Identification and characterization of two novel SH2 domain ...

Oncogene (1997) 15, 1255 ± 1262  1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00

Identi®cation and characterization of two novel SH2 domain-containing proteins from a yeast two hybrid screen with the ABL tyrosine kinase Tsukasa Oda1, Jody Kujovich1, Margaret Reis1, Brenda Newman2,3 and Brian J Druker1,2,4 1

Division of Hematology and Medical Oncology, 2Department of Cell and Developmental Biology, 3Vollum Institute for Advanced Biomedical Research and 4Department of Biochemistry and Molecular Biology, Oregon Health Sciences University, Portland, Oregon 97201, USA

To further our understanding of the molecular mechanism of Bcr ± Abl mediated transformation, a yeast two hybrid screen was used to identify proteins binding to the Abl tyrosine kinase. Two partial cDNAs encoding novel SH2 domain-containing proteins were cloned and designated Shd and She. Both have homology to Shb, a previously reported SH2 domain-containing protein. Northern blot analysis showed that She is expressed in heart, lung, brain, and skeletal muscle, while expression of Shd is restricted to the brain. The deduced amino acid sequence of the full length mouse Shd cDNA contains an amino-terminal proline-rich region, and a carboxyterminal SH2 domain. A bacterially expressed Shd domain bound multiple tyrosine-phosphorylated proteins with relative molecular weights of 200, 170, 130, 100, 90, 78, 72 and 32 kDa from K562 cell lysates. Shd contains ®ve YXXP motifs, a substrate sequence preferred by Abl tyrosine kinases. Shd was tyrosine phosphorylated in COS-7 cells co-transfected with Shd and c-Abl or Bcr ± Abl. These results suggest that Shd may be a physiological substrate of c-Abl and may function as an adapter protein in the central nervous system. Keywords: c-Abl; tyrosine kinase; SH2 domains; yeast two hybird screen

Introduction Src homology 2 (SH2) domains are phosphotyrosine binding motifs implicated in the regulation of proteinprotein interactions. SH2 domains are found in a diverse group of proteins involved in signaling pathways for cell growth and dierentiation, and are thought to function as molecular adhesives facilitating the formation of protein complexes (Koch et al., 1991). SH2 domain binding to speci®c phosphotyrosine containing sequences may transmit intracellular signals by two dierent mechanisms. First, the interaction may induce conformational changes that alter an enzyme's catalytic activity. For example, binding of the Src SH2 domain to a regulatory tyrosine phosphorylation site on Src (Tyr 527) represses its tyrosine kinase activity (Cooper and Howell, 1993). Other SH2 domain-containing enzymes such as the tyrosine phosphatase Syp are activated by binding to phosphopeptides (Cohen et al., 1995). Secondly, the interaction of an SH2 protein with phosphotyrosine Correspondence: BJ Druker Received 28 February 1997; accepted 23 May 1997

may alter its subcellular localization. Many growth factor receptors, such as epidermal growth factor (EGF) receptor, platelet-derived growth factor (PDGF) receptor, and colony-stimulating factor-1 receptor, undergo autophosphorylation upon binding of their respective ligands, which creates speci®c tyrosine binding sites for the SH2 domains of cytoplasmic signaling proteins (Schlessinger and Ullrich, 1992). It is well known that several SH2 proteins, including Grb2, Shc, PI-3K and GAP are recruited to the plasma membrane by binding to phosphotyrosine residues on activated growth factor receptors (Cohen et al., 1995). Although many dierent SH2 proteins have been cloned and partially characterized, the identi®cation of novel SH2 proteins is important for understanding cell signaling pathways. c-Abl is a ubiquitously expressed tyrosine kinase whose physiological function is largely unknown. Although c-Abl is primarily localized to the nucleus, some is associated with the plasma membrane or with actin ®laments (McWhirter and Wang, 1991; Van Etten et al., 1989). c-Abl contains amino terminal SH3 and SH2 domains followed by the kinase domain. The Cterminus of Abl is unique among tyrosine kinases, being over 600 amino acids long and containing a DNA-binding domain, an actin-binding domain and a nuclear localization signal (Jackson and Baltimore, 1989; McWhirter and Wang, 1991; Van Etten et al., 1989). c-Abl kinase activity appears to be regulated during the cell cycle (Welch and Wang, 1993) and may be activated by DNA damage or cellular adhesion (Kharbanda et al., 1995; Lewis et al., 1996). Oncogenic activation of c-Abl can occur by deletion of the SH3 domain (Jackson and Baltimore, 1989) or fusion to amino terminal sequences, most commonly Bcr (BenNeriah et al., 1986; Groen et al., 1984; Heisterkamp et al., 1985). The transforming variants of Abl are largely cytoplasmic (Van Etten et al., 1989) and are constitutively active as tyrosine kinases (Daley et al., 1987; Davis et al., 1985; Lugo et al., 1990). Although numerous proteins, including Bcr ± Abl itself, are tyrosine phosphorylated in Abl transformed cells, the role of many of these proteins in Bcr ± Ablmediated transformation is not clear. Of these tyrosine phosphorylated proteins, only the binding sites for Grb-2, which binds to Tyr 177 of Bcr and CrkL, which binds to a proline rich sequence in the C-terminus of Abl, have been mapped (Heaney et al., 1997; Pendergast et al., 1993; Puil et al., 1994; Ren et al., 1994). Several proteins that bind to c-Abl have been reported, including the retinoblastoma gene product and p53 (Goga 1995; Welch and Wang, 1993; Wen et al., 1996). In addition, RNA polymerase II has been

Shd and She, novel SH2 domain-containing proteins T Oda et al

1256

demonstrated to bind to and be tyrosine phosphorylated by c-Abl (Baskaran et al., 1993, 1996). However, the precise role of these proteins in mediating Abl functions is not clear. Since proteins interacting with Abl are likely to be required for Abl function, the ®rst step in understanding Abl function is to identify proteins that interact directly with Abl or are substrates for phosphorylation. We used the yeast two hybrid system to identify proteins that bind to the Abl kinase domain. We previously reported that most of the interacting proteins were CrkL and Crk II (Oda et al., 1994). Other positive clones included the SH3/SH2 domains of Src, Fyn and Abl. We also identi®ed two novel SH2 domain-containing proteins which we designated Shd and She. The cDNA sequences of Shd and She have homology to Shb, a previously reported SH2 protein (Welsh et al., 1994). However, in contrast to Shb which is widely expressed, She is expressed in heart, lung, brain, and skeletal muscle, and the expression of Shd is restricted to the brain. The deduced amino acid sequence of the full length Shd contains ®ve YXXP motifs, which is the preferred phosphorylation site for the Abl kinase (Songyang et al., 1995). Shd was tyrosine phosphorylated in COS 7 cells co-transfected with Shd and the kinase active form of c-Abl or Bcr ± Abl. These results suggest that Shd may be a physiological substrate of c-Abl, and may function as an adapter protein in the central nervous system.

Results Identi®cation of two novel SH2 proteins, Shd and She We used a yeast two hybrid screen to identify proteins that interact with c-Abl. The c-Abl fragment fused to the LexA DNA binding domain used as a bait is constitutively active as a tyrosine kinase. Analysis of phosphotyrosine immunoblots of yeast lysates after transformation with the bait plasmid showed phosphorylation of the LexA ± Abl hybrid protein, as well as a variety of endogenous proteins, compared to control lysate (data not shown). Approximately 6.76106 recombinants from a mouse embryo pVP16 cDNA library were screened. A yeast strain containing both HIS3 and lacZ gene coding sequences driven by promoters fused to LexA DNA binding sites was used as a reporter strain (Vojtek et al., 1993). In this system, transactivation of the reporter constructs requires an interaction between the proteins fused to the LexA and VP16 domains. A yeast strain expressing two interacting hybrid proteins becomes prototrophic for histidine and expresses b-galactosidase. Interacting clones were identi®ed by selection on plates lacking histidine and analysed for b-galactosidase activity. After elimination of false positive clones by mating analysis, 113 clones interacting with LexA ± Abl were identi®ed by this screen (Table 1). Sequence analysis of positive interacting clones indicated that the majority were CrkII and CrkL. Others included the SH3/SH2 domains of Src, Fyn and Abl. Two non-identical novel sequences were also identi®ed which both encoded SH2 domains with homology to Shb, a previously reported SH2 domain-containing protein

(Welsh et al., 1994). We named these novel SH2 domain-containing proteins `Shd' and `She'. The Shd clone and several other SH2 domain-containing proteins (Src, Fyn, Abl) did not interact with the kinase defective Abl mutant (Table 1). This indicates that the interaction between Shd and Abl is kinase dependent, and most likely involves binding of the Shd SH2 domain to phosphotyrosine on Abl kinase. In contrast, the She clone, as well as the CrkII and CrkL clones, did interact with the kinase defective Abl mutant in this system (Table 1). Analysis of the amino acid sequence of She did not reveal an obvious SH3 domain which was previously shown to be responsible for the kinase independent interaction between CrkL, CrkII and Abl (Heaney et al., 1997; Ren et al., 1994). Since both novel cDNAs identi®ed by the yeast two hybrid screen were partial sequences encoding the SH2 domain, we attempted to clone the complete cDNA sequences of the Shd and She proteins. Using the partial Shd sequence as a probe, we cloned the full length Shd cDNA. Despite several library screening attempts, we have not yet cloned the full length She cDNA. The complete 1543 bp Shd cDNA sequence and the deduced amino acid sequence is shown in Figure 1. The open reading frame contains 343 amino acids. Since there is an in-frame stop codon located in the 5' non-coding region upstream of the open reading frame encoding the SH2 domain, we identi®ed the ®rst ATG codon after the stop codon as the translation initiation site. A previous study demonstrated that the Shb protein is expressed in both 67 kD and 56 kD forms which most likely result from two alternative sites of initiation of translation (Welsh et al., 1994). Comparison of the amino acid sequences of Shd and Shb suggests that the translation initiation site of Shd corresponds to that of the 56 kD Shb protein. The deduced amino acid sequence of the full length mouse Shd cDNA contains an amino-terminal proline-rich region, and a carboxy-terminal SH2 domain. Comparison of the Shd SH2 sequence with other SH2 proteins revealed the greatest degree of homology with Shb (Figure 2).

Tissue expression of Shd and She Previous studies showed that Shb is ubiquitously expressed in all tissues and cell lines tested (Welsh et al., 1994). We therefore investigated the tissue distribution of Shd and She expression. Northern blot analysis of poly (A)+ RNA from a variety of murine tissues showed expression of Shd in the brain as an

Table 1 Results of the two hybrid screen with the Abl kinase domain Kinase independent Protein Crk II CrkL Abl Ubiquitin conjugating enzyme U1snRNP She Unknown

Number of clones 40 39 7 3 5 3 1

Kinase dependent Protein

Number of clones

Src Fyn Abl gag-related protein

7 10 2 2

Shd Unknown

1 10


1257

Figure 1 Nucleotide and deduced amino acid sequence of murine Shd. A full length mouse SHD cDNA was cloned by screening a lSHlox 16 day mouse embryo cDNA library (Novagen) with a NotI fragment (in italics) of pVP16 Shd. Both strands of the Shd cDNA were sequenced. The open reading frame, containing 343 amino acids, is shown with a one-letter symbol. An in-frame stop codon in the 5'-¯anking non-coding region (TGA at position 259) and the SH2 domain are underlined

Figure 2 Amino acid sequence comparison between Shb, Shd, and She. Full length Shb, Shd and partial sequences of She are aligned. Identical or similar amino acids are boxed. A proline rich region is double-underlined and YXXP motifs, substrate sequences preferred by the ABL kinase, are underlined. The beginning of the SH2 domain is marked with an arrow


approximately 1.6 kb transcript, and expression of She in heart, lung, brain, and skeletal muscle as an approximately 7.3 kb transcript (Figure 3). These results suggest more restricted expression of both Shd and She compared with that of Shb. We also examined several dierent cell lines for expression of Shd. A rat

adrenal pheochromocytoma cell line (PC12), a mouse myeloid cell line (32D) and a mouse ®broblast cell line (A31), were individually screened by immunoblotting with anti-Shd and by Northern blotting and neither protein nor RNA was detected.

testis

kidney

skeletal muscle

liver

lung

spleen

brain

Shd contains a functional SH2 domain

heart

1258

7.5 kb —

She

2.4 kb — Shd 1.35 kb — Figure 3 Northern Blot Analysis of the Expression of She and Shd. A multiple tissue Northern Blot (CLONTECH) was hybridized with 32P-labeled anti-sense Shd or She RNA as a probe. The migration of molecular weight markers is indicated on the left of the panel

GST

GST-SH2 /SHD 200 (kD) 170 130 100 90 78 72

The kinase dependent interaction between Abl and Shd in the yeast two hybrid system suggests the SH2 domain of Shd mediates binding to phosphotyrosine residues. In order to con®rm that Shd is a functional SH2 protein, the SH2 domain was expressed as a GST fusion protein and assessed for the ability to bind to tyrosine phosphorylated proteins from K562 cell lysates. As shown in Figure 4, the Shd SH2 domain bound to several phosphotyrosine containing proteins with molecular weights of 200, 170, 130, 100, 90, 78 and 72 kD. Together with the results of the yeast two hybrid screen, these ®ndings indicate that the Shd SH2 domain mediates binding to phosphotyrosine motifs. When the immunoblotting experiment was repeated using an anti-Abl antibody, no 210 kD Bcr ± Abl protein was detected. p140 c-ABL was also not detected, however, c-Abl is not phosphorylated in K562 cells (B Druker, unpublished data). Thus, although an interaction with Abl kinase occurred in the yeast two hybrid system, the GST-SH2 domain fusion protein did not bind p210 Bcr ± Abl in K562 cell lysates. These discrepant results may be due to dierences in the sensitivity of these two assays, a dierence in binding of Shd to full length Bcr ± Abl, expressed in K562 cells versus the kinase domain used in the yeast two hybrid screen, or to a dierence in tyrosine phosphorylation sites on these two constructs. We have expressed a full length Bcr ± Abl in the yeast two hybrid system and retain binding of Shd to this construct (data not shown), suggesting that full length Bcr ± Abl can bind to Shd. Shd is phosphorylated by the Abl kinase Although the exact mechanism by which Abl kinase recognizes a speci®c tyrosine residue for phosphorylation is unclear, there is evidence that YXXP sequences are preferred by the Abl kinase in vitro (Songyang et al., 1995). The phosphorylation of CrkL and CrkII by Abl kinase is consistent with this model. CrkL and CrkII bind to the proline rich region in the C-terminus of Abl through their SH3 domains and can then be

* 34 c-ABL p210bcr-abl SHD

*

pTyr-blot Figure 4 Binding of the SH2 domain of Shd to tyrosinephosphorylated proteins. The SH2 domain of Shd was subcloned into pGEX-KG and expressed as a GST-fusion protein. The GST-Shd fusion protein was immobilized on glutathione sepharose and incubated with lysates of K562 cells. Bound proteins were separated by SDS ± PAGE and tyrosine-phosphorylated proteins were detected using a monoclonal anti-phosphoty rosine antibody. The relative molecular weights of the bound tyrosine-phosphorylated proteins are indicated on the right of the panel. The asterisks indicate non-speci®c binding of antibodies to GST whose migration was determined by Coomassie staining

WT – –

KD – +

WT – +

– KD +

– WT +

44 kD pTyr blot Figure 5 Phosphorylation of Shd by Abl kinases. A full length Shd cDNA was co-transfected with cDNAs for c-Abl (wild type ± WT), c-Abl kinase defective (KD), p210bcr-abl WT or p210bcr-abl KD into COS 7 cells using lipofectamine (GIBCO). Two days after transfection, cell lysates wre analysed for tyrosine phosphorylation of Shd by immunoblotting with a monoclonal anti-phosphotyrosine antibody. The expression of Shd in each of the transfected COS 7 cells was approximately the same as assessed by immunoblotting with Shd antisera (data not shown)


phosphorylated by the Abl kinase (Feller et al., 1994a, b). The tyrosine phosphorylation site identi®ed on Crk II (Y221) (Feller et al., 1994a) and the corresponding site of CrkL (Y207) both have a YXXP motif. Since Shd contains ®ve YXXP motifs and interacts directly with Abl kinase in the yeast two hybrid system, we investigated whether Shd is phosphorylated by Abl kinase in vivo. Shd and c-Abl kinase were co-expressed in COS-7 cells, followed by immunoblotting with an anti-phosphotyrosine antibody (Figure 5). Although cAbl kinase activity is negatively regulated by a putative inhibitor in mammalian cells (Pendergast et al., 1991), overexpression of c-Abl in COS-7 cells can overcome the inhibition (Goga et al., 1995). Analysis of whole cell lysate phosphotyrosine immunoblots of COS-7 cells overexpressing c-Abl showed that multiple proteins including c-Abl were tyrosine phosphorylated, thus con®rming the kinase activity of c-Abl in these cells (data not shown). p210Bcr-Abl, a constitutively active form of Abl kinase, also showed a high level of tyrosine kinase activity when expressed in COS-7 cells (data not shown). Phosphotyrosine immunoblots of COS-7 cells co-expressing Shd and either c-Abl or p210Bcr-Abl demonstrated phosphorylation of Shd (Figure 5). In contrast, when Shd was co-expressed with kinase-defective mutants of c-Abl or p210Bcr-Abl, Shd was not phosphorylated. When COS-7 cells were co-transfected with Shd and the pcDNA3 expression vector alone, phosphorylation of Shd also did not occur (data not shown). These results suggest that Shd is tyrosine phosphorylated by active forms of Abl kinase in vivo. Overexpression of Shd In order to analyse the eect of expression of the Shd protein, the Shd cDNA was transfected into several cell lines. A protein with an approximate molecular weight of 45 kD was detected by immunoblotting PC12/Shd, A31/ Shd and 32D/Shd cell lysates with an antibody raised against a synthetic oligopeptide corresponding to the carboxy-terminal sequence of Shd (data not shown). The apparent MW on SDS-polyacrylamide electrophoresis is consistent with the estimated MW of 41 kD based on the amino acid sequence of Shd. Despite our ability to achieve high levels of Shd expression, we were unable to identify a corresponding phenotype. Discussion We identi®ed two novel SH2 domain-containing proteins that interact directly with c-Abl in the yeast two hybrid system. We designated these proteins Shd and She, since both cDNA sequences encode SH2 domains with signi®cant homology to the recently cloned SH2 protein, Shb. Using the partial Shd sequence as a probe, we cloned the complete mouse Shd cDNA. Comparison of the amino acid sequences of Shd and Shb indicates they share several structural motifs. Both sequences contain an N-terminal prolinerich region and a single C-terminal SH2 domain. These results suggest that Shb, Shd and She form a family of SH2 adapter proteins similar to the Shc family consisting of Shc A, Shc B and Shc C (O'Bryan et al., 1996). Both protein families contain a single SH2

domain localized to the C terminus. Shc proteins have a central proline-rich region with a conserved tyrosine phosphorylation site (Salcini et al., 1994). However, Shc proteins also contain an N-terminal phosphotyrosine binding domain (PTB), a second phosphotyrosine recognition sequence with a distinctly dierent binding speci®city (Blaikie et al., 1994; Kavanaugh and Williams, 1994). The amino acid sequences of Shb and Shd do not contain an obvious PTB domain. The Shc and Shb protein families also share similar patterns of tissue expression. Shc A is widely expressed except for extremely low levels in the brain (Nakamura et al., 1996; O'Bryan et al., 1996; Pelicci et al., 1992). Shc C is expressed primarily in brain and other tissues of neural origin (Nakamura et al., 1996; O'Bryan et al., 1996). Similarly, Shb is expressed ubiquitously, while Shd expression is restricted to brain. The molecular interactions of the Shb SH2 and proline-rich motifs have been characterized, and are consistent with a physiological role as an adapter molecule in signal transduction. The SH2 domain of Shb was shown to bind to autophosphorylated b-PGDF receptors as well as a series of tyrosine phosphorylated peptides encompassing the dierent tyrosines in the PDGF b receptor (Karlsson et al., 1995; Welsh et al., 1994). The observed high degree of homology with Shb suggested that the SH2 domain of Shd might also mediate binding to phosphotyrosine motifs. As predicted, the GST-Shd SH2 fusion protein bound to multiple tyrosine phosphorylated proteins in a Bcr ± Abl expressing cell line. Overexpression of Shc A transforms rodent fibroblasts, and induces neurite outgrowth in the rat pheochromocytoma PC12 cell line (Pelicci et al., 1992; Rozakis-Adcock et al., 1992). Both eects of Shc A are dependent on the Ras pathway (RozakisAdcock et al., 1992; Salcini et al., 1994). We expressed Shd in a variety of cell lines in order to determine whether Shd had similar biological eects. Overexpression of Shd did not induce transformation of A31 ®broblasts in a focus-forming assay. PC12 cells dierentiate in response to nerve growth factor (NGF) through activation of the Ras pathway (Thomas et al., 1992; Wood et al., 1992). Transfection of Shd into the PC12 neuronal cell line neither induced nor inhibited dierentiation in response to NGF. The Shd cDNA was also transfected into the IL-3 dependent murine 32D cell line and assayed for IL-3 dependent growth. Although several oncogenes, have been shown to induce growth factor independence in this myeloid cell line, overexpression of Shd did not induce IL-3 independent growth. Thus, despite stable expression in several cell lines, we have been unable to determine the biological function of Shd. Shc A participates in intracellular signaling pathways for growth and dierentiation and is inducibly phosphorylated in cells stimulated with a variety of growth factors or transformed by oncogenic tyrosine kinases (Burns et al., 1993; Cutler et al., 1993; Dilworth et al., 1994; Matsuguchi et al., 1994; McGlade et al., 1992). We therefore investigated whether Shd could be inducibly phosphorylated by a variety of stimuli. However, stimulation of the stable Shd-expressing 32D, PC12 and A31 cell lines with IL-3, NGF and fetal bovine serum, respectively, did not induce tyrosine phosphorylation on Shd. These results

1259


1260

suggest Shd is not a substrate of JAK kinases, which are activated in response to IL-3 signals (Ihle and Kerr, 1995; Taniguchi, 1995). They also suggest Shd is not a substrate for NGF receptor or PDGF and EGF receptor tyrosine kinases, which are activated by NGF and serum and in turn, phosphorylate several cytoplasmic target proteins in order to transmit their biological signals. We previously reported that CrkII and CrkL interact directly with Abl kinase in the yeast two hybrid system and that CrkL is highly phosphorylated by Abl kinase (Oda et al., 1994). Other groups have also reported that CrkII and CrkL bind and are phosphorylated by Abl kinase (Feller et al., 1994a). According to the current model, the N-terminal SH3 domain of CrkL binds to the proline-rich region of Abl, followed by tyrosine phosphorylation (Feller et al., 1994b). The CrkL phosphorylation site (Y207) is part of a YXXP sequence preferred by Abl kinases (Songyang et al., 1995). Our ®nding that Shd contains ®ve YXXP sequences and interacts directly with active Abl kinase in the yeast two hybrid system suggested that Shd was a substrate for Abl. Co-expression of Shd with either c-Abl or Bcr ± Abl in COS-7 cells resulted in tyrosine phosphorylation of Shd, con®rming our prediction. In contrast, Shd was not phosphorylated when co-expressed with the kinase defective forms of Abl or Bcr ± Abl. These results indicate that Shd is tyrosine phosphorylated by the kinase active forms of Abl and Bcr ± Abl. Although the tyrosine phosphorylation site(s) of Abl have not been mapped, one would predict that the Shd SH2 domain binds to a phosphotyrosine residue on Abl, followed by phosphorylation at one or more of the YXXP motifs on Shd. Although Shd is phosphorylated by Bcr ± Abl when both proteins are co-expressed in COS-7 cells, it is an unlikely physiological substrate for Bcr ± Abl, given their distinctly dierent tissue expression. The Bcr ± Abl tyrosine kinase is found in the leukemia cells of virtually all cases of CML and a subgroup of patients with ALL. Since expression of Shd is restricted to brain, it is not likely to be involved in Bcr ± Abl mediated transformation of cells. c-Abl is a ubiquitously expressed tyrosine kinase whose physiological function is largely unknown. Overexpression of c-Abl was shown to arrest growth of ®broblasts, suggesting it may function as a negative regulator of cell growth (Sawyers et al., 1994). c-Ablde®cient mice homozygous for deletions of the entire cAbl gene or the carboxy-terminal third of the protein have abnormal head and eye development, lymphopenia, and a high neonatal mortality (Schwartzberg et al., 1991; Tybulewicz et al., 1991). Flies de®cient in the Drosophila Abl homolog have abnormal neural development (Henkemeyer et al., 1987), suggesting a role for c-Abl in the CNS. Since Shd is phosphorylated by c-Abl and is primarily expressed in the brain, it is possible that Shd phosphorylation transmits an c-Abl signal in the CNS. Materials and methods Cell culture K562 cells, a human CML cell line expressing p210Bcr-Abl (Lozzio and Lozzio, 1975), were cultured in RPMI 1640 medium (Gibco) supplemented with 10% fetal bovine

serum, penicillin, and streptomycin. The 32Dc13 (Greenberger et al., 1983) cell line was obtained from Joel Greenberger, University of Massachusetts Medical Center, Worcester, MA. Cells were cultured as per the K562 cells, with the addition of 15% WEHI-3B conditioned medium as a source of IL-3. PC12 cells were cultured in DMEM supplemented with 5% fetal bovine serum, 5% horse serum, penicillin, and streptomycin. A31 mouse ®broblast cells, COS-7 and C-2 cells were cultured in DMEM supplemented with 10% fetal bovine serum, penicillin and streptomycin. Transfections A pGD Shd retrovirus was produced from C2 packaging cells transfected with pGD Shd by the calcium phosphate method as described (Druker and Roberts, 1991). PC12 and A31 cells were infected with this retrovirus or a control retrovirus without insert as described (Druker and Roberts, 1991). Infected cells were selected in media containing 0.5 mg/ml G418 (Gibco). 32D/Shd and 32D/Neo cells were produced by transfection of the pGD Shd and pGD retroviral vectors into parental 32D cells by electroporation under conditions of 260 V, 960 mF. Following electroporation, cells were grown in RPMI 1640 medium for 48 h and then cultured in the presence of 1 mg/ml of G418 (Gibco). COS-7 cells were transfected with pcDNA3 expression plasmids using Lipofectin (Goga et al., 1995). After 2 days, the cells were harvested for immunological analysis. Plasmids The LexA fusion vector pBTM116 (courtesy of S Hollenberg), which contains the yeast TRP1 gene as a selectable marker, was used to construct the bait plasmid (pBTM116Abl) for the yeast two hybrid screen as follows: The KpnI to BclI fragment of human c-Abl cDNA encoding the SH2, tyrosine kinase and a small portion of the C-terminal domain was subcloned into pUC9 digested with KpnI and BamHI. The EcoRI to SalI fragment of pUC9 Abl (KpnI ± BclI) was then inserted between the EcoRI and SalI sites downstream of the LexA coding sequence of pBTM116. The pBTM116Abl plasmid was digested with EcoRI, ®lled in with Klenow and religated to adjust the frame. A kinase defective pBTM116Abl vector containing a Lys 271 to Arg amino acid substitution (Oda et al., 1995) was similarly constructed. pGD Shd was produced by subcloning the BamHI fragment from pSHlox Shd (which contains the full length Shd cDNA) into the BclI site of the recombinant retroviral vector pGD. pGD (courtesy of G Daley) contains a neomycin phosphotransferase gene allowing for selection in G418 (Daley et al., 1990). pcDNA3 Shd was made by inserting the EcoRI to NotI fragment of pSHlox Shd into pcDNA3 (InVitrogen). The expression plasmids for wild type and kinase defective c-Abl and p210Bcr-Abl kinases were made by subcloning the corresponding cDNAs into the EcoRI site of pcDNA3. Yeast two hybrid screening The yeast two hybrid screening was performed as previously described (Vojtek et al., 1993), using a c-Abl fragment fused in frame to the LexA DNA binding domain (LexA-Abl). A mouse embryo cDNA library fused to the acidic activation domain of VP16 in the pVP16 fusion vector was a gift from Stan Hollenberg. An L40 yeast strain expressing the LexAAbl construct was transformed with the pVP16 cDNA library. Transformants were plated to synthetic (THULL) medium lacking tryptophan, histidine, uracil, leucine and lysine. Interacting clones were identi®ed by selection on plates lacking histidine and analysed for b-galactosidase


production by a ®lter assay (Vojtek et al., 1993). The His+, LacZ+ colonies were then `cured' of the LexA fusion plasmid by growth on synthetic medium lacking leucine (leu7, trp+). Plasmid loss was con®rmed by plating individual colonies to both leu7, trp+ and leu7, trp7 plates. Colonies which grew only on trp+ plates were usd for mating to yeast strains that had been transformed with one of the following pLexA plasmids: (1) the pBTM116 LexA fusion vector; (2) pLexA-Lamin; (3) pLexA-Abl; or (4) a pLexA-Abl mutant. The LexA-Abl mutant construct contains a kinase defective Abl (Oda et al., 1995). The Leu+Trp+ diploids were assayed for transactivation of the LacZ reporter gene by a b-galactosidase ®lter assay, as described (Vojtet et al., 1993). The library pVP16 plasmids were recovered from desired clones, ampli®ed in HB101 E. coli and sequenced by the dideoxy method, using oligonucleotides corresponding to the 5' or 3' ¯anking regions of the pVP16 cloning site. Cloning of Shd A 16 day mouse embryo cDNA library in lSHlox vector (Novagen) (approximately 16106 recombinants) was screened with a 32P-labeled NotI fragment of pVP16 Shd under high stringency conditions. The NotI fragment encodes a partial Shd sequence. Positive plaques were incubated with the host bacterial strain BM25.8 according to the manufacturer's protocol in order to excise the pSHlox Shd plasmid. Both strands of the Shd cDNA were sequenced using dideoxy sequencing (USB). Northern blot analysis Mouse multiple tissue Northern blots purchased from Clontech were prehybridized and hybridized in Rapid hybbuer (Amersham) at 658C with a 32P-labeled mouse antisense Shd RNA probe transcribed from pSHlox by SP6 RNA polymerase. The blots were washed twice for 30 min at room temperature in 0.3 M NaCl, 0.03 M Na Citrate, 0.1% SDS, and then for 1 h at 688C in 15 mM NaCl, 1.5 mM Na Citrate, 0.1% SDS before autoradiography. The Shd RNA probe was removed from the hybridized blot by boiling. The same blot was then used for Northern analysis of She using a 32Plabeled mouse anti-sense She RNA transcribed from pGEM3Zf(+)-She which was prepared by subcloning the BamHI ± EcoRI fragment from pVP16-She into the pGEM3Zf(+) vector. The blot was hybridized with the labeled She RNA probe under the same conditions used for Shd, followed by autoradiography. Antisera and immunoblotting Rabbit polyclonal antisera against Shd was generated using an amino acid sequence CTLAAKPERGQGDP corresponding to the C-terminus of Shd. The peptide was crosslinked to KLH as described (Harlow and Lane, 1988) and mixed with Freund's adjuvant (Gibco) for injection into rabbits. The K-12 Abl antibody was purchased from Santa Cruz Biotechnology. The anti-phosphotyrosine monoclonal antibody 4G10 was generated using KLHphosphotyrosine as the immunogen as previously described (Druker et al., 1989; Kanakura et al., 1990). 32D, PC12,

and COS-7 cells used for immunoblotting experiments were lysed in NP40 lysis buer (20 mM Tris, pH 8.0, 1 mM EDTA, 150 mM NaCl, 1% NP40, 10% glycerol) containing 10 mg/ml aprotinin, 1 mM Na3VO4 and 1 mM phenylmethylsulfonyl ¯uoride. Whole cell lysates were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS ± PAGE), followed by either transfer onto PVDF membranes (Immobilon-P) for antibody detection using enhanced chemiluminescence (Pierce, Rockford, IL), or transfer onto nitrocellulose membranes (Schleicher & Schuell) for development using the NBT/BCIP method. The blots were incubated for 4 h with the anti-Shd antibody, and for 16 h with the anti-phosphotyrosine antibody, followed by incubation with the secondary antibody and development. Glutathione-S-Transferase (GST) fusion constructs and binding assays pGEX-SH2/Shd, an expression plasmid for the GST-SH2/ Shd fusion protein was constructed by subcloning the sequence encoding the SH2 domain of Shd into the pGEXKG vector (Guan and Dixon, 1991). The Shd SH2 domain sequence was obtained by PCR using the primers 5'CCGGAATTCCAGCCGTTGCCCTGGAG-3' and 5'CCGGAATTCGCTCTGGCTGCTGGTGA-3'. The PCR product was digested with EcoRI and ligated into pGEXKG that had been digested with EcoRI. The GST-SH2/Shd construct was sequenced to verify that no mutations were introduced by the PCR reaction. GST-fusion constructs were expressed in DH5a E. coli and the fusion proteins isolated from bacterial lysates using glutathione sepharose beads as previously described (Heaney et al., 1997). For the binding assays, the immobilized GST-SH2/Shd fusion proteins were incubated with K562 cell lysates. The beads were then washed twice in both NP40 lysis buer and PBS. Proteins were separated by SDS ± PAGE and transferred onto PVDF membranes for immunoblot analysis using anti-phosphotyrosine or anti-Abl antibodies. Phosphorylation of Shd by Abl kinases COS-7 cells were co-transfected with full length Shd cDNA and one of the following Abl kinase cDNA sequences: (1) c-Abl WT; (2) c-Abl kD; (3) p210 Bcr ± Abl WT; or (4) p210 Bcr ± Abl, using the lipofectamine method (GIBCO). (WT-wild type; kD-kinase defective mutant) COS-7 cells were also transfected with Shd cDNA, c-Abl, p210 Bcr ± Abl or the pcDNA3 expression vector alone. After 2 days, the cells were lysed with NP40 lysis buer and the proteins separated by SDS ± PAGE. The electrophoresis gel was cut for immunoblotting with either phosphotyrosine antibodies or Shd antibodies.

Acknowledgements We are grateful to Stan Hollenberg for providing the yeast two hybrid plasmids and the mouse embryo cDNA library. This work was supported by NIH grant CA65823 (to BJD). BJD is a recipient of a Translational Resarch Award from the Leukemia Society of America.

References Baskaran R, Chiang GC and Wang JY. (1996). Mol. Cell. Biol., 16, 3361 ± 3369. Baskaran R, Dahmus ME and Wang JY. (1993). Proc. Natl. Acad. Sci. USA, 90, 11167 ± 11171.

Ben-Neriah Y, Daley GQ, Mes-Masson A-M, Witte ON and Baltimore D. (1986). Science, 233, 212 ± 214. Blaikie P, Immanuel D, Wu J, Li N, Yajnik V and Margolis B. (1994). J. Biol. Chem., 269, 32031 ± 32034.

1261


1262

Burns LA, Karnitz LM, Sutor SL and Abraham RT. (1993). J. Biol. Chem., 268, 17659 ± 17661. Cohen GB, Ren R and Baltimore D. (1995). Cell, 80, 237 ± 248. Cooper JA and Howell B. (1993). Cell, 73, 1051 ± 1054. Cutler RL, Liu L, Damen JE and Krystal G. (1993). J. Biol. Chem., 268, 21463 ± 21465. Daley GQ, McLaughlin J, Witte ON and Baltimore D. (1987). Science, 237, 532 ± 535. Daley GQ, Van Etten RA and Baltimore D. (1990). Science, 247, 824 ± 830. Davis RL, Konopka JB and Witte ON. (1985). Mol. Cell. Biol., 5, 204 ± 213. Dilworth SM, Brewster CEP, Jones MD, Lanfrancone L, Pelicci G and Pelicci PG. (1994). Nature, 367, 87 ± 90. Druker BJ, Mamon HJ and Roberts TM. (1989). N. Engl. J. Med., 321, 1383 ± 1391. Druker BJ and Roberts TM. (1991). Nucl. Acids Res., 19, 6855 ± 6861. Feller SM, Knudsen B and Hanafusa H. (1994a). EMBO J., 13, 2341 ± 2351. Feller SM, Ren R, Hanafusa H and Baltimore D. (1994b). TIBS, 19, 453 ± 458. Goga A, Liu X, Hambuch TM, Senechal K, Major E, Berk AJ, Witte ON and Sawyers CL. (1995). Oncogene, 11, 791 ± 799. Greenberger JS, Sakakeeny MA, Humphries RK, Eaves CJ and Eckner RJ. (1983). Proc. Natl. Acad. Sci. USA, 80, 2391 ± 2395. Groen J, Stephenson JR, Heisterkamp N, de Klein A, Bartram CR and Grosveld G. (1984). Cell, 36, 93 ± 99. Guan K and Dixon JE. (1991). Anal. Biochem., 192, 262 ± 267. Harlow E and Lane D. (1988). Antibodies: a laboratory manual. Cold Spring Harbor Laboratory Press: Cold Spring Harbor, NY. Heaney C, Kolibaba K, Bhat A, Oda T, Ohno S, Fanning S and Druker BJ. (1997). Blood, 89, 297 ± 306. Heisterkamp N, Stam K, Groen J, de Klein A and Grosveld G. (1985). Nature, 315, 758 ± 761. Henkemeyer MJ, Gertler FB, Goodman W and Homann FM. (1987). Cell, 51, 821 ± 828. Ihle JN and Kerr IM. (1995). Trends in Genetics, 11, 69 ± 74. Jackson P and Baltimore D. (1989). EMBO J., 8, 449 ± 456. Kanakura Y, Druker B, Cannistra SA, Furukawa Y, Torimoto Y and Grin JD. (1990). Blood, 76, 706 ± 715. Karlsson T, Songyang Z, Landgren E, Lavergne C, Di Fiore PP, Ana® M, Pawson T, Lantley LC, Claesson-Welsh L and Welsh M. (1995). Oncogene, 10, 1475 ± 1483. Kavanaugh WM and Williams LT. (1994). Science, 266, 1862 ± 1865. Kharbanda S, Ren RB, Pandey P, Shafman TD, Feller SM, Weichselbaum RR and Kufe DW. (1995). Nature, 376, 785 ± 788. Koch CA, Anderson D, Moran MF, Ellis C and Pawson T. (1991). Science, 252, 668 ± 674. Lewis JM, Baskaran R, Taagepera S, Schwartz MA and Wang JYJ. (1996). Proc. Natl. Acad. Sci. USA, 93, 15174 ± 17179. Lozzio CB and Lozzio BB. (1975). Blood, 45, 321 ± 334. Lugo TG, Pendergast AM, Muller AJ and Witte ON. (1990). Science, 247, 1079 ± 1082.

Matsuguchi T, Salgia R, Hallek M, Eder M, Druker B, Ernst T and Grin JD. (1994). J. Biol. Chem., 269, 5016 ± 5021. McGlade J, Cheng A, Pelicci G, Pelicci PG and Pawson T. (1992). Proc. Natl. Acad. Sci. USA, 89, 8869 ± 8873. McWhirter JR and Wang JYJ. (1991). Mol. Cell. Biol., 11, 1553 ± 1565. Nakamura T, Sanokawa R, Sasaki Y, Ayusawa D, Oishi M and Mori N. (1996). Oncogene, 13, 1111 ± 1121. O'Bryan JP, Songyang Z, Cantley L, Der CJ and Pawson T. (1996). Proc. Natl. Acad. Sci. USA, 93, 2729 ± 2734. Oda T, Heaney C, Hagopian J, Okuda K, Grin JD and Druker BJ. (1994). J. Biol. Chem., 269, 22925 ± 22928. Oda T, Tamura S, Matsuguchi T, Grin JD and Druker BJ. (1995). Leukemia, 9, 295 ± 301. Pelicci G, Lanfrancone L, Grignani F, McGlade J, Cavallo F, Forni G, Nicoletti I, Grignani F, Pawson T and Pelicci PG. (1992). Cell, 70, 93 ± 104. Pendergast AM, Muller AJ, Havlik MH, Clark R, McCormick F and Witte ON. (1991). Proc. Natl. Acad. Sci. USA, 88, 5927 ± 5931. Pendergast AM, Quilliam LA, Cripe LD, Bassing CH, Dai Z, Li N, Batzer A, Rabun KM, Der CJ, Schlessinger J and Gishizky ML. (1993). Cell, 75, 174 ± 185. Puil L, Liu J, Gish G, Mbamalu G, Botwell D, Pelicci PG, Arlinghaus R and Pawson T. (1994). EMBO J., 13, 764 ± 773. Ren R, Ye Z-S and Baltimore D. (1994). Genes & Devel., 8, 783 ± 795. Rozakis-Adcock M, McGlade J, Mbamalu G, Pelicci G, Daly R, Li W, Batzer A, Thomas S, Brugge J, Pelicci PG, Schlessinger J and Pawson T. (1992). Nature, 360, 689 ± 692. Salcini AE, McGlade J, Pelicci G, Nicoletti I, Pawson T and Pelicci PG. (1994). Oncogene, 9, 2827 ± 2836. Sawyers CL, McLaughlin J, Goga A, Havlik M and Witte O. (1994). Cell, 77, 121 ± 131. Schlessinger J and Ullrich A. (1992). Neuron, 9, 383 ± 391. Schwartzberg PL, Stall AM, Hardin JD, Bowdish KS, Humaran T, Boast S, Harbison ML, Robertson EJ and Go SP. (1991). Cell, 65, 1165 ± 1175. Songyang Z, Carraway KL III, Eck MJ, Harrison SC, Feldman RA, Mohammadi M, Schlessinger J, Hubbard SR, Smith DP, Eng C, Lorenzo MJ, Ponder BAJ, Mayer BJ and Cantley LC. (1995). Nature, 373, 536 ± 539. Taniguchi T. (1995). Science, 268, 251 ± 255. Thomas SM, DeMarco M, D'Arcangelo G, Halegoua S and Brugge JS. (1992). Cell, 68, 1031 ± 1040. Tybulewicz VL, Crawford CE, Jackson PK, Bronson RT and Mulligan RC. (1991). Cell, 65, 1153 ± 1163. Van Etten RA, Jackson P and Baltimore D. (1989). Cell, 58, 669 ± 678. Vojtek AB, Hollenberg SM and Cooper JA. (1993). Cell, 74, 205 ± 214. Welch PJ and Wang JY. (1993). Cell, 75, 779 ± 790. Welsh M, Mares J, Karlsson T, Lavergne C, BreÂant B and Claesson-Welsh L. (1994). Oncogene, 9, 19 ± 27. Wen ST, Jackson PK and Van Etten RA. (1996). EMBO J., 15, 1583 ± 1595. Wood KW, Sarnecki C, Roberts TM and Blenis J. (1992). Cell, 68, 1041 ± 1050.