Involvement of NH2-terminal Sequences in the Negative Regulation of Vav Signaling and Transforming Activity*

Deletion of the NH2-terminal 65 amino acids of proto-Vav (to form onco-Vav) activates its transforming activity, suggesting that these sequences serve a negative regulatory role in Vav function. However, the precise role of these NH2-terminal sequences and whether additional NH2-terminal sequences are also involved in negative regulation have not been determined. Therefore, we generated additional NH2-terminal deletion mutants of proto-Vav that lack the NH2-terminal 127, 168, or 186 amino acids, and assessed their abilities to cause focus formation in NIH 3T3 cells and to activate different signaling pathways. Since Vav mutants lacking 168 or 186 NH2-terminal residues showed a several 100-fold greater focus forming activity than that seen with deletion of 65 residues, residues spanning 66 to 187 also contribute significantly to negative regulation of Vav transforming activity. The increase in Vav transforming activity correlated with the activation of the c-Jun, Elk-1, and NF-κB transcription factors, as well as increased transcription from the cyclin D1 promoter. Tyrosine 174 is a key site of phosphorylation by Lck in vitro and Lck-mediated phosphorylation has been shown to be essential for proto-Vav GEF function in vitro. However, we found that an NH2-terminal Vav deletion mutant lacking this tyrosine residue (ΔN-186 Vav) retained the ability to be phosphorylated by Lckin vivo and Lck still caused enhancement of ΔN-186 Vav signaling and transforming activity. Thus, Lck can stimulate Vav via a mechanism that does not involve Tyr174 or removal of NH2-terminal regulatory activity. Finally, we found that NH2-terminal deletion enhanced the degree of Vav association with the membrane-containing particulate fraction and that an isolated NH2-terminal fragment (residues 1–186) could impair ΔN-186 Vav signaling. Taken together, these observations suggest that the NH2 terminus may serve as a negative regulator of Vav by intramolecular interaction with COOH-terminal sequences to modulate efficient membrane association.

Vav is a member of the Dbl family of proteins (greater than 22 mammalian members) (reviewed in Refs. 1 and 2), and like a majority of Dbl family proteins was identified initially as transforming protein. All Dbl family proteins possess a tandem Dbl homology (DH) 1 domain followed by an invariant COOHterminal pleckstrin homology (PH) domain. The DH domain acts as a guanine nucleotide exchange factor (GEF) for specific members of the Rho family of small GTPases. For example, Vav is an activator of Rac1, RhoA, and Cdc42 (3,4). Rho family proteins are members of the Ras superfamily of GDP/GTPregulated proteins that function as binary switches (5). Dbl family proteins promote GDP/GTP exchange to cause formation of the active GTP-bound protein, whereas Rho GTPase activating proteins stimulate the intrinsic GTPase activity to convert these small GTPases to their inactive GDP-bound form. Rho guanine nucleotide dissociation inhibitors constitute a third class of regulators of GDP/GTP cycling.
Rho family proteins are regulators of a wide range of cellular activities that include the control of actin cytoskeletal organization (reviewed in Refs. 6 and 7). Whereas Rac1 promotes membrane ruffling, Cdc42 causes the formation of actin microspikes and the development of filopodia, and RhoA stimulates the assembly of actin stress fibers and focal adhesions. Rho family proteins are also regulators of gene expression (reviewed in Refs. 6 and 7). Rac1 and Cdc42 are activators of the c-Jun NH 2 -terminal kinases (8 -11), that in turn phosphorylate and activate the c-Jun, ATF-2, and Elk-1 nuclear transcription factors. Rac1, Cdc42, and RhoA also stimulate the activities of the serum response factor (SRF) (12) and NF-B transcription factors (13,14), that in turn regulate the expression of genes that control cell growth and apoptosis. Rho family proteins are also regulators of cell growth (reviewed in Refs. 6 and 7). Rac1, RhoA, and Cdc42 function is required for cell cycle progression and they are stimulators of transcription from the cyclin D1 promoter (10,11,15). Furthermore, the aberrant up-regulation of RhoA, Rac1, or Cdc42 can promote tumorigenic transformation, invasion, and metastasis. Specific Rho family proteins (RhoA, RhoB, RhoG, Rac1, Cdc42, and TC10) are also required for the transforming activity of oncogenic Ras protein and other oncoproteins. The transforming activities of Dbl family proteins are mediated presumably by constitutive activation of their specific Rho family substrates.
The invariant association of a PH domain with all DH domains suggests an interdependent relationship. Present evidence supports a dual role for the PH domain in the regulation of DH domain function (1,2). First, mutation of the PH domain abolishes the transforming activity of many Dbl family proteins and addition of a plasma membrane targeting sequence can restore this loss of PH function (16 -18). Thus, the PH domain may promote the association of Dbl family proteins with membranes, where their small GTPase substrates reside. Second, there is evidence that the PH domain, via an intramolecular interaction, can negatively or positively regulate DH domain function (3,19,20). For example, mutation of the Vav PH domain results in constitutive activation of Vav GEF function, suggesting that the PH domain negatively regulates Vav GEF function (3,21).
In addition to the DH and PH domains, Vav also contains a diverse array of other protein-protein or protein-lipid interaction domains that can potentially serve as negative or positive regulators of Vav function (reviewed in Ref. 22). The NH 2 terminus of Vav contains a leucine-rich, calponin homology domain (CH) (amino acids 3-115) (reviewed in Ref. 22), an acidic amino acid-rich domain (designated AD; 132-176), and three putative tyrosine phosphorylation sites (Tyr 142 , Tyr 160 , and Tyr 174 ). Tyr 174 has been shown to be a site of phosphorylation by the Src and Syk protein tyrosine kinases (23,24). Since NH 2 -terminal truncation of the first 65 residues of Vav (to form onco-Vav) unmasks Vav transforming activity, these sequences are believed to serve as negative regulators of Vav function. Finally, COOH-terminal sequences following the tandem DH/PH domains include a cysteine-rich domain, and a Src homology 2 (SH2) domain flanked by two Src homology 3 (SH3) domains (reviewed in Ref. 22). Mutagenesis studies showed that intact cysteine-rich domain, SH2, and SH3 domains are required for Vav transforming activity (25)(26)(27).
Recent studies have proposed a mechanism of Vav activation that involves phosphoinositide binding to the PH domain and phosphorylation of the NH 2 terminus. Like other PH domains (28,29), the Vav PH domain can bind phosphatidylinositol 4,5-biphosphate (PIP 2 ) and phosphatidylinositol 3,4,5-triphosphate (PIP 3 ) (3). Whereas PIP 2 association with proto-Vav does not promote its GEF function, PIP 3 -mediated dissociation of PIP 2 binding enhances its GEF function in vitro. PIP 3 binding also further promotes Lck-mediated tyrosine phosphorylation of proto-Vav, which in turn leads to further enhancement of GEF function. Lck-mediated phosphorylation of proto-Vav and onco-Vav, as well as the closely related Vav2, was found to be essential for GEF activity in vitro (3,22,23). Thus, a model has been proposed where phosphatidylinositol 3-phosphate kinasemediated conversion of PIP 2 to PIP 3 promotes Lck-mediated phosphorylation and activation of Vav. How phosphorylation, presumably at residue Tyr 174 , promotes Vav GEF function, is not known. Since this residue is positioned just upstream of the DH domain, phosphorylation of Tyr 174 may serve to overcome the negative regulatory action of the NH 2 -terminal sequences and promote the interaction of the DH domain with its GTPase substrates.
Deletion of the NH 2 -terminal 65 amino acids of proto-Vav unmasks its transforming activity, suggesting that these sequences serve a negative regulatory role in Vav function. However, the precise role of these NH 2 -terminal sequences and whether additional NH 2 -terminal sequences are also involved in negative regulation have not been determined. Therefore, we generated additional NH 2 -terminal deletion mutants of proto-Vav, that lack the NH 2 -terminal 127, 168, or 186 amino acids, and assessed their abilities to cause focus formation in NIH 3T3 cells and to activate different signaling pathways. Since Vav mutants lacking 168 or 186 NH 2 -terminal residues showed a 20 -30-fold greater focus forming activity than that seen with deletion of 127 residues, residues spanning 127-187 contribute significantly to negative regulation of Vav transforming activity. Furthermore, we found that an NH 2 -terminal Vav deletion mutant lacking Tyr 174 (⌬N-186) retained the ability to be phosphorylated by Lck in vivo and Lck could still enhance the signaling and transforming activity of this mutant. Thus, Lck can stimulate Vav by a mechanism that does not involve removal of NH 2 -terminal regulatory activity. Finally, we found that NH 2 -terminal deletion enhanced Vav membrane association and that an isolated NH 2 -terminal fragment of Vav could impair ⌬N-186 Vav function. Taken together, these observations suggest that the NH 2 terminus may serve as a negative regulator of Vav by intramolecular interaction with COOH-terminal sequences to modulate efficient membrane association.

EXPERIMENTAL PROCEDURES
Molecular Constructs-The pAX142 and pCTV3HA mammalian expression vectors have been described previously (16). To subclone the mouse proto-Vav and oncogenic Vav cDNA sequences into the pCTV3HA expression vector, the original pMEXneo-vav constructs PJC11 (encoding proto-Vav) and PJC12 (encoding onco-Vav) (provided by Mariano Barbacid) (26) were partially digested by EcoRI and ligated into the EcoRI site of pBS-SK ϩ (pBS-vav). The ClaI/HpaI fragment from pBS-vav was isolated and then fused in-frame with linkers (proto-Vav, ACGAATTCACGCGTAGTACTCCACCATGGAGCTCTGGCGACA; onco-Vav, ACGAATTCGTTAACCTCGCGAAGA) at the NH 2 terminus to a sequence encoding an initiating methionine followed by the hemagglutinin (HA) epitope tag contained in pCTV3.
The NH 2 -terminal truncation mutants of Vav were generated by polymerase chain reaction-mediated approaches using pCTV3HAproto-vav plasmid DNA as the template. 5Ј-Oligonucleotides with EcoRV restriction enzyme sites were generated at the chosen NH 2terminal truncation sites: for ⌬N-127 Vav, 5Ј-CAGATATCCCTTC-CCAACAGAGGACAGT; for ⌬N-168, 5Ј-CAGATATCTTTATGACTGCG-TGGAAAATGAG; and for ⌬N-186 Vav, 5Ј-CAGATATCCAAGATGACA-GAGTATGATAAG. These oligonucleotides were used to amplify 837, 744, or 660 base pair fragments, respectively, using a 3Ј-oligonucleotide (3Ј-CAGATATCCCTGGTAGAATGTGCCTCT) containing an EcoRV site corresponding to an EcoRV site within vav. Each fragment was ligated to the pCR 2.1 vector from the TA-cloning kit (Stratagene) and sequenced. The fragments were then digested with EcoRV and fused to the HpaI site of pCTV3HA. The remaining COOH-terminal sequences of the NH 2 -terminal truncated versions of Vav were added by digesting the truncation mutants and pCTV3-proto-vav with FseI/BsiWI and ligating the FseI/BsiWI ⌬N-Vav fragments to the FseI/BsiWI vector fragment of pCTV3HA proto-vav. cDNA sequences encoding proto-Vav⌬C were generated by polymerase chain reaction-mediated DNA amplification using a 5Ј-oligonucleotide sequence corresponding to a unique FseI restriction enzyme site (5Ј-CAGTTAACGGCCGGCCCAAGATTG-ACGGTGAG) and a 3Ј-oligonucleotide sequence (3Ј-CAGTTAACCTAG-AATTCCTTAGGCAGACCCAATTC) encoding a stop codon just prior to the start of the coding sequences for the NH 2 -terminal SH3 domain (at residue 616). This fragment was ligated to the HpaI site in pCTV3HA and then digested with FseI/BsiWI. This FseI/BsiWI fragment was then ligated to the appropriate FseI/BsiWI pCTV3HA-proto-vav fragment. The plasmid pCTV3HA-proto-vav⌬C and the pCTV3HA-⌬N-vav constructs were digested with MluI/BsiWI, and then subcloned into the corresponding sites within the pAX142 mammalian expression vector, where expression is under the control of the EF-1␣ promoter.
A cDNA sequence encoding proto-Vav that lacks the COOH-terminal 186 to 845 residues and fused to an NH 2 -terminal FLAG epitope tag was generated by polymerase chain reaction-mediated DNA amplification. A 5Ј-oligonucleotide primer (5Ј-CACCCGGGACCATGGAGGAC-TACAAGGACGACGATGACAAGGAGCTCTGGCGACAGTGC) containing a SmaI site, a FLAG epitope, and sequences corresponding to the ATG start site of proto-vav was used in conjunction with a 3Јoligonucleotide primer (3Ј-CACCCGGGTCATCATGGCGTAGGCAC-CGACTC) containing a SmaI site and sequences prior to the DH domain (up to residue 186) to generate an 186-amino acid fragment. The polymerase chain reaction fragment was digested with SmaI and ligated to the SmaI site of pAX142. The fragment was then sequenced to verify the accuracy of the amplified sequences. Mammalian expression vectors encoding constitutively activated human Rac1 (pAX142rac(61L)) and constitutively activated Lck (pLXSN-lck(Y505F)) have been described previously (30,31).
Cell Culture and Transformation Assays-NIH 3T3 mouse fibroblasts were cultured in Dulbecco's modified Eagle's medium supple-mented with 10% calf serum. 293T human kidney epithelial cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. DNA transfections were performed by calcium phosphate precipitation as described previously (32). For each assay, cognate empty vector was used as a control. For focus formation analyses, transfected NIH 3T3 cells were maintained in growth medium for 12 to 14 days. The cultures were then stained with crystal violet (0.5%) and the number of foci of transformed cells was then quantitated. To generate cells lines stably expressing each Vav protein, the transfected cultures were selected in growth medium supplemented with hygromycin (400 g/ml) and maintained under selection for 10 to 12 days. Multiple drug-resistant colonies (Ͼ100) were then pooled together to establish mass populations of cells stably expressing wild type or mutant Vav protein and used for the analyses described.
Transient Expression Reporter Gene Assays-Transient expression transcriptional assays were performed as described previously (33). Briefly, NIH 3T3 cells were transfected by calcium phosphate precipitation in 6-well dishes (5 ϫ 10 5 cells/well). The cells were serum-starved (0.5% calf serum) 14 to 15 h prior to lysing with luciferase lysis buffer (Amersham Pharmacia Biotech). Lysates were analyzed using Enhanced Chemiluminescent reagents and a Monolight 2010 luminometer (Analytical Luminescence). All the assays were performed at least three times.
The reporter plasmid constructs used for these assays have been described previously. To determine c-Jun and Elk-1 activity, we used a plasmid encoding the Gal4 DNA-binding domain fused to either the c-Jun NH 2 -terminal transactivation domain (Gal4-Jun) or the transcriptional activation domain of Elk-1 (Gal4-Elk-1), together with a luciferase gene reporter plasmid where expression is under the control of a minimal promoter containing five tandem Gal4 DNA-binding sites (5ϫGal4-Luc). To measure SRF and NF-B activation, we used reporter plasmids where luciferase gene expression was either under the control of a fos minimal promoter that contains tandem copies of a mutated serum response element (from the c-fos promoter), or tandem copies of the NF-B-binding site from the human immunodeficiency virus long terminal repeat, respectively. Cyclin D1 activation was done using the CD1-Luc reporter plasmid where the expression of the luciferase gene is under the control of the human cyclin D1 promoter (provided by Richard Pestell) (34).
Subcellular Fractionation Analyses-293T cells (2 ϫ 10 6 cells per 100-mm dish) were transiently transfected with 3 g of the pAX142 empty vector, or pAX142 constructs encoding proto-Vav, NH 2 -Vav, or ⌬N-186 Vav. Forty-eight h after transfection, the cells were serumstarved in Dulbecco's modified Eagle's medium supplemented with 0.1% fetal calf serum for 12-14 h, washed with ice-cold phosphatebuffered saline, and resuspended in cold TSA buffer (2 mM Tris, pH 8.0, 0.14 M NaCl) for 1 h. The lysates were homogenized in TSA buffer supplemented with 0.25 M sucrose, 1 mM EDTA, 100 g/ml phenylmethylsulfonyl fluoride, 25 g/ml leupeptin, and 1 mM NaVO 4 , and then centrifuged at 3,000 rpm to acquire total protein (200 l of supernatant). The remaining supernatant was then centrifuged at 100,000 ϫ g to separate into crude cytosolic (S100) and membrane (P100) fractions. The P100 fraction also contains cytoskeletal proteins. The protein concentration of the total, cytosolic, and membrane fractions was determined using a BCA protein assay kit (Pierce), with 40 g of protein for each fraction then resolved by SDS-polyacrylamide gel electrophoresis (SDS-PAGE), then transferred to Immobilon-P membranes (Millipore), and probed with anti-HA epitope antibody (Babco).
Immunoprecipitation and Western Blot Analyses-Expression of the HA epitope-tagged Vav proteins in stably transfected NIH 3T3 or in transiently transfected 293T cells was determined by standard immunoprecipitation and Western blotting techniques as described previously (35). Briefly, cells were lysed in PLC-LB (50 mM HEPES, pH 7.5, 150 mM NaCl, 10% glycerol, 1% Triton X-100, 1 mM EGTA, 1.5 mM MgCl 2 , 100 mM sodium fluoride) supplemented with 1 mM vanadate and various protease inhibitors (10 g/ml leupeptin and 10 g/ml aprotinin). Protein concentration was determined (Pierce) and then 500 g of the protein lysate was incubated with anti-HA epitope antibody (HA.11; Babco), and followed by the addition of 50 l of a 50% slurry (v/v in lysis buffer) of protein A/G-Sepharose beads (Santa Cruz Biotech). Expression of the FLAG epitope-tagged NH 2 -Vav protein was detected using the anti-FLAG epitope antibody (␣-FLAG M2; Sigma). All transfected cultures were serum-starved (0.1% fetal calf serum) for 12-14 h before lysing. For the NH 2 -terminal binding studies, 1 g of pAX142 NH 2 -Vav with cognate empty vector pAX142 (1 g) or pAX142 ⌬N-186 Vav (1 g) were transiently transfected into 293T cells. Protein samples were resolved by SDS-PAGE (7.5% for the HA epitope-tagged proteins and 12.5% for FLAG epitope-tagged Vav), then transferred to Immobilon-P membranes (Millipore), and probed with the appropriate primary antibodies. Membranes were then incubated with anti-mouse IgG secondary antibodies and protein was detected by chemiluminescence (Amersham Pharmacia Biotech).

NH 2 -terminal Truncation Mutants of Vav Exhibit Different
Levels of Transforming Activity in NIH 3T3 Cells-Since deletion of the NH 2 -terminal 65 residues of proto-Vav (designated onco-Vav) causes activation of its transforming potential, a negative regulatory role for these sequences has been proposed (36). However, whether the remaining NH 2 -terminal sequences that lie upstream of the DH domain (amino acids 66 -185) also contribute to the negative regulation of Vav transforming activity has not been determined. These upstream sequences contain a portion of the CH domain (37), an AD, and a phosphorylation site (Tyr 174 ) for Src or Syk family protein tyrosine kinase (4,22). To evaluate a possible role for these sequences in the regulation of Vav function, we generated a series of NH 2terminal truncation mutants of Vav, in which we deleted 127, 168, or 186 amino acids from the NH 2 terminus of proto-Vav (Fig. 1A). cDNA sequences encoding proto-Vav, onco-Vav, and each truncation mutant (designated ⌬N-Vav) were introduced into the pCTV3HA mammalian expression vector, which resulted in the addition of an NH 2 -terminal HA epitope tag to each Vav sequence that we used to monitor expression of the resulting HA epitope:Vav fusion proteins.
We first evaluated the expression of each mutant protein to determine if these NH 2 -terminal truncations altered protein stability. We transfected NIH 3T3 cells with the empty retroviral vector pCTV3HA, and pCTV3HA encoding proto-Vav and  the three NH 2 -terminal ⌬N-Vav mutant proteins to determine the levels of protein expression in stably transfected cells. Transfected cells were selected with hygromycin and multiple drug-resistant colonies (Ͼ100 colonies) were then pooled together and assessed for levels of Vav protein expression. We detected equivalent levels of protein expression for the truncation mutants in the stably selected NIH 3T3 cell lines (Fig. 1B). We also transiently transfected 293T cells using pAX142 expression vectors encoding the different Vav proteins and saw similar levels of expression (Fig. 1C). The comparable levels of protein seen for each Vav protein suggest that the different NH 2 -terminal truncations did not cause a significant alteration in protein stability.
Our preliminary transfection analyses showed that onco-Vav exhibited a weak focus forming activity (5-10 foci/dish) when transfected at 1 g of plasmid DNA/dish. Although this activity was 5-fold greater than that seen with proto-Vav (data not shown), we found that the three NH 2 -terminal deletion mutants exhibited very potent focus forming activities when transfected at these high plasmid DNA levels that were several 100-fold greater than onco-Vav (Fig. 2). Therefore, to provide a more accurate quantitation of their focus forming potencies, we performed our assays at plasmid DNA concentrations where no focus forming activity was seen for onco-Vav (30 ng of plasmid DNA/60-mm dish). Under these assay conditions, we observed a dramatic increase in focus forming activity as increasing amounts of the NH 2 -terminal sequences were removed. ⌬N-127 Vav displayed only a slight elevation of focus forming activity (Ͻ0.5 ϫ 10 3 foci/pmol) when compared with proto-Vav. The ⌬N-186 Vav mutant, which lacks all sequences upstream of the DH domain, showed the highest levels of focus forming activity (Ͼ3.5 ϫ 10 3 foci/pmol) and was 35-fold more potent than ⌬N-127 Vav. ⌬N-168 Vav possessed an intermediate activity that was 20-fold greater than that seen with ⌬N-127 Vav (Ͼ2 ϫ 10 3 foci/pmol). These results suggest that the sequences between residues 66 and 186 contain additional negative regulatory elements and that the putative NH 2 -terminal phosphorylation sites are dispensable for Vav focus forming activity.
Truncation of the NH 2 Terminus of Vav Activates Signaling Pathways That Are Similar to the Pathways Activated by RhoA, Rac1, and Cdc42-Constitutively activated Rho family members have been shown to activate a variety of transcription factors such as SRF, c-Jun, NF-B, and Elk-1, and to stimulate transcription from the cyclin D1 promoter (6, 7). Since Vav exhibits GEF activity in vitro for RhoA, Rac1, and Cdc42 (3, 4), we determined whether NH 2 -terminal truncation and activa-tion of Vav transforming activity correlated with activation of signaling pathways controlled by these Rho family proteins.
We compared the ability of wild type and ⌬N-Vav mutants to stimulate transcriptional activation of c-Jun, Elk-1, SRF, and NF-B and to stimulate expression of cyclin D1. For these analyses, NIH 3T3 cells were transiently transfected with pAX142 expression vectors encoding wild type and mutant Vav proteins. In general, we found that the transforming potential of a Vav mutant correlated directly with its ability to activate c-Jun, Elk-1, NF-B, and cyclin D1 (Fig. 3). Proto-Vav showed little, or no, activation of any of these activities. Onco-Vav exhibited limited, or no, enhanced ability to stimulate these signaling activities (data not shown). In contrast, the three NH 2 -truncation mutants caused greatly enhanced activities in all four assays done. SRF activation showed a direct correlation with their transforming activity, with ⌬N-168 and N-186 showing a 2-fold greater ability to activate SRF than that seen with ⌬N-127 Vav. Activation of Elk-1 also exhibited a direct correlation with transforming activity, with ⌬N-186 Vav causing the strongest activation (12-fold). In contrast, although ⌬N-168 Vav and ⌬N-186 Vav displayed much greater focus forming activity than ⌬N-127 Vav, all three showed comparable abilities to activate c-Jun, NF-B, and cyclin D1. Thus, these analyses suggest that the degree of activation of c-Jun, Elk-1, SRF, NF-B, or cyclin D1 may not account for the greatly increased transforming activity seen when 168 or 186 NH 2 -terminal sequences are deleted. However, since these assays were done under conditions of transient overexpression of each Vav mutant, it is possible that subtle changes in signaling activity are being obscured by the greater overexpression seen in transiently versus stably transfected cells. proto-Vav. If so, then ⌬N-186 Vav should no longer be sensitive to Lck-mediated enhancement of its activity. Therefore, we determined if ⌬N-186 Vav signaling and transformation can still be regulated by Lck.
We showed previously that co-expression of activated Lck(Y505F) caused synergistic enhancement of onco-Vav signaling and transforming activity (23). In this study, we found that co-expression of Lck(Y505F) was still able to enhance the activity of truncation mutants independent of the presence of Tyr 174 . Co-expression of Lck(Y505F) caused comparable 2-fold enhancement of the transforming activities of ⌬N-168 and ⌬N-186 Vav (Fig. 4A). Lck(Y505F) also cooperated with both ⌬N-168 and ⌬N-186 Vav and enhanced their ability to stimulate c-Jun and Elk-1 transcriptional activation (Fig. 4, B and C).
The ability of Lck(Y505F) to enhance the activity of ⌬N-186 Vav suggested that an additional site(s) of phosphorylation, other than Tyr 174 , may be present in Vav. To evaluate this possibility, we determined the ability of Lck to phosphorylate proto-Vav and all the ⌬N-Vav mutants in vivo. For these analyses, we transiently co-transfected 293T cells with expression vectors for Lck(Y505F) and proto-Vav or each of the NH 2terminal truncation mutants. Equal amounts of protein were then immunoprecipitated with anti-HA epitope antibody and both protein expression and tyrosine phosphorylation were determined by Western blot analyses using an anti-phosphoty-rosine (Tyr(P)) antibody or anti-HA epitope antibody, respectively (Fig. 5). Proto-Vav and all truncated proteins showed comparable levels of tyrosine phosphorylation when expressed in the absence of Lck(Y505F). Co-expression of Lck(Y505F) caused comparable increases in the level of tyrosine phosphorylation of all Vav proteins, including the ⌬N-186 Vav mutant that lacks Tyr 174 , suggesting that Tyr 174 may not be a key target for Lck in vivo. Thus, Lck can phosphorylate a tyrosine residue(s) other than those present in the NH 2 terminus of Vav. Furthermore, the ability of Lck to enhance the activity of the ⌬N-186 Vav truncation mutant suggests that Lck phosphorylation does not promote Vav activity by removing the negative actions of NH 2 -terminal sequences.
Truncation of the NH 2 Terminus Causes Increased Association of Vav with the Particulate Fraction-How the NH 2 terminus of Vav serves a negative regulatory role is not known. One possibility is that NH 2 -terminal truncation increases the intrinsic catalytic activity of the DH domain. However, it has been shown that proto-Vav and onco-Vav displayed comparable GEF activity in vitro (4). Thus, NH 2 -terminal truncation must increase Vav GEF function in vivo by a mechanism other than increasing the intrinsic catalytic function of the DH domain. A second possible mechanism may involve regulation of protein turnover. For example, NH 2 -terminal truncation and activation of Dbl is associated with increased protein stability (38,39). However, since proto-Vav and all NH 2 -terminal truncation mutants showed comparable levels of expression, in either transiently or stably transfected cells, this appears unlikely to be the case for Vav (Figs. 1, B and C, and 5). A third possibility, where NH 2 -terminal truncation allows greater phosphorylation of Vav, also does not appear to be involved, since the various forms of Vav did not show significant differences in their level of tyrosine phosphorylation, either in the absence or presence of Lck co-expression (Fig. 5). A fourth possibility involves NH 2 -terminal regulation of subcellular location. Since membrane association of various Dbl family proteins is essential for their transforming activities (reviewed in Refs. 1 and 2) we investigated whether the increase in focus forming activity seen with truncation of the Vav NH 2 terminus correlated with an increase in membrane association.
To determine whether the subcellular location of Vav is altered upon NH 2 -terminal deletion, we performed crude sub- cellular fractionation (100,000 ϫ g) with cell lysates of 293T cells transiently transfected with pAX142 expression vectors encoding HA epitope-tagged proto-Vav or the highly transforming ⌬N-186-Vav truncation mutant. The resulting cytosolic S100 and particulate P100 (membrane and cytoskeleton) fractions were then assessed by Western blot analyses using anti-HA antibody. We also evaluated the subcellular location of a COOH-terminal truncation mutant that lacks the SH3/SH2/ SH3 domains, to assess the contribution of these protein-protein interaction domains to Vav membrane association (designated proto-Vav⌬C). Proto-Vav and proto-Vav⌬C showed a similar distribution and were present predominantly in the S100 cytosolic fraction (ϳ60%), indicating that the COOH terminus is not involved in regulating Vav membrane or cytoskeletal association in non-stimulated cells. In contrast, ⌬N-186 Vav was associated predominantly (67%) with the P100 crude membrane fraction (Fig. 6). This increased association with the membrane-containing fraction suggests that the NH 2 terminus acts as a negative regulator of Vav function by decreasing its translocation to the plasma membrane, where its small GTPase substrates reside. Finally, immunofluorescence analyses did not show a drastic difference in the subcellular distribution of the two Vav proteins (data not shown).
An Isolated Vav NH 2 -terminal Fragment Inhibits ⌬N-Vav 186 Signaling Activity-One mechanism for the negative regulatory function of the NH 2 terminus of Vav may be to form an intramolecular association with COOH-terminal sequences, thereby preventing Vav membrane association and interaction with its GEF target substrates. To address this possibility, we determined if we could detect a stable association between the isolated NH 2 and COOH termini when co-expressed in vivo. For these analyses we generated a FLAG epitope-tagged NH 2terminal fragment that contained amino acids 1-185 (designated NH 2 -Vav) (Fig. 1A). We then transiently co-transfected 293T cells with expression vectors encoding FLAG epitope-tagged NH 2 -Vav and HA epitope-tagged ⌬N-186 Vav. Coimmunoprecipitation studies were performed using either the anti-HA or anti-FLAG epitope antibody. These analyses did not reveal a significant association of these two isolated fragments (data not shown).
The failure to detect stable association between the isolated NH 2 -Vav and ⌬N-186 Vav fragments did not exclude the possibility that such an interaction may still occur transiently to impair Vav function. To address this possibility we assayed the ability of NH 2 -Vav to inhibit ⌬N-186 Vav signaling. We found that co-expression of NH 2 -Vav was able to greatly decrease (ϳ70% inhibition) ⌬N-186 Vav-mediated activation of SRF (Fig. 7). The inhibitory activity of NH 2 -Vav was specific, since its co-expression did not inhibit the ability of activated Rac1(Q61L) to promote activation of SRF. These results support the possibility that the NH 2 terminus can form an intramolecular association with Vav COOH-terminal sequences and that this interaction serves as negative regulator of Vav function.

DISCUSSION
The removal of the NH 2 -terminal 65 amino acids of proto-Vav results in the formation of a constitutively activated and transforming protein, designated onco-Vav (36). Thus, these NH 2 -terminal sequences serve a negative regulatory role in Vav function. However, how these sequences regulate Vav activity, and whether additional NH 2 -terminal sequences also contribute to this regulation, have not been elucidated. The NH 2 -terminal sequences that remain in onco-Vav and are upstream of the DH domain include the remaining sequences of the leucine-rich CH domain, the acidic amino acid-rich AD, and a phosphorylation site for various tyrosine kinases such as Lck and Syk. Therefore, we removed additional NH 2 -terminal sequences to further delineate the involvement of the sequences upstream of the DH domain in the regulation of Vav function. We found that the removal of residues between 168 and 186 resulted in mutant Vav proteins with several 100-fold greater transforming activity than that seen with proto-Vav or onco-Vav. In our analyses we have detected only a weak focus forming activity with onco-Vav that is only 5-fold greater than that seen with proto-Vav. Thus, the loss of sequences between residues 66 and 186 caused a much more significant activation of Vav transforming activity than the loss of the first 65 residues. Removal of the putative Lck phosphorylation site at Tyr 174 did not abolish the ability of Lck to phosphorylate and stimulate Vav signaling and transforming activity. Therefore, an additional Lck phosphorylation site(s) may be present in Vav that contributes to Vav activation. The highly transforming ⌬N-186 Vav truncation mutant showed enhanced association with membranes, suggesting that the NH 2 -terminal sequences may regulate Vav membrane association. Finally, coexpression of an isolated NH 2 -terminal fragment (residues 1-185) caused inhibition of ⌬N-186 Vav signaling activity. Taken together, our observations support a model where mul- tiple residues within the region spanning the NH 2 -terminal 186 residues of Vav serve a negative regulatory role, in part, by regulating the subcellular location of Vav via an intramolecular interaction. NH 2 -terminal removal of sequences upstream of the DH domain represents the most common mechanism to unmask the transforming activation of Dbl family proteins. In addition to Vav, NH 2 -terminal deletion leads to constitutively activated mutants of Dbl (40), Ost (41), Tiam1 (42), Ect2 (43), and Net1 (44). Thus, although no significant sequence homology is seen among the NH 2 -terminal sequences of these Dbl family proteins (reviewed in Refs. 1 and 2), they share a common negative regulatory role. To further delineate the contribution of the NH 2 -terminal sequences of Vav in regulating Vav signaling and transforming activity, we deleted variable amounts of the NH 2 -terminal residues upstream of the DH domain that remain in onco-Vav. Our studies were also prompted, in part, because our assays detected only a weak activation of transforming activity with onco-Vav when compared with proto-Vav. We found that the sequential removal of sequences further enhanced onco-Vav transforming activity, suggesting that multiple NH 2 -terminal sequences contributed to negative regulation of Vav. In particular, the dramatic increase in Vav focus forming activity seen with the ⌬N-186 Vav or ⌬N-168 Vav mutants suggests that residues between amino acids 65 and 168 play a more critical role than the first 65 amino acids in this negative regulation. This possibility is also consistent with observations that NH 2 -terminal deletion of residues 1-65 of the highly related Vav2 protein did not unmask its transforming activity (45). Instead, it was found that NH 2 -terminal deletion of essentially all sequences upstream of the DH domain (residues 1-183) was required to create a transforming mutant of Vav2.
Presently, little is known regarding any specific biochemical function(s) of the NH 2 -teminal sequences that are upstream of the DH/PH domains of Vav. The Vav homology to the actinbinding domain of calponin (CH domain) suggested that this Vav sequence may promote association with actin (46). However, there is some suggestion that the CH domain of Vav does not play such a role (37,47). This region of Vav was found to associate with a novel protein, ENX-1, which is a putative transcriptional regulator of homeobox gene expression (48). Whether ENX-1 contributes to Vav function has not been determined. Essentially nothing is known regarding a specific function of the AD, where 22 of 45 amino acids are either aspartate or glutamate residues. Whether the CH or AD sequences interact with other proteins to promote the negative regulatory function of the NH 2 terminus will be important to determine.
We also evaluated the ability of Vav to stimulate various signaling pathways to determine if activation of a specific component correlated with Vav transforming potential. Vav has been shown to be a GEF for RhoA, Rac1, and Cdc42 in vitro (3,4). Therefore, we anticipated that Vav transforming potency should correlate with an increased ability to activate signaling pathways regulated by these Rho family proteins. Indeed, we found that Vav activation of the c-Jun, SRF, Elk-1, and NF-B transcription factors, as well as stimulation of the cyclin D1 promoter, increased with further removal of NH 2 -terminal sequences. However, the greatest increase for most of these signaling activities was seen with the ⌬N-127 Vav mutant and only limited (or no) further increase in stimulation was seen for the ⌬N-168 or ⌬N-186 Vav truncation mutants. One possible interpretation of these data is that the activation of these signaling components are not responsible for the very potent transforming activities of the ⌬N-168 and ⌬N-186 Vav mu-tants. Instead, Vav activation of other signaling components may provide a better correlation with Vav transforming activity. However, an alternative explanation for our results is that our signaling analyses were done by transient overexpression of the different proteins at levels that were severalfold higher than was seen in stably transfected NIH 3T3 cells. Thus, the transient high overexpression of the less potent ⌬N-127 mutant may have led to maximum stimulation of these signaling pathways and the overexpression of the ⌬N-168 or ⌬N-186 Vav mutants could not further stimulate these pathways.
Several mechanisms may be envisioned for how the NH 2 terminus of Vav can serve as a negative regulator of Vav function. First, the loss of the NH 2 terminus may promote protein stability and lead to an increase in the steady-state levels of protein expression. The NH 2 terminus of Dbl has been proposed to serve such a role (38,39). The p115 proto-Dbl protein showed a 5-fold higher turnover rate than did its NH 2terminal truncated and activated p66 counterpart. However, we found that NH 2 -terminal truncation did not cause any significant alteration in the level of expression, when compared with proto-Vav, when expressed either transiently (NIH 3T3 or 293T) or stably (NIH 3T3) in transfected cells. Second, the loss of the NH 2 terminus may increase the intrinsic GEF activity of the Vav DH domain. However, Crespo et al. (4) found that proto-Vav and onco-Vav showed comparable GEF activity in vitro. A similar observation was also made for Dbl, where the NH 2 -terminal truncated and activated Dbl protein did not show enhanced GEF function in vitro (49,50). Thus, NH 2terminal truncation is not likely to alter the intrinsic GEF activity of Vav. We have attempted to measure the GEF activity of our different Vav mutants by using anti-HA antibodyimmunoprecipitated proteins. However, we have not been able to reproducibly measure any significant GEF activity in these analyses. Thus, it remains a possibility that our ⌬N-186 Vav mutant does have increased intrinsic GEF activity when compared with onco-Vav. A third possibility may involve phosphorylation. Since Lck-mediated phosphorylation of Vav was found to be essential for proto-Vav GEF activity in vitro (3,23), NH 2 -terminal truncation may allow increased Vav phosphorylation in vivo. However, we found that proto-Vav and the NH 2terminal truncated mutants showed comparable degrees of tyrosine phosphorylation when expressed stably in NIH 3T3 cells.
Finally, a fourth mechanism may be that the NH 2 terminus regulates Vav association with membranes, where its GTPases targets reside. In support of this possibility, our fractionation analyses showed that whereas proto-Vav was found predominantly in the crude cytosolic fraction, the highly transforming ⌬N-186 Vav mutant was present predominantly in the crude P100 membrane and cytoskeletal protein fraction. Membrane translocation has been shown to be a positive regulator of the transforming activity of Lfc, Tiam1, and Dbs (16 -18). For each of these Dbl family proteins, deletion of their PH domain caused a loss of function, and the addition of a membranetargeting sequence restored transforming activity. Membrane targeting has also been established as an important positive regulator of the function of GEFs for Ras proteins as well (51,52). At present, we do not know to what degree this increased membrane association is responsible for the highly transforming activity of ⌬N-186 Vav. In addition to quantitative increases in membrane association, qualitative alterations in membrane association may also be important. It remains possible that NH 2 -terminal deletion may promote Vav association with specific membrane components that would not be obvious from our fractionation or immunofluorescence analyses. Finally, since optimal activation of Vav transforming activity required the removal of both the CH and AD sequences, each may serve a distinct negative regulatory function. Thus, the NH 2 terminus may also involve regulation of Vav properties distinct from regulation of membrane association. We are presently evaluating the contribution of membrane association to Vav function by generating mutants of Vav that are rendered constitutively associated with the plasma membrane.
Phosphorylation may be one mechanism to overcome the negative regulatory activity of the Vav NH 2 terminus in proto-Vav. Broek and colleagues (3) showed that phospholipid binding to the PH domain of Vav could modulate Lck-mediated phosphorylation of Vav, at Tyr 174 , in vitro. Whereas PIP 2 association with the PH domain inhibited Vav GEF activity in vitro, PIP 3 competed for PIP 2 binding and stimulated Vav GEF activity. PIP 3 binding also stimulated further Lck phosphorylation of proto-Vav, leading to additional enhancement of GEF activity. This and another study (4) showed that only phosphorylated Vav displayed GEF activity in vitro. Thus, a model was proposed whereby phosphatidylinositol 3-phosphate kinase conversion of PIP 2 to PIP 3 would lead to PIP 3 interaction with the Vav PH domain, thus potentiating Lck phosphorylation at Tyr 174 . How this phosphorylation event modulates Vav GEF activity is not known. The location of Tyr 174 just NH 2 terminus to the DH domain suggests the possibility that phosphorylation of this residue leads to an alleviation of the negative regulatory action of the NH 2 terminus. However, in our studies, we found that the ⌬N-186 Vav mutant, which lacks Tyr 174 , and proto-Vav were phosphorylated by Lck to a similar degree in vivo. Furthermore, co-expression of activated Lck(Y505F) caused the synergistic enhancement of the signaling and transforming activity of this truncation mutant. These results argue that Tyr 174 may not be the only phosphorylation site in proto-Vav and that an additional phosphorylation event(s) may promote Vav function even when the NH 2 terminus has been deleted. The DH domain of Vav does contain multiple tyrosines that may be phosphorylated by Lck to promote Vav GEF function. Alternatively, it is possible that NH 2 -terminal truncation may have unmasked tyrosine residues to serve as Lck phosphorylation sites, but they do not normally act as Lck regulatory sites in the full-length protein. However, we found that a full-length proto-Vav protein, that contains a Y174F mutation to prevent its phosphorylation, could be activated by Lck(Y505F) at levels comparable to those seen with the proto-Vav (data not shown). Thus, our observed Lck-mediated phosphorylation of ⌬N-186 Vav may be indicative of physiologically relevant sites of phosphorylation by Lck in full-length proto-Vav.
By analogy to protein kinases such as Raf-1, the negative regulatory action of the NH 2 terminus of Vav may be a consequence of its ability to form an intramolecular association with COOH-terminal sequences that then inhibits the catalytic function of the DH domain. Such a role has been proposed for the NH 2 terminus of Raf-1, where NH 2 -terminal deletion promotes the formation of constitutively activated and highly transforming kinase mutants (53,54). Morrison and colleagues (55) showed recently that the isolated NH 2 -terminal domain of Raf-1 could form a stable complex with the isolated and constitutively active COOH terminus of Raf-1 in vivo and that this NH 2 -terminal fragment could also inhibit its biological activity. We did observe that co-expression of the isolated NH 2 terminus of Vav selectively impaired ⌬N-186 Vav activation of SRF. However, we did not find that the isolated NH 2 terminus formed a stable complex with the activated ⌬N-186 Vav protein. Our failure to detect a stable association between the two fragments of Vav does not exclude the possibility that they can form an intramolecular association in the intact Vav protein.
Such intramolecular interactions are not expected to be too strong to allow it to be readily reversible. We are considering other approaches to detect a transient interaction between the NH 2 and COOH termini of Vav. Nevertheless, while other interpretations are possible for the inhibitory action of NH 2 -Vav, our results are consistent with the ability of the NH 2 terminus to associate with COOH terminus.
Although Vav expression is restricted to hematopoietic cells, our analyses of Vav function were performed in fibroblasts, where Vav transformation can be monitored. Thus, whether our observations concerning the role of the NH 2 terminus reliably predict its role in regulation of Vav function in hematopoietic cells remains to be determined. However, since the highly related and ubiquitously expressed Vav2 protein is also activated by NH 2 -terminal truncation, and exhibits the same GTPase specificity and signaling as Vav, 2 we suspect that our observations with Vav can be extended to Vav2.
In summary, our studies provide further characterization of the role of the NH 2 terminus in regulating Vav function. While our observations support a model whereby the NH 2 terminus regulates Vav membrane association via an intramolecular interaction, clearly more studies will be required to validate such a mechanism. The NH 2 -terminal sequences of other Dbl family proteins, including Vav2, Dbl, Ect2, Tiam1, and Net1, play a similar negative regulatory role. However, since there is a lack of primary sequence similarity among these NH 2 -terminal sequences, whether their mechanisms of negative regulation will be shared or distinct from those that we have seen with the NH 2 terminus of Vav remains to be determined. Finally, since either NH 2 -terminal truncation or membrane targeting of Raf-1 also converts this serine/threonine kinase into a potent transforming protein (55), there are apparent parallels with how Vav and Raf-1 activities are regulated. Despite intensive analyses, we still do not fully understand all the complexities of Raf-1 regulation. The regulation of Vav is likely to be equally complex.