodes a single product, the v-Abl protein tyrosine kinase. This protein contains amino-terminal determinants derived from the Mo-MLV gag gene fused to abl-derived sequences which specify the SH2 and catalytic domains and the long carboxyl-terminal region characteristic of Abl protein family members (18, 42). Although the protein tyrosine kinase activity of v-Abl is absolutely required for all transforming functions of the virus, other portions of the protein play important roles. The SH2 domain facilitates phosphotyrosine-mediated protein-protein interactions (35), and the carboxyl terminus is important for transformation of lymphoid cells (23, 40, 50, 58).

Mutations affecting different regions of the v-Abl protein have shed light on their functions; however, the way these regions may work together to orchestrate transformation is less clear. For the modular SH2 domain, similar domains from other signaling molecules can partially substitute for its function in transformation. For example, the SH2 domains of Crk and several other proteins can substitute for endogenous SH2 sequences in activated c-abl alleles (33, 38). However, the pattern of tyrosine-phosphorylated cellular proteins is not identical (33), and transformation usually occurs at a reduced frequency. In contrast, chimeras in which the entire amino terminus of v-Src has been fused to the v-Abl catalytic domain are fully transforming for NIH 3T3 cells (22), suggesting that the presence of an SH3 domain alters the response, perhaps by providing docking sites for cellular signaling molecules (34, 41).

The carboxyl terminus of Abl proteins is also involved in protein-protein interactions (2, 8, 29, 36, 48, 65, 68), some of which appear to involve cellular proteins which also interact with other portions of the Abl protein (6, 64). In addition, the carboxyl terminus also appears to signal to the Ras pathway (15, 45, 54), an event that is critical for Ab-MLV transformation (62). Other studies have identified a Jak-interactive region and shown that Jak activation is important for v-Abl-induced factor-independent growth of hematopoietic cell lines (7, 8). In addition to these functions, the carboxyl terminus contains a DNA binding domain, an RNA polymerase II binding site, multiple sites of serine phosphorylation, and regions that interact with the cytoskeleton (2, 29, 36, 48, 68).

To more fully understand the ways in which the different regions of the v-Abl protein work together to induce transformation, we have examined the ability of the v-Src SH2 domain to substitute for that of v-Abl in the presence and absence of the v-Abl carboxyl terminus. In contrast to results obtained with chimeric proteins containing the entire amino-terminal region of v-Src, our data demonstrate that the v-Src SH2 domain can functionally substitute for the v-Abl SH2 domain only in the presence of a complete carboxyl terminus. Chimeric proteins lacking the extreme carboxyl terminus fail to transform NIH 3T3 cells in vitro, a feature that is correlated to decreased activation of the mitogen-activated protein (MAP) kinase pathway and diminished activation of the c-fos promoter. These data highlight a novel function of this region and suggest that one role of these sequences is to facilitate signaling through the MAP kinase pathway.

MATERIALS AND METHODS

Cells and viruses.

NIH 3T3 cells, Ab-MLV-transformed NIH 3T3 cells, and 293T cells (12) were grown in Dulbecco's modified Eagle's medium (Life Technologies) supplemented with 10% fetal calf serum (Sigma) and 2 mM l-glutamine (Gibco). Ab-MLV-transformed pre-B cells were grown in RPMI 1640 medium (Gibco) supplemented with 10% fetal calf serum, 2 mM l-glutamine, and 50 μM 2-mercaptoethanol (Sigma). Helper-free viral stocks were prepared using Ab-MLV strains in the pSRαMSVtkneo retroviral vector and the pSV-Ψ-E-MLV retroviral packaging plasmid (39). Briefly, 293T cells were plated at 4 × 10⁶ cells per 100-mm-diameter plate the day before transfection. The Ab-MLV plasmid and the packaging plasmid were precipitated, washed, and resuspended in sterile double-distilled water; CaCl₂ was added to a final concentration of 235 mM, and the DNA was added dropwise to the dish. Virus was harvested 36 to 72 h later, pooled on ice, filtered through a 0.45-μm-pore-size filter, and stored at −70°C. To assess the amounts of infectious Ab-MLV in different viral stocks, 10⁵ NIH 3T3 cells were plated in 60-mm-diameter dishes and infected 24 h later with virus in the presence of Polybrene (8 μg/ml; Aldrich). After 48 h, the cells were lysed and the amount of v-Abl protein expressed by the cells was assessed by Western blotting (46). Transformation of NIH 3T3 cells was assessed by growth in soft agar. Cells were infected as described above, trypsinized 2 h later, and plated in an RPMI 1640-based medium containing 10% fetal calf serum and 0.3% agar (Difco) onto a 0.6% agar layer containing RPMI 1640 medium and 10% fetal calf serum. The plates were fed every 7 days, and macroscopic colonies were scored 3 to 4 weeks postinfection.

Construction of viral strains.

pSRαMSVtkneo-Ab-MLV-P120 (pSRα-P120) and pSRαMSVtkneo-Ab-MLV-P70 (pSRα-P70) were constructed by replacing the sequence encoding the v-Abl carboxyl terminus in pSRαMSVtkneo-P160 (39) with the corresponding 2,831-bp BstEII-BspEI fragment from pUC120 and pUC70 (14), respectively. P120Δ668-819 (Fig. 1A) was constructed by PCR using primers that amplified the sequences encoding amino acids 496 to 671 of P120 v-Abl. The 3′ primer contained a SalI site and the amplified material contained the DraIII site at bp 2106 of the Ab-MLV-P120 genome. The amplified material was cloned into the TA vector and sequenced, and the DraIII-SalI fragment was used to replace the sequence encoding amino acids 496 to 819 of P120v-Abl in pSRα-P120. Ab-MLV-P90A has been described previously (40). To construct Ab-MLV strains in which the sequences encoding the v-Abl SH2 domain were replaced by those encoding the v-Src SH2 domain, pSRα viruses and pSAK (22) were used. pSAK encodes a chimeric v-Abl–v-Src protein in which the v-Abl kinase domain is surrounded by v-Src flanking sequences (22). Specifically, the sequences encoding the v-Src SH1 domain (proviral bases 7901 to 8658 of the B77 strain of Rous sarcoma virus) have been replaced with those encoding the v-Abl SH1 domain (proviral bases 2052 to 2956 of Ab-MLV) by using EcoRI linkers. To facilitate exchange of the SH2 domains between pSAK and the pSRα viruses, the intermediate vector TA/P70 was generated by cloning the 2,426-bp SacI fragment (proviral bases −36 to 2391) from pSRα-P120 into the unique SacI site in the TA cloning vector (Invitrogen). The 1.3-kb HincII-EcoRI pSAK fragment (bases 7459 to 8658) encoding the v-Src SH2 domain linked to the v-Abl SH1 domain was used to replace the 1,100-bp HincII-EcoRI fragment of TA/P70, generating TA/P70S2. pSRα-P120S2, pSRα-P120S2Δ668-819, pSRα-P90S2, and pSRα-P70S2 were subsequently generated by replacing the 1.4-kb BstEII-DraIII (bases 725 to 2106) fragment from pSRα-P120, pSRα-P120Δ668-819, pSRα-P90, and pSRα-P70 with the corresponding 1.6-kb BstEII-DraIII fragment from TA/P70S2 which contains the sequences encoding the v-Src SH2 domain linked to v-Abl sequences. The final constructs express chimeric v-Abl–v-Src proteins in which 88 amino acids of v-Abl (residues 241 to 328) have been replaced with 146 v-Src SH2 amino acids plus the four amino acids GINS. The P120ΔSH2 mutant in which the sequences encoding the SH2 domain are deleted was created by replacing the 1,386-bp BstEII-DraIII fragment in pSRα-P120 vector with the corresponding fragment from pAMΔSH2, a plasmid containing the first 708 bases of Mo-MLV gag fused to human c-abl type IV 15 bases into the c-abl coding sequence; the c-abl sequences in this plasmid contain a deletion of c-abl bases which encode the SH2 domain (A. M. Pendergast, personal communication). The Myc-expressing retroviral vector, pSRα-Myc, was created by inserting the 1.8-kb fragment encoding human c-Myc from pBS-myc-mut3 (19) into the EcoRI site of pSRαMSVtkneo.

FIG. 1.

Expression of enzymatically active chimeric viral proteins. (A) Structures of the P120, P120/S2, P70, P70/S2, and P120/K− v-Abl proteins. P120K− encodes a v-Abl protein in which the aspartic acid at position 484 has been replaced by an asparagine (P120/D484N). (B) NIH 3T3 cells were infected with the Ab-MLV strains; lysates were prepared 48 h later and analyzed by Western blotting with the H548 anti-Gag/v-Abl monoclonal antibody (5). The blot was stripped and reprobed with antiphosphotyrosine antibody.

Protein analyses.

Cell lysates were prepared as described previously (45). Briefly, the cells were washed twice with phosphate-buffered saline (PBS), and the cell pellets were treated with lysis buffer (10 mM Tris [pH 7.4], 1% sodium dodecyl sulfate [SDS], 1 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride). The lysates were boiled immediately and sheared through a 25-gauge needle. The amount of protein in each lysate was quantitated using a bicinchoninic acid protein assay kit (Pierce), and 50 μg of each sample was fractionated through an SDS-polyacrylamide gel. In some experiments, equivalent amounts of total cell protein were immunoprecipitated with 1 μg of purified antibody or serum on ice for 1 h. The immune complexes were recovered using protein A- or protein G-Sepharose beads (Pharmacia) and washed with buffer (10 mM sodium phosphate, 150 mM sodium chloride, 1% NP-40, 2 mM EDTA, 1 mM sodium vanadate, 1 mM phenylmethylsulfonyl fluoride). Bound proteins were eluted by heating the beads at 95°C in sample buffer (1% SDS, 50 mM Tris [pH 6.8], 10% glycerol, 0.03% bromophenol blue) for 5 min and analyzed by SDS-polyacrylamide gel electrophoresis. The proteins were electrotransferred to polyvinylidene difluoride membranes (U.S. Biochemicals) which were blocked with PBS containing 0.2% I-block (Tropix) and 0.1% Tween 20 for at least 1 h. Blotting was performed according to the Western-Light kit protocol (Tropix), utilizing alkaline phosphatase-conjugated secondary antibodies with a CSPD substrate (Tropix). Blots were exposed to Kodak XAR-5 film and subsequently stripped by incubating in a pH 2.2 solution containing 0.2 M glycine and 1% Tween 20 for 3 h at 80°C. After stripping, blots were washed with PBS containing 0.1% Tween 20 and treated with blocking solution prior to reprobing. Proteins were analyzed using anti-Gag/v-Abl (H548) (5); antiphosphotyrosine (05-321; Upstate Biotechnology); anti-Shc, anti-Grb2, and anti-Ras (S14630 or S14620, G16720, and R02120, respectively; Transduction Laboratories); anti-Myc (OP10L; Calbiochem); anti-p42/44 MAP kinase, anti-phospho-p42/44 MAP kinase, anti-stress-activated (SAP) kinase/Jun N-terminal kinase (JNK), anti-phospho-SAP kinase/JNK (9120, 9106, 9252, and 9251, respectively; New England Biochemical); and alkaline phosphatase-conjugated anti-mouse and anti-rabbit immunoglobulin G (S372B and S373B, respectively; Promega).

Ras assay.

Levels of RasGTP were assessed using an assay in which RasGTP is recovered through its interaction with a glutathione S-transferase (GST) fusion protein containing the Ras binding domain of Raf (RBD) (10, 21). To prepare the GST fusion proteins, log-phase Escherichia coli BL-21 cells containing pGEX-2T plasmids expressing either GST or GST-RBD were grown for 3 h in the presence of 0.1 mM isopropyl-1-thio-β-d-galactosidase and 12.5 μg of ampicillin per ml. The cells were pelleted, lysed in ice-cold lysis buffer (1× PBS, 1% Triton, 5 mM sodium fluoride, 1 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, 10 μg of leupeptin [Boehringer Mannheim] per ml) and subjected to three cycles of freeze-thawing. The extracts were centrifuged at 12,000 rpm for 30 min at 4°C, and the resulting supernatants were stored at −20°C. The GST proteins were coupled to protein G-Sepharose beads (Pharmacia) according to the manufacturer's protocol. The day before the experiment, 293T cells were plated at 10⁶ cells per 100-mm-diameter plate; the cells were fed with fresh medium 3 h before transfection and then transfected with 15 μg of a v-Abl expression plasmid in the pSRαMSVtkneo vector and 15 μg of a c-Ha-Ras expression vector (62). For precipitation, 200 to 1,000 μg of cell lysate was incubated with 25 μl of packed beads prebound to either GST or GST-RBD and rocked at 4°C for 90 min. The beads were washed three times with wash buffer (50 mM Tris [pH 7.5], 20 mM NaCl, 10 mM MgCl₂, 0.5% NP-40), and the bound proteins were eluted with SDS sample buffer, resolved by electrophoresis, and analyzed by Western blotting.

Transfection assays.

The day before the experiment, 293T cells were plated at 2.5 × 10⁵ cells per 60-mm-diameter plate and fed with fresh medium 3 h before transfection. For experiments using luciferase expression vectors, the cells were transfected with 1 μg of pSVOAΔ5′, which contains a 379-bp murine c-fos promoter upstream of the firefly luciferase gene (20), 4 μg of a v-Abl expression plasmid in the pSRαMSVtkneo vector, and 0.7 μg of pRL-TK, an internal control reporter plasmid expressing renilla luciferase (Promega). In some experiments, pSVOAΔ5′ was replaced with plasmids that contained mutations in three of the cis-acting elements of the murine c-fos promoter as described previously (28) (see Results). Other experiments included 4 μg of a fourth DNA which encoded either dominant-negative (DN) Ras (DN-Ras; Ras/S17N) (62) or DN-Akt (Akt K179M) (13). The transfections were performed as described earlier. The medium was changed 8 to 12 h posttransfection, and the cells were fed with 5 ml serum-free medium 24 h posttransfection. Luciferase activity was measured using the Dual-Luciferase reporter assay system (Promega) according to the manufacturer's protocol. Results were normalized to the expression of renilla luciferase. These transfections were carried out in triplicate. In other experiments, 293T cells were transfected with 4 μg of pJ3M-Erk, expressing Myc-tagged Erk from the simian virus 40 promoter (L. Feig, personal communication), or pcDNA, expressing Flag-Jnk1 (17), and 4 μg of a v-Abl expression plasmid in the pSRαMSVtkneo vector. Cells were serum starved, and lysates were prepared 24 h later. The samples were analyzed by Western blotting.

RESULTS

P70/S2 fails to transform NIH 3T3 cells.

Previous studies had demonstrated that the v-Src amino-terminal region, comprising the SH3 and SH2 domains, can function in concert with the v-Abl kinase domain and mediate transformation of NIH 3T3 cells (22). To test the ability of the v-Src SH2 domain to supply functions normally provided by the v-Abl SH2 domain, chimeric Ab-MLV strains in which the v-Abl SH2 domain was replaced with the SH2 domain of v-Src were constructed using both the P120 wild-type background and that of the P70 carboxyl-terminal truncation mutant (Fig. 1A). The v-Abl protein encoded by this latter strain is similar to v-Src in that it terminates at the end of the kinase domain and lacks the carboxyl-terminal region characteristic of Abl proteins (14). To confirm that all of the viruses expressed proteins of the expected size, NIH 3T3 cells were infected and 48 h postinfection cell lysates were analyzed by Western blotting. All of the samples contained proteins of the expected size and the levels of expression were similar (Fig. 1B). In addition, total cellular phosphotyrosine levels were elevated in all infected cells except those infected with P120/D484N, a control virus expressing a kinase-negative v-Abl protein (52). These data suggest the kinase activity of the chimeric proteins is similar to that of v-Abl protein.

Although both the P120/S2- and P70/S2-infected cells displayed elevated levels of cellular phosphotyrosine, only those infected with the P120/S2 strain displayed the transformed morphology characteristic of cells infected with Ab-MLV-P120 or P70 (Fig. 2). Cells infected with the P70/S2 strain failed to form foci of rounded up cells characteristic of Ab-MLV-transformed cells. However, the cells assumed a more fusiform appearance and grew in a more irregular pattern compared to uninfected cells or cells infected with the P120/D484N strain (Fig. 2). To investigate this difference further, virus stocks were standardized by assessing levels of Abl proteins expressed in NIH 3T3 cells after a 48 h infection and tested for the ability to induce transformation in soft agar. Consistent with the appearance of the infected cells, similar numbers of colonies were obtained from cultures infected with the P120, P120/S2, and P70 strains (Table 1); in contrast, no colonies were evident in cells infected with the P70/S2 strain. These results suggest that the v-Abl SH2 domain plays a vital role in v-Abl-mediated fibroblast transformation and that the carboxyl terminus is able to complement this function in the P120/S2 strain.

FIG. 2.

NIH 3T3 cells infected with the P70/S2 are not morphologically transformed. NIH 3T3 cells were infected with the Ab-MLV strains and photographed 48 h later.

TABLE 1.

P70/S2 fails to transform NIH 3T3 cells^a

Virus	Colonies/105 cells ± SEM
	Expt 1	Expt 2
P120	121 ± 16	86.5 ± 6.9
P120/S2	226.5 ± 12	124.8 ± 4.3
P70	103.5 ± 13	69.5 ± 4.3
P70/S2	<1	<1
P120K−	<1	<1

^aCells were infected with virus stocks of matched titer and plated in agar following virus adsorption. Macroscopic colonies were counted 4 weeks after the cells were plated. Values given as <1 indicate that no colonies were observed and reflect the minimum number of colonies that could have been detected. 

Shc phosphorylation and Grb-2 association are decreased in NIH 3T3 cells infected with the P70/S2 strain.

v-Abl association with Shc and the formation of a Shc-Grb2 complex has been implicated as one pathway by which the v-Abl protein activates Ras and mediates transformation (44, 52). To examine whether the chimeric viral proteins were associated with Shc, lysates were prepared from infected NIH 3T3 cells. Western analysis of samples immunoprecipitated with anti-Abl antibody and probed with anti-Shc antibody revealed that Shc associated with all of the v-Abl proteins, including P70/S2 (Fig. 3A). To examine whether Shc was tyrosine phosphorylated and associated with Grb2, NIH 3T3 cells were infected and lysates were prepared 48 h later. Western analysis of samples immunoprecipitated with anti-Shc antibody and probed with antiphosphotyrosine and anti-Grb2 antibodies revealed that Shc was tyrosine phosphorylated and associated with Grb2 in cells that had been infected with the P120, P120/S2, and P70 strains but not in cells that had been infected with the P70/S2 strain (Fig. 3B). Thus, the presence of a full COOH terminus enhances the ability of a v-Abl protein containing the Src SH2 domain to interact with Shc. These results raise the possibility that the absence of the Shc/Grb2 signal contributes to the absence of transformation in cells expressing P70/S2.

FIG. 3.

P70/S2 does not mediate Shc phosphorylation and Grb2 association. NIH 3T3 cells were infected and lysed. (A) The v-Abl proteins were immunoprecipitated with the anti-Gag/v-Abl monoclonal antibody H548 (5) and analyzed by Western blotting with the H548 antibody; the blot was then stripped and reprobed with an anti-Shc antibody. (B) The lysates were immunoprecipitated with anti-Shc antibodies and analyzed by Western blotting. The upper portion of the blot was probed with anti-Shc antibodies; the blot was stripped and reprobed with an antiphosphotyrosine (α-pTyr) antibody. The lower portion of the blot was probed with anti-Grb2 antibody. CTRL, control.

P70/S2 can stimulate increased RasGTP levels.

Expression of DN-Ras inhibits transformation of NIH 3T3 and bone marrow cells by the virus (62). To determine if expression of P70/S2 is capable of activating Ras, 293T cells were transfected with DNAs encoding the different v-Abl proteins and a c-Ha-Ras expression plasmid and serum starved for 24 h. Cell lysates were prepared, and the levels of activated Ras were assessed using a GST-RBD fusion protein (21). This protein contains the RBD of Raf, a region which binds preferentially to RasGTP. Western analysis of the GST-RBD complexes revealed that all of the kinase-active v-Abl proteins tested, when expressed at comparable levels (Fig. 4), stimulated Ras activity above the background level observed with the P120/D484N mutant. Although this assay does not monitor Ras activation in NIH 3T3 cells, these results suggest that differences in Ras activation do not explain the different biological phenotypes associated with the different chimeric proteins.

FIG. 4.

P70/S2 stimulates increased RasGTP levels. 293T cells were transfected with 15 μg of plasmids expressing the different v-Abl forms and 15 μg of a c-Ha-Ras expression vector. The cells were serum starved for 24 h, and extracts were prepared 48 h after transfection. (A) Portions of the lysate were analyzed by Western blotting with the H548 monoclonal antibody (5). (B) Portions of the lysates were incubated with GST or GST-RBD, and the affinity precipitates were analyzed by Western blotting with anti-Ras antibody (10, 21). The signals obtained were analyzed by densitometry, and the RasGTP levels were normalized to the levels of viral protein. The value for the P120 sample was set as 1.

P70/S2 is compromised in activation of ERK but not JNK.

Activated Ras transmits signals which activate both the MAP and SAP kinase pathways (11, 37). Both of these pathways are activated by v-Abl (53, 55). To test the ability of the chimeric v-Abl proteins to activate ERK and JNK, 293T cells were cotransfected with DNAs encoding the viral proteins and expression plasmids encoding either ERK or JNK. Analysis of cell extracts prepared from serum-starved cells revealed that cells expressing P70, P120, and P120/S2 contained elevated levels of phosphorylated ERK. However, levels of phosphorylated ERK recovered from cells expressing P70/S2 were similar to those recovered from cells expressing P120/D484N (Fig. 5A). In contrast to these results, levels of phosphorylated JNK were recovered from all cells expressing kinase active v-Abl proteins (Fig. 5B). These results suggest P70/S2 is compromised in its ability to activate ERK but not JNK and that the carboxyl terminus is able to complement this function. These results also suggest that activation of the JNK pathway alone is not sufficient to induce NIH 3T3 cell transformation.

FIG. 5.

P70/S2 activates JNK but not ERK. 293T cells were transfected with 4 μg of plasmids expressing the v-Abl proteins and 4 μg of either pJ3M-ERK (A) (L. Feig, personal communication) or pcDNA expressing Flag-Jnk1 (17) (B). The cells were serum starved for 24 h, and extracts were prepared 48 h after transfection. Equivalent amounts of extract were analyzed by Western blotting. The blots were probed with the anti-Gag/v-Abl H548 monoclonal antibody (5) and anti-ERK or anti-JNK antibodies and reprobed with either anti-phospho-ERK or anti-phospho-JNK antibodies. The pattern of degradation observed for the v-Abl proteins in panel A is not typical.

P70/S2 fails to transactivate the c-fos promoter.

c-fos is one important gene stimulated by signals from the MAP and SAP kinase pathways (27, 66). To test the ability of the chimeric proteins to activate transcription from the c-fos promoter, 293T cells were transfected with DNAs encoding the viral proteins and a plasmid expressing firefly luciferase under the control of the c-fos promoter. The pRL-TK vector encoding renilla luciferase was included as a transfection control. The cells were serum starved, and luciferase activity was analyzed using the Dual-Luciferase reporter assay system (Promega). As expected, the P120 protein stimulated luciferase activity (55), as did expression of P120/S2 and P70 (Fig. 6A). However, in these latter instances, levels of luciferase activity were twofold lower than those recovered from cells expressing the P120 protein. In contrast to these results, expression of P70/S2 failed to stimulate luciferase expression above the background levels obtained with P120/D484N. These data suggest that P70/S2 is compromised in its ability to transactivate the c-fos promoter. The ability of P120/S2 to stimulate the c-fos promoter suggests that the carboxyl terminus is able to complement functions normally provided by the v-Abl SH2 domain with respect to c-fos activation.

FIG. 6.

v-Abl transactivates the c-fos promoter via Ras and the SRE. 293T cells were transfected with 4 μg of a pSRα v-Abl expression plasmid, 1 μg of different reporter plasmids, and 0.7 μg of pRL-TK, an internal control reporter plasmid. The cells were serum starved for 24 h, and extracts were prepared 48 h after transfection. Luciferase activity was measured using the Dual-Luciferase reporter assay system (Promega). Each transfection was done in triplicate. For each replicate, firefly luciferase activity was normalized to renilla luciferase activity over three or more serial dilutions and averaged for each sample. The error bars reflect standard deviations. (A) Cells were transfected as described above with the pSVOAΔ5′ reporter plasmid, which contains a 379-bp murine c-fos promoter upstream of the firefly luciferase gene (20). (B and C) Cells were transfected with plasmids that contained mutations in various cis-acting elements (28): mSIE, mutant SIE; mCRE, mutant CRE; mAP-1, mutant AP-1 binding site; mSRE, mutant SRE; mTCF, mutant TCF binding site. WT, wild type. (D) Cells were transfected with pSVOAΔ5′ and plasmids encoding DN-Ras (62) or DN-Akt (13).

P120/S2 and P70 mediate c-fos activation via the cis-acting SRE element and Ras.

The promoter region of the c-fos gene contains various cis-acting elements including a cyclic AMP-responsive element (CRE), an AP-1-responsive element, a c-sis–platelet-derived growth factor-inducible element (SIE), and a serum-responsive element (SRE) (3, 61, 69). The SRE contains binding sites for both ternary complex factors (TCFs) and serum response factors (SRFs) (43, 63, 67). The TCFs bind to the SRE only in the presence of SRF or its core subdomain (63), while the SRFs can activate c-fos independently of the TCFs (24). The SRE is the major site at which Ras-dependent MAP kinase signals are integrated at the c-fos promoter (9, 27, 37), and others have demonstrated the ability of v-Abl to transactivate promoters containing SREs (55, 56).

To determine which cis-acting elements in the c-fos promoter are required for transactivation by the P70 v-Abl protein, 293T cells were transfected with DNAs encoding v-Abl proteins and luciferase reporter plasmids under the control of c-fos promoters containing mutations in the various upstream elements. Mutation of the SIE had no effect on luciferase expression, and mutation of either the AP-1 or CRE site had minimal effects. In contrast, mutation of the SRF binding site within the SRE ablated transaction of the c-fos promoter completely (Fig. 6B). These results indicate that the SRE is the principal site at which signals from v-Abl proteins integrate at the c-fos promoter. Mutation of the TCF binding site within this region also reduced transactivation significantly (Fig. 6C), suggesting that an SRF/TCF-dependent mechanism is involved in v-Abl-mediated transactivation of the c-fos promoter. Consistent with the role of Ras in transducing signals to SREs (58), expression of DN-Ras but not DN-Akt ablated induction of the c-fos promoter (Fig. 6D). This result suggests that signals transmitted via Ras play a critical role in the response.

Complementation by the carboxyl terminus maps to the extreme COOH terminus.

The ability of P120/S2, but not P70/S2, to transform NIH 3T3 cells and stimulate c-fos expression demonstrates that functions contributed by the carboxyl terminus of v-Abl can complement functions normally provided by the v-Abl SH2 domain. To define the region within the carboxyl terminus required for SH2 complementation, the P90/S2 and P120/S2Δ668-819 Ab-MLV strains were constructed. The P90/S2 protein contains the v-Src SH2 domain and the first 144 amino acids of the carboxyl terminus present in P120 v-Abl; the P120/S2Δ668-819 protein contains the v-Src SH2 domain but lacks the carboxyl-terminal amino acids 668 to 819 as a consequence of an in-frame coding sequence deletion (Fig. 7). The Ab-MLV strains expressing these proteins were tested for the ability to induce NIH 3T3 transformation in soft agar assays. As expected, colonies were readily detected following infection with Ab-MLV-P90 (59). However, the P90/S2 strain was similar to the P70/S2 strain and did not induce colony formation (Table 2). In contrast, P120/S2 and P120/S2Δ668-819 induced similar numbers of colonies. Consistent with these data, both P120/S2 and P120/S2Δ668-819 activated expression of the c-fos promoter, while P90/S2 failed to stimulate activity above the background levels obtained with the kinase-inactive P120/D484N protein (data not shown). Thus, sequences within the last 162 carboxyl-terminal amino acids are required for SH2 domain complementation, and at least one function mediated by these sequences leads to activation of c-fos expression.

FIG. 7.

Structures of the P90, P90/S2, P120Δ668-819, P120/S2Δ668-819, and P120 viral proteins.

TABLE 2.

Carboxyl-terminal sequences are required in cis^a

Virus	Colonies/105 cells ± SEM
	Expt 1	Expt 2
P120	130.8 ± 9.2	229.6 ± 6.81
P70/S2	<0.1	<0.1
P120/ΔSH2	<0.1	<0.1
c-Myc	<0.1	<0.1
P120 + c-Myc	367.2 ± 29.32	468.4 ± 19.46
P70/S2 + c-Myc	10.4 ± 1.78	2.1 ± 0.5
P120/ΔSH2 + P70/S2	<0.1	<0.1

^aCells were infected with virus stocks of matched titer and plated in agar following virus adsorption. Macroscopic colonies were counted 4 to 5 weeks after the cells were plated. Values given as <0.1 indicate that no colonies were observed and reflect the minimum number of colonies that could have been detected. 

Carboxyl-terminal sequences are required in cis.

To determine whether the carboxyl-terminal v-Abl sequences were required in cis or trans for complementation of SH2 function, the ability of the transformation-defective P120/ΔSH2 strain to restore transforming function to P70/S2 was tested. The P120/ΔSH2 strain expresses a v-Abl protein that contains a complete carboxyl terminus but from which sequences encoding the SH2 domain have been deleted (G. Raffel, personal communication). NIH 3T3 cells were also infected with pSRα-Myc, a retrovirus which expresses the c-Myc protein, since overexpression of c-Myc has been shown to complement the transformation-defective phenotype of a variety of abl alleles (1). NIH 3T3 cells were infected, either singularly or in combination, with P70/S2, P120/ΔSH2, P120, and pSRα-Myc and plated in soft agar (Fig. 8A). Colonies were scored 3 to 4 weeks later.

FIG. 8.

P70/S2 is not complemented by P120ΔSH2. NIH 3T3 cells were infected, either singularly or in combination, with retroviruses viruses expressing different v-Abl proteins or c-Myc protein. (A) Samples of infected cells were lysed 48 h postinfection, and levels of the different v-Abl proteins were assessed by Western blotting with the anti-Gag/v-Abl monoclonal antibody H548 (5). (B) Colonies obtained from doubly infected populations were expanded and analyzed by Western blotting with the anti-Gag/v-Abl monoclonal antibody H548 and an anti-Myc antibody.

When used individually, the P70/S2, P120/ΔSH2, and pSRα-Myc strains all failed to transform NIH 3T3 cells. Coexpression of both P120/ΔSH2 and P70/S2 failed to induce transformation in the agar assay, while coexpression of c-Myc weakly complemented P70/S2 in the agar transformation assay (Table 3). Colonies obtained from cells coinfected with c-Myc and P70/S2 were screened via Western analysis to confirm the expression of both proteins (Fig. 8B). Although some samples contained lower amounts of both proteins, probably reflecting smaller colony size, all expressed both proteins. These data demonstrate that carboxyl-terminally mediated complementation of v-Abl SH2 function requires the carboxyl terminus in cis. In addition, overexpression of c-Myc only partially complemented P70/S2, suggesting that defective signaling to c-Myc may be partly, but not wholly, responsible for the transformation defect of P70/S2.

TABLE 3.

P90/S2 fails to transform NIH 3T3 cells^a

Virus	Colonies/105 cells ± SEM
	Expt 1	Expt 2
P90	171.5 ± 16.0	97.5 ± 7.6
P90/S2	<1 ± 0.1	<1 ± 0.1
P120Δ668-819	112.0 ± 19.0	94.3 ± 9.0
P120/S2Δ668-819	190.0 ± 12.0	98.8 ± 5.8
P120K−	<1 ± 0.1	<1 ± 0.1

^aCells were infected with virus stocks of matched titer and plated in agar following virus adsorption. Macroscopic colonies were counted 4 to 5 weeks after the cells were plated. Values given as <1 indicate that no colonies were observed and reflect the minimum number of colonies that could have been detected. 

DISCUSSION

Our analysis of chimeric v-Abl/v-Src proteins demonstrates that sequences within the carboxyl terminus of the v-Abl protein can complement functions normally supplied by the v-Abl SH2 domain. As shown previously for chimeras in which the v-Src SH2 domain was inserted into other full-length, transforming Abl proteins (33, 35), the v-Src SH2 region functions well in the context of the P120 strain. However, even though Ab-MLV strains encoding carboxyl-terminally truncated v-Abl proteins transform NIH 3T3 cells efficiently (14, 23, 50, 58), substitution of the v-Src SH2 domain in this context abolishes transformation competency. Although the carboxyl terminus has long been appreciated as playing an important role in lymphoid transformation (23, 40, 50, 58), this is the first demonstration that the region can influence transformation of immortalized fibroblast cell lines.

The inability of the P70/S2 strain to transform NIH 3T3 cells demonstrates that the v-Abl and v-Src SH2 domains are not functionally equivalent. Earlier work (22), using different chimeras, reached the opposite conclusion. However, in these experiments, the entire amino terminus of v-Src was appended to a truncated v-Abl protein similar to P70. Consequently, these chimeras also contained the v-Src SH3 domain. Several Src SH3-binding proteins are phosphorylated in v-Src transformed cells but not in cells transformed by a variant v-Src allele lacking the SH3 domain (70). Thus, the Src SH3 domain plays a role in v-Src-mediated tyrosine phosphorylation of substrates, and its presence may influence transformation potential in NIH 3T3 cells.

Consistent with the inability of P70/S2 to transform cells, several downstream signals associated with Ab-MLV transformation are missing in cells expressing this protein. For example, tyrosine phosphorylation of the Shc adapter protein and association with Grb-2 does not occur. This interaction is believed to be one way in which Abl proteins activate Ras (16, 44, 52). The p42 and p44 ERK proteins, elements that function downstream of Ras and are critical components of the MAP kinase cascade, are not phosphorylated in cells expressing P70/S2. Consistent with this, P70/S2 is compromised in its ability to activate transcription from the c-fos promoter. All of these events are thought to be important for v-Abl-mediated transformation (45, 62). These data contrast with the ability of P70/S2 to stimulate Ras activation. Because Ras activation was examined using a transfection system, the results may not reflect events occurring in the NIH 3T3 target cells. Alternatively, Ras activation can be achieved in many ways, raising the more intriguing possibility that the way in which Ras is activated influences transmission of downstream signals.

SH2 domains contribute significantly to the specificity with which tyrosine kinases activate downstream effectors (31, 66), and the SH2 domains of v-Src and v-Abl have been placed in different subgroups based on their ability to interact with tyrosine phosphorylated peptides (66). The ability of the v-Abl SH2 domain but not the Src SH2 domain to interact with the Shc adapter protein (31, 44, 52) may reflect this property. The inability of P70/S2 to mediate phosphorylation of Shc suggests that one pathway by which v-Abl normally signals to Ras is not functional in these cells, a feature that may contribute to the transformation defective nature of this strain.

Although Shc phosphorylation and assembly of the Shc-Grb2-Sos complex is classically associated with Ras activation, these proteins may mediate other types of responses. For example, complex formation does not always correlate with MAP kinase activation (44, 49), and injection of Grb2 antibodies inhibits membrane ruffing and cell growth in response to epidermal growth factor, even though microinjection of anti-Ras antibodies affects only growth (32, 57). In addition, dShc, the Drosophila homologue, lacks the residue analogous to the mammalian Grb-2 interaction site, and no complex between dShc and Drk, the Drosophila Grb-2 homologue, can be detected (30). Considered in combination with the finding that Ras can be activated by P70/S2, this information raises the possibility that the absence of Shc-Grb2-Sos complexes in cells expressing P70/S2 reflects the loss of other functions important for v-Abl-mediated transformation.

P70/S2 was unable to activate transcription from a c-fos promoter, a response that involves interactions at both the SRF and TCF sites within this element. SRF activation does not appear to require the JNK and ERK MAP kinases; TCF can be activated by both of these kinases (4). The inability of P70/S2 to activate ERK may influence the TCF activity and contribute to the transformation-defective phenotype of the P70/S2 strain. If this is indeed the case, interactions with SAP-1 might be involved. This protein appears to be important in murine cell lines such as NIH 3T3 cells and is activated in a Ras-independent manner with minimal influence of JNK (25).

The final 162 amino acids of the v-Abl carboxyl terminus are sufficient to complement the transformation defect in P70/S2 and to restore Shc phosphorylation. Analyses of other mutants suggest that the carboxyl terminus does not bind Shc (52); because the region appears to be required in cis, it may simply stabilize the interaction between v-Abl and Shc. Some proteins, including Bcr/Abl, can recruit Grb2-Sos complexes directly (47, 51), bypassing a need for Shc interaction. However, Bcr/Abl-Grb2 interaction is mediated by residues in the Bcr portion of the protein (47, 51), and sequences within the Abl carboxyl terminus have not been shown to interact with Grb2. Although the effects on Shc and Grb2 are striking, they may not be the critical feature involved. The extreme carboxyl terminus contains motifs that mediate interactions with the cytoskeleton (36, 68). Other studies have shown that this region binds RNA polymerase II, facilitating phosphorylation of this protein by c-Abl (2). While this nuclear event is probably not involved in Ab-MLV-mediated transformation, it does highlight a role for this region in protein-protein interaction. Unraveling the mechanism by which these sequences affect transformation should shed light on the function of the v-Abl carboxyl terminus. Because the chimeric v-Abl proteins studied here display strong phenotypic differences in a readily manipulable cell type, they provide an excellent model to uncover the mechanism underlying the function of the carboxyl-terminal residues.