HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 16, 13463-13472, April 19, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/16/13463    most recent M110981200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Himmel, K. L. Articles by Largaespada, D. A. Search for Related Content PubMed PubMed Citation Articles by Himmel, K. L. Articles by Largaespada, D. A. Social Bookmarking                      What's this?

Activation of Clg, a Novel Dbl Family Guanine Nucleotide Exchange Factor Gene, by Proviral Insertion at Evi24, a Common Integration Site in B Cell and Myeloid Leukemias^*

Karen L. Himmel^§, Feng Bi^¶, Haifa Shen, Nancy A. Jenkins, Neal G. Copeland, Yi Zheng^¶, and David A. Largaespada^**

From the  University of Minnesota Cancer Center, Institute of Human Genetics, and the Department of Genetics, Cell Biology, and Development, University of Minnesota, Minneapolis, Minnesota 55455, the ^¶ Department of Molecular Sciences, University of Tennessee Health Science Center, Memphis, Tennessee 38163, and the  Mouse Cancer Genetics Program, NCI-Frederick Cancer Research and Development Center, Frederick, Maryland 21702

Received for publication, November 15, 2001, and in revised form, January 17, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Retroviruses induce leukemia in inbred strains of mice by activating cellular proto-oncogenes and/or inactivating tumor suppressors. The proviral integration sites in these leukemias provide powerful genetic tags for disease gene identification. Here we show that Evi24, a common site of retroviral integration in AKXD B cell and BXH-2 myeloid leukemias, contains a novel Dbl family guanine nucleotide exchange factor gene. We have designated this gene Clg (common-site lymphoma/leukemia guanine nucleotide exchange factor). Proviral integrations on chromosome 7 at Evi24 are located 7.6-10.3 kb upstream of Clg and increased Clg expression 2-5-fold compared with leukemias lacking proviral integrations at Evi24. Clg contains Dbl/pleckstrin homology domains with substantial sequence homology to many Rho family activators, including the transforming Dbl and Dbs/Ost oncogenes. Nucleotide exchange assays indicated that Clg specifically activated nucleotide exchange on Cdc42, but not RhoA or Rac1, in vitro. NIH 3T3 transfection studies showed that overexpression of full-length and carboxyl-terminally truncated forms of Clg morphologically transformed NIH 3T3 cells. This study and studies showing that the human homolog of EVI24 is located in a region of 19q13 frequently amplified in B cell lymphomas and pancreatic and breast cancers implicate Clg and Cdc42 activation in mouse and human cancers.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Slow-transforming retroviruses cause cancer in experimental organisms in part by acting as insertional mutagens (1, 2). Proviruses can insert within tumor suppressor genes and inactivate them (3, 4). Alternatively, activation of proto-oncogenes occurs in many different ways due to proviral insertion, including promoter insertion, truncation of the open reading frame, mRNA stabilization, and enhancement of the transcriptional activity of the endogenous promoter by elements in the long terminal repeats (LTRs)¹ of the provirus. This transcriptional enhancement mechanism often occurs in the activation of proto-oncogenes in the BXH-2 and AKXD models of acute myeloid leukemia (AML) and B and T cell malignancies (5, 6).

BXH-2 is one of a series of recombinant inbred strains created by crosses of the C57BL/6J and C3H/HeJ strains. Unlike either parental strain, ~95% of BXH-2 strain mice will develop AML by 1 year of age (7, 8). AML induction is causally associated with chronic infection with a B ecotropic murine leukemia virus (MuLV), which is passed from mother to offspring (7). A large number of common integration sites, chromosomal sites that harbor an integrated provirus in multiple independent tumors, have been identified in BXH-2 strain AML (2, 5) and in B cell tumors, developing as a consequence of MuLV infection, in various AKXD recombinant inbred strains made from crosses between the AKR and DBA strains (5, 9). Many of these common insertion sites were discovered using a powerful PCR-based system for cloning proviral insertions near CpG islands (5). In two BXH-2 and four AKXD mice, a common integration site called Evi24 (ecotropic viral integration site 24) was mapped to the promoter region of the Zfp36/tristetraprolin gene. Evi24 thus appears to contain a novel oncogene that predisposes B cell and myeloid leukemias.

In this work, we demonstrate that a novel Dbl family gene downstream of Zfp36 appears to be the target of proviral insertions at Evi24. We have designated the gene Clg (for common-site lymphoma/leukemia GEF) and show that it encodes a transforming oncogene with exchange activity for Cdc42, but not RhoA or Rac1. DBL is the prototypical member of a class of genes that encode GEFs for the Rho subfamily of the Ras superfamily small GTPases (10). All members of this class of GEFs are characterized by the presence of a similar ~180-amino acid Dbl homology (DH) domain, which catalyzes the exchange of bound GDP for GTP on Rho subfamily members such as Rho, Rac, and Cdc42 (11, 12). In all Dbl-like GEFs described to date, the DH domain is located amino-terminal to a pleckstrin homology (PH) domain, which has been shown to bind to other proteins or lipids produced as a consequence of phosphatidylinositol 3-kinase (PI3-K) activity (13). Many of the Dbl family GEFs have been shown to be transforming oncogenes in fibroblasts, and this transforming activity has been shown to require GEF activity and signaling via one or more of the small GTPases in the Rho subfamily (10). These proteins, like Ras, act as binary molecular switches in signal transduction pathways and are turned on when bound to GTP, interacting with various downstream effector molecules. They are inactive in the GDP-bound form and require Dbl family GEFs to exchange bound GDP for cytosolic GTP, resulting in activation. The Rho, Rac, and Cdc42 activities have been shown to be intimately involved in cytoskeletal organization (14-16); cell cycle control (17); and transcription activation, endocytosis and exocytosis, and transformation of fibroblasts (18, 19). At least two DH domain genes, LARG and BCR, are partners in chromosomal translocations that cause human myeloid leukemia (20, 21). The discovery of Clg activation by proviral insertion in myeloid and B cell malignancies implicates Dbl family genes and possibly Cdc42 activity in the hematological malignancies as well. Indeed, the CLG gene is a candidate oncogene in human diffuse large cell lymphoma, as it maps to 19q13.1, a region of frequent amplification in these B cell cancers and in ovarian, breast, and pancreatic adenocarcinomas (22).

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Molecular Cloning of Clg Proviral Insertions-- Proviral insertions at Evi24 were discovered as part of a large-scale cloning effort that utilized a long template, inverse PCR method described in detail elsewhere (5). Briefly, 5 µg of genomic DNAs from individual BXH-2 AML and AKXD malignant clones were digested with SacII endonuclease (New England Biolabs Inc.) overnight. The enzyme was then inactivated by heating at 65 °C for 30 min, and the DNA fragments were ligated in 500-µl reactions using 25 units of T4 DNA ligase (Roche Molecular Biochemicals) at 4 °C overnight to produce circular provirus/cellular DNA templates for PCR amplification. The ligated material was precipitated in ethanol and resuspended in 20 µl of Tris/EDTA (pH 8.0). Primary PCR was performed using a 5'-end-biotinylated forward primer, which allowed purification of the PCR products on M-280 streptavidin Dynabeads^® (Dynal, Inc., Lake Success, NY) before subsequent secondary PCR. 3.5 µl of the precipitated ligated tumor DNA were used as template for primary PCR in a 50-µl reaction volume containing 20 nmol of each dNTP, 15 pmol of forward (biotinylated) primer, 45 pmol of reverse primer, 1× Expand Long Template^TM buffer 2, and 2.5 units of enzyme mixture in the Expand^TM Long Template PCR system (Roche Molecular Biochemicals). Amplification was performed with an Omnigene Hybaid thermocycler with the following program: 92 °C for 2 min; 10 cycles at 92 °C for 10 s, 63 °C for 30 s, and 68 °C for 10 min; and 20 cycles at 92 °C for 10 s, 63 °C for 30 s, and 68 °C for 10 min with 20 s of autoextension. Primary PCR products were then subjected to bead purification following the manufacturer's instructions (Dynal, Inc.). Briefly, for each sample, 200 µg of beads were washed and resuspended in 40 µl of binding and washing buffer before adding to 40 µl of primary PCR products. After incubation at room temperature for 15 min, the tubes were placed in a Magnetic Particle Concentrator^® for 1 min, and the solution was removed. The beads were then washed twice with 50 µl of binding and washing buffer before being resuspended in 20 µl of Expand Long Template^TM buffer 2. Secondary PCR was performed as described by Li et al. (5), except that the template DNA was bound to beads. 5% of the purified, bead-bound PCR products (1 µl of 20) and 15 pmol of each secondary primer were used in 50-µl PCRs carried out exactly as described for the primary PCRs. The secondary PCR products were separated on a 1% agarose gel, purified using the Geneclean II^® kit (BIO 101, Inc.), and directly cloned using the CloneAmp^® pAMP1 system (Invitrogen) according to the supplied protocol. The primers used in the primary PCRs were AKV5-For 5' biotinylated (5'-CGAAGTAGTGTTACAGAATCGTAGAGGC-3') and SacII-Rev1 (5'-GCAACTGACCATTACCCCCC-3'). The primers used in the secondary PCRs were U5-For (5'-CUACUACUACUAGAGTGATTGACTGCCCAGCC-3') and SacII-Rev2 (5'-CAUCAUCAUCAUGAAAGCCCGAGAGGTGGTGG-3'). The cloned PCR products were sequenced using a PRISM BigDye^TM cycle sequencing kit (PerkinElmer Life Sciences) on an ABI Model 373A DNA Sequencer (Applied Biosystems). SP6 and T7 sequencing primers were purchased from Invitrogen.

Cloning Full-length Clg and Clg Clone 1e cDNAs-- Using BLAST analysis at the NCBI Database,² it was discovered that Evi24 is located on a segment of mouse chromosome 7 that had been deposited in the GenBank^TM/EBI Data Bank (accession number AC002327). Genomic sequence ~20 kb on either side of Evi24 was analyzed for putative exons using BLAST against the mouse expressed sequence tag data base as well as by employing the GRAIL³ and Gene Finder⁴ algorithms. The full-length Clg used in the studies herein was generated in two segments by RT-PCR using RNA from BXH-2 tumor 134 (GenBank^TM/EBI accession number AF465238). However, multiple cDNA clones were generated and sequenced from both BXH-2 tumor 134 and normal bone marrow RNAs, and it was determined that there are no sequence differences between the two. The first segment of the full-length Clg cDNA was cloned as follows: the 5'-primer (5'-CAGGCAGCCACCACCAT-3') and 3'-primer (5'-GTTGCATCTGCTGGATACGC-3') were used to generate a 2353-bp PCR product that was cloned into pcDNA3.1/V5/His-TOPO (Invitrogen). The resultant plasmid was then digested with KpnI restriction endonuclease (New England Biolabs), and the 1902-bp fragment was cloned into the KpnI site of pMSCVneo (CLONTECH). This plasmid was then digested in a SmaI/HindIII double digest, and the resultant 1384-bp fragment was cloned into SmaI/HindIII-cut pCMVTag2A (Stratagene), which put the cDNA in-frame with an upstream FLAG epitope. The second segment of the full-length Clg cDNA was generated using the 5'-primer (5'-GCCGCAGGCAGTCTGAGCCAGCAA-3') and 3'-primer (5'-CTCCTGGGCTAGAAAGCTACAG-3'). The 2742-bp PCR product was cloned into pcDNA3.1/V5/His-TOPO. The resultant vector was digested with HindIII, and the 2615-bp fragment was cloned into the HindIII site of the pCMVTag2A vector containing the first segment of the cDNA (see above). After sequencing, it was determined that the full-length cDNA clone exactly matched the sequence of all exons from the genomic clone (GenBank^TM/EBI accession number AC002327). Neither 5'-rapid amplification of cDNA ends nor RT-PCR revealed the presence of any additional 5'-exons. Clg clone 1e was generated using the same 5'-primer as was used in generating the first segment of full-length Clg above (5'-CAGGCAGCCACCACCAT-3') and a different 3'-primer (5'-GGGTTCTGGGATATGTTTGCTC-3'). The 1208-bp PCR product was cloned into pcDNA3.1/V5/His-TOPO in-frame with the downstream V5/His₆ epitope.

Southern Blot Analysis-- Genomic DNAs from BXH-2 and AKXD frozen tissues were isolated, and 5 µg were digested overnight using KpnI endonuclease. The resultant DNA fragments were electrophoresed through a 1% agarose gel and transferred to Hybond^TM N⁺ membrane (Amersham Biosciences) before hybridization. Labeling of the SacII/BlpI fragment from the inverse PCR product in pAMP1 from tumor 134 was performed by random-primed [³²P]dCTP labeling (Roche Molecular Biochemicals) essentially as described previously (7).

Northern Blot Analysis-- Total RNA was isolated from tumor samples using the RNA STAT-60^TM isolation reagent (TEL-TEST, Inc., Friendswood, TX). Subsequent purification of poly(A)⁺ mRNA on oligo(dT)-cellulose columns was performed with an Amersham Biosciences mRNA purification kit. 3.0 µg of poly(A)⁺ tumor RNA were electrophoresed through a formaldehyde-containing 1.0% agarose gel and transferred to Hybond^TM N⁺ membrane before hybridization. A 541-bp Clg fragment probe from the coding region was generated by RT-PCR using 5'-primer C (5'-CAGGCAGCCACCACCAT-3') and 3'-primer C (5'-GGGCCTGACGCTCCTGT-3'). The PCR products were directly radiolabeled by random-primed [³²P]dCTP labeling and hybridized to the blot. Blots were then stripped and rehybridized with a GAPDH probe as a loading control.

Western Blot Analysis-- Transiently transfected HEK 293 cells (calcium phosphate-transfected using a CellPhect^TM transfection kit from Amersham Biosciences) were lysed by vortexing in radioimmune precipitation assay buffer (10 mM Tris-HCl (pH 8.0), 150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40, 0.5% deoxycholate, and 0.1% SDS) and protease inhibitors (2 mM phenylmethylsulfonyl, 2 µg/ml leupeptin, 2 µg/ml pepstatin, and 2.5 µg/ml aprotinin) at a concentration of 10 × 10⁶ cells/ml at 4 °C for 30 min. The lysates were clarified by centrifugation for 30 min at 14,000 rpm to remove insoluble material. The concentration of protein was determined using the Pierce Coomassie Plus^® protein assay. 60 µg of protein were loaded onto 8% SDS-polyacrylamide gels, electrophoresed, and transferred to nitrocellulose as described (3). The blots were hybridized with anti-FLAG (M2; Sigma) or anti-V5 (Invitrogen) antibody according to the manufacturers' instructions and visualized with an enhanced chemiluminescence kit (ECL, Amersham Biosciences).

Focus Formation Assays-- NIH 3T3 fibroblasts were calcium phosphate-transfected (CellPhect^TM transfection kit) with 10 µg of Clg expression constructs or empty vector and selected in Dulbecco's modified Eagle's medium (Invitrogen), 10% fetal bovine serum, 1% penicillin/streptomycin, and 0.4 mg/ml G418 for 14 days. Cells (0.5 × 10⁶) were then plated in triplicate in 6-cm plates and grown for 7 days before staining with 50% methanol and 0.01 g/ml methylene blue.

Luciferase Reporter Assays-- A Dual-Luciferase^® reporter assay system (Promega) and a Mercury^TM pathway profiling system (CLONTECH) were used to assess signal transduction pathways that could be influenced by Clg. Experiments were conducted as outlined in the manufacturers' protocols. Briefly, HEK 293 cells were plated at 0.5 × 10⁶ cells/well using six-well culture plates. 24 h later, 1.3 µg of reporter vector (CLONTECH) and 0.2 µg of Renilla luciferase reporter vector (Promega) to control for transfection efficiency and 0.7 µg of Clg vector or corresponding empty vector were combined in a total volume of 120 µl and calcium phosphate-transfected in triplicate using the CellPhect^TM transfection kit as recommended by the manufacturer. Transfected cells were grown in Dulbecco's modified Eagle's medium without added serum for 16 h before lysis and luciferase detection. Levels of protein in each lysate were determined using the Pierce Coomassie Plus^® protein assay, and separate luciferase readings were taken for both the reporter plasmid and the Renilla luciferase reporter plasmid to control for transfection efficiency of each sample. For analysis, luciferase levels for empty vectors were set equal to 1, and all sample readings were normalized to protein content and transfection efficiency.

Clg-Small GTPase Complex Formation Assay-- COS-7 cells were transfected with the V5/His₆-Clg 1e construct by the LipofectAMINE method (Invitrogen). 48 h post-transfection, complex formation between V5/His₆-tagged Clg 1e and GST-fused dominant-negative Cdc42 (N17Cdc42), Rac1 (N17Rac1), RhoA (N19RhoA), or Ha-Ras protein was carried out by incubation of the Clg 1e-expressing cell lysates with the GST fusion proteins (1 µg). The coprecipitated complexes with immobilized Ni²⁺-agarose beads were washed three times with ice-cold lysis buffer. The coprecipitates containing the V5/His₆-Clg 1e input and GST fusion proteins were probed with anti-V5 and anti-GST monoclonal antibodies, respectively, on Western blots visualized by chemiluminescence reagents (Amersham Biosciences).

In Vitro GDP/GTP Exchange Assay-- The time courses for [³H]GDP/GTP exchange of Rho family GTPases in the presence or absence of purified His₆-Clg 1e, HA₃-Dbl (where HA is hemagglutinin), or His₆-TrioN were determined as previously described using the nitrocellulose filtration method (23). The GEF reaction buffer contained [³H]GDP-loaded Rho proteins with 20 mM Tris-HCl (pH 7.6), 100 mM NaCl, 10 mM MgCl₂, 0.5 mM GTP, and 1 mM dithiothreitol supplemented with purified His₆-Clg 1e.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Previous experiments using a long template, inverse PCR cloning procedure identified two BXH-2 myeloid leukemias with proviral integration in the promoter region of the Zfp36/tristetraprolin gene (GenBank^TM/EBI accession numbers AC002327 and L42317) (5). This common viral integration site was named Evi24. More recently, four additional AKXD malignancies (three B cell lymphomas and one mixed T and B cell lymphoma) were identified that have proviral integrations at Evi24 (Fig. 1). All but one viral integration (A034) is oriented in the same transcriptional direction as the Zfp36 gene. Zfp36 encodes an RNA-binding protein that regulates the stability of mRNAs such as TNF- (tumor necrosis factor-) and GM-CSF (granulocyte-macrophage colony-stimulating factor) by binding to their AU-rich elements (24, 25). A Zfp36 knockout has been made and has symptoms characteristic of TNF- transgene overexpression, including myeloid hyperplasia (26).

View larger version (5K): [in this window] [in a new window]   Fig. 1.   Gene content near Evi24. In two independent BXH-2 strain AML (gray arrows) and four independent AKXD strain B cell (black arrows) lymphomas, proviruses integrated into a 2.6-kb region of chromosome 7 (GenBankTM/EBI accession number AC002327) upstream of the Zfp36 gene. Three additional putative genes also reside in this region. A novel Dbl family GEF gene, which was named Clg, lies 7.5-10 kb downstream of the proviral insertions. Two as yet uncharacterized genes, named UsgA and UsgB, lie upstream of the insertions. UsgB appears to be the mouse homolog of a putative human gene, PD2 (GenBankTM/EBI accession number CAC20564). Large arrows indicate direction of transcription.

Additional genomic sequence around the Zfp36 gene, which maps to mouse chromosome 7 (27), has been deposited in the GenBank^TM/EBI Data Bank (accession number AC002327). By comparing sequences within 20 kb on each side of the Zfp36 gene with the dbEST Database as well as using the GRAIL and Gene Finder algorithms (28, 29), three additional genes in the Evi24 region were identified. Two genes, designated UsgA (upstream gene A) and UsgB (upstream gene B), are located 6-17 kb upstream of Zfp36, whereas Clg (see below) is located 4 kb downstream of Zfp36 (Fig. 1). Analysis of the predicted nucleotide and amino acid sequences of UsgA and UsgB did not reveal any information about their potential functions. However, UsgB appears to be the mouse homolog of a human gene called PD2, which is deposited in the GenBank^TM/EBI Data Bank without a publication, but the annotation of which indicates that this gene is amplified and overexpressed in pancreatic cancer (accession number CAC20564). Analysis of putative exons by GRAIL and Gene Finder in the gene downstream of Zfp36 revealed the possible presence of DH and PH domains. Subsequent RT-PCR and an independent cDNA cloning project in human (GenBank^TM/EBI accession number AK024429) have confirmed that this gene encodes a Dbl family GEF. We thus named the gene Clg, for common-site lymphoma/leukemia GEF. We detected no other known protein motifs in Clg. Southern blot analysis of the six tumor DNAs containing Evi24 viral integrations using an Evi24-specific probe showed that two tumors (BXH-2 tumors 134 and 154) contained rearranged bands that were equal in intensity to the unrearranged wild-type Evi24 allele (Fig. 2A). Thus, these two tumors harbor clonally integrated proviruses at the Evi24 locus. The other four tumor DNAs did not show rearrangements at the Southern blot level and presumably harbor subclonal proviral insertions at Evi24 (Fig. 2A and data not shown).

View larger version (46K): [in this window] [in a new window]   Fig. 2.   Expression of Clg. A, rearrangement at Evi24. Southern blot analysis was performed on BXH-2 and AKXD tumor DNAs digested with KpnI endonuclease. The cloned inverse PCR product from tumor 134 was used as a probe. Due to a KpnI site in the LTR of the provirus, fragments 2.3 and 1.8 kb smaller than germ line KpnI fragments were detected in the tumor 134 and 154 (BXH-2 tumors) lanes, respectively, showing that the locus is rearranged in a high percentage of tumor 134 and 154 cells. Rearrangement was not detected in tumors 10928 and X090 (AKXD tumors). Arrows indicate rearranged bands. B, Northern blot analysis of Clg expression levels. 3.0 µg of BXH-2 tumor poly(A)+ RNAs were run on a formaldehyde-containing 1.0% agarose gel. A Clg probe was generated by RT-PCR using primers designed from coding sequences. The major transcript is ~5 kb in size. The blot was then stripped and rehybridized with a GAPDH probe as a loading control. Samples with Evi24 insertions are indicated with asterisks. C, tissue distribution of Clg mRNA. A CLONTECH mouse poly(A)+ RNA Master BlotTM was hybridized with a Clg cDNA BlpI fragment probe. Expression levels were highest in embryonic day (E) 7, 11, and 17 tissues as well as in pancreas, thymus, uterus, skeletal muscle, lung, and heart. In addition, the level of Clg expression in testis was higher than it appears because subsequent hybridization with a hypoxanthine-guanine phosphoribosyltransferase control cDNA probe revealed underloading of the testis RNA sample. skel. mus., skeletal muscle; smooth mus., smooth muscle; submax. gland, submaxillary gland; prost., prostate; epidid., epididymis.

To determine whether any of the nearby genes are affected by proviral integration at Evi24, the two tumors with clonally integrated proviruses were analyzed by Northern blotting for UsgA, UsgB, Zfp36, and Clg expression. Surprisingly, Zfp36 expression was not affected by proviral integration when normalized to GAPDH and compared with other tumors without proviral insertions at Evi24. Likewise, the expression of UsgA and UsgB was also unaffected by proviral integration (data not shown). Instead, both tumors showed increased Clg expression or an additional Clg transcript that was larger than normal (Fig. 2B). The nature of these larger transcripts is not currently understood. They may represent fusion transcripts between the provirus and the Clg gene. However, attempts to amplify such a fusion using Clg- and virus-specific primers were not successful (data not shown).

Northern blot analysis of a multiple-tissue mouse mRNA dot blot with a Clg probe revealed expression in thymus, skeletal muscle, lung, testis, uterus, pancreas, and heart and during embryogenesis (Fig. 2C). A combination of RT-PCR and rapid amplification of cDNA ends was used to obtain a cDNA clone for the Clg gene, which was deposited in the GenBank^TM/EBI Data Bank (accession number AF465238). Multiple Clg cDNA clones from normal bone marrow and BXH-2 AML tumor 134 were obtained and sequenced. No differences were observed between any of these cDNA clones and the sequences of the corresponding exons in the genomic DNA clone (GenBank^TM/EBI accession number AC002327). The 5'-end of this cDNA maps 4.2 kb downstream of the last exon of Zfp36. The Clg coding region contains 19 exons and covers 13 kb of genomic DNA (Fig. 3A). The putative Clg start site is in exon 3, and the predicted Clg open reading frame codes for a 1298-amino acid protein (GenBank^TM/EBI accession number AF465238) that terminates in exon 19 (Fig. 3, A and B). The human CLG protein (GenBank^TM/EBI accession number AC011500) is 70% identical to the mouse protein overall, with 94% identity in the DH/PH domains. The Clg DH and PH domains are located in the N-terminal region of the protein. Very similar DH domains were also found in hypothetical proteins from human (GenBank^TM/EBI accession numbers BAB15719, BAB15364, CAB54806, and BAA86523), Caenorhabditis elegans (accession number T21663), and Drosophila (accession number AAF57673). The Clg DH domain is more distantly related to those from Dbl, Vav2, and Dbs/Ost (Fig. 3, C and D).

View larger version (161K): [in this window] [in a new window]   Fig. 3.   Clg structure, predicted amino acid sequence, and comparison with other DH domain-containing proteins. A, predicted structure of Clg. Clg has 19 exons, encoding a transcript of 5.2 kb, starting 4.2 kb downstream of the end of Zfp36. The coding sequence is shown in black, and the noncoding sequence is shown in gray. The 5'-untranslated region is ~1400 nucleotides, and the 3'-untranslated region is ~200 nucleotides. The exons encoding amino acid sequences similar to DH/PH domain-containing proteins are shown. There are no other known protein motifs in Clg. B, predicted amino acid sequence of Clg. The Clg protein is 1298 amino acids in length, with a predicted molecular mass of 139.2 kDa. The DH domain (~180 amino acids) is shown in white lettering on a gray background, and the PH domain (~110 amino acids) is shown in black lettering on a gray background. C, alignment of the Clg DH domain with other DH domains. Vector NTI Align X software (InforMax, Inc., North Bethesda, MD) was used to apply the ClustalW algorithm to align DH domains from various proteins. Blocks of amino acid consensus sequence are indicated in black lettering on a gray background. Weakly homologous residues are indicated in white lettering on a gray background. Other proteins in the alignment are as follows: C. elegans (Ce) hypothetical protein F32F2.1 (GenBankTM/EBI accession number T21663), Ciona savignyi (Cs) PEM-2 (accession number BAA36290), Drosophila melanogaster (Dm) CG5503 (accession number AAF57673), mouse (Mus musculus (Mm)) Dbl (accession number BAB40664), mouse Vav2 (accession number AAC52761), and rat Dbs/Ost (accession number Q63406). D, phylogenetic tree of Clg DH and related domains. The mouse Clg DH domain is compared with various DH domains from other Dbl family members using Vector NTI Align X software to apply the Neighbor Joining algorithm (InforMax, Inc.). All entries in the alignment are mouse, except for rat Dbs/Ost and human (Homo sapiens (Hs)) BCR.

To determine whether Clg encodes a true guanine nucleotide exchange factor that can activate various signaling pathways and transform cells in culture, as might be expected for a putative oncogene, we generated Clg expression vectors. Full-length Clg was N-terminally tagged with a FLAG epitope and expressed from the CMV immediate-early enhancer/promoter. In addition, the tagged cDNA was cloned into the pMSCVneo retroviral vector (30), in which expression is driven by the viral LTR. Similarly, a carboxyl-terminally truncated form of Clg called 1e, which contains the DH/PH domains and 33 adjacent amino acids, was produced and tagged with a V5/His₆ epitope at the C terminus and cloned into CMV expression and MSCVneo plasmid vectors (Fig. 4A). The CMV-driven vectors were tested for the ability to drive expression of the tagged proteins by transient transfection into HEK 293 cells. Recombinant tagged proteins of the anticipated size were detected by Western blotting (Fig. 4B).

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Clg constructs. A, Full-length and C-terminally truncated versions of Clg were generated by RT-PCR. Both clones contain the DH and PH domains. Full-length Clg was tagged with a FLAG epitope at the N terminus, and 1e was tagged with a V5 epitope at the C terminus. Full-length Clg (without the FLAG tag) is predicted to be ~140 kDa, and 1e (without the V5 tag) is predicted to be ~48 kDa. B, expression of Clg. Lysates from transiently transfected HEK 293 cells were subjected to SDS-PAGE and blotted. ECL was used to detect the full-length and 1e versions of Clg using anti-FLAG antibody M2 and anti-V5 antibody, respectively. Positive controls (+) and empty vector-negative controls are indicated.

To investigate the function of the DH and PH domains of Clg, we transfected Clg 1e cDNA, encoding amino acid residues 1-397, encompassing the DH domain (residues 60-236) and the PH domain (residues 256-364), into COS-7 cells. To test whether the Clg 1e fragment can directly bind to Ras or Rho family GTP-binding proteins, we employed the complex formation assay to detect the association between the Clg 1e polypeptide in cell lysates and GST fusion proteins of Cdc42, Rac1, RhoA, and Ras. The dominant-negative form of the Rho proteins (N17Cdc42, N17Rac1, and N19RhoA), which bear a Thr-to-Asn mutation at the corresponding positions, were used in these assays because of their relative higher affinity for known Dbl family GEFs. The anti-V5 and anti-GST antibodies were employed to probe the Ni²⁺-agarose-bound protein coprecipitates. Among the panel of small GTP-binding proteins, GST-N17Cdc42, GST-N17Rac1, and GST-N19RhoA were detected in the precipitates of V5/His₆-Clg 1e, whereas the GST control alone or GST-Ras showed no detectable affinity (Fig. 5A), suggesting that the Clg polypeptide containing the DH and PH domains is capable of specifically associating with Rho proteins in vitro.

View larger version (27K): [in this window] [in a new window]   Fig. 5.   Guanine nucleotide exchange activity of Clg. A, binding of the V5/His6-tagged Clg DH/PH domain module to Ras and the members of the Rho family of small GTP-binding proteins. 1 ml of COS-7 cell lysates containing V5/His6-tagged Clg 1e were incubated with 1 µg of GST-fused dominant-negative N17Cdc42, N17Rac1, or N19RhoA or with 1 µg of GST control alone or GST-Ras. The Ni2+-agarose precipitates were washed thoroughly, resolved by 10% SDS-PAGE, and probed by anti-GST immunoblotting (upper panel). V5/His6-Clg 1e in the coprecipitates were detected by anti-V5 antibody (lower panel). B, the Clg 1e polypeptide stimulates the guanine nucleotide exchange activity of Cdc42. Time courses are shown of the [3H]GDP dissociation from Cdc42, Rac1, or RhoA stimulated by purified V5/His6-Clg 1e. , buffer control; , Clg 1e; , purified onco-Dbl (Cdc42 and RhoA) or TrioN (Rac1). Data are representative of three independent assays.

To further determine whether the Clg polypeptide could act as a GEF toward the Rho GTPases, we expressed the V5/His₆-Clg 1e fusion fragment product in COS-7 cells and purified it by Ni²⁺-agarose affinity chromatography. The ability of purified His₆-Clg 1e to stimulate guanine nucleotide exchange of Cdc42, Rac1, and RhoA was examined in a [³H]GDP dissociation assay. As shown in Fig. 5B, isolated Clg 1e could efficiently promote the exchange of [³H]GDP for GTP only for Cdc42, very weakly for Rac1, and not at all for RhoA, indicating that Clg may function as a Cdc42-specific exchange factor.

To discover signal transduction pathways that can be activated by Clg and that are required for its ability to function as an oncoprotein, we performed transient transfection experiments utilizing Clg vectors and luciferase reporter plasmids containing promoters sensitive to various different signaling pathways. These experiments were performed in the HEK 293 cell line. Reporter vectors responsive to the p53, Rb, GAS (-interferon activation sequence), STAT3 (signal transducer and activator of transcription-3), SRE (serum-response element), AP1 (activator protein 1), NFB (nuclear factor of B cells), NFAT (nuclear factor of activated T cells), Myc, E2F, CRE (cAMP-response element), and ISRE (interferon-stimulated response element) pathways (Promega) were used in these experiments. Signal values were normalized to total protein in the lysate and to the level of expression of a Renilla luciferase gene included in the transfection mixture to control for differences in cell number and transfection efficiency in each sample (Fig. 6). The level of reporter activation was compared between triplicate samples of empty vector and full-length Clg or Clg 1e for each reporter. Transfection of the Clg vectors significantly up-regulated expression from the promoter containing the SRE. The truncated Clg 1e vector also more modestly up-regulated the Myc-, E2F-, and AP1-responsive promoters. These results suggest that Clg can signal through small GTPase proteins, which in turn activate the MAPK and/or JNK pathways.

View larger version (19K): [in this window] [in a new window]   Fig. 6.   Signaling pathway activation by Clg. A MercuryTM pathway profiling system and a Dual-Luciferase® reporter assay system were used to determine signaling pathways affected by overexpression of Clg. CMV-driven Clg clone 1e or empty vector was cotransfected into HEK 293 cells with each of 12 luciferase reporters (CLONTECH) and a transfection efficiency control reporter (Promega). CMV-driven full-length Clg or empty vector was cotransfected with six of the 12 luciferase reporters and the transfection efficiency control vector. The level of reporter activation was compared between triplicate samples of empty vector and full-length Clg or Clg 1e for each reporter. Transfection of the Clg vectors significantly up-regulated expression from the promoter containing the SRE. More modest up-regulation from the Myc, AP1, and E2F reporter vectors was seen with expression of Clg 1e. Background expression from cells transfected with empty vector control is set equal to 1. The -fold change over or under background is shown for each reporter. Error bars are ±1 S.E.

NIH 3T3 cells were also transfected with the different Clg constructs to determine whether they are capable of transforming NIH 3T3 cells in vitro. As a positive control, NIH 3T3 cells were also transfected with an activated form of the prototypical DH/PH domain-containing Dbl oncogene. As shown in Fig. 7, the various Clg constructs were able to transform NIH 3T3 cells in vitro. In contrast, very few colonies were detected when NIH 3T3 cells were transfected with an empty vector control.

View larger version (80K): [in this window] [in a new window]   Fig. 7.   Effect of Clg overexpression in fibroblasts. Full-length Clg and Clg 1e clones were expressed from CMV and LTR promoters in NIH 3T3 cells in focus formation assays. Cells transfected with full-length Clg, clone 1e, or empty vector were selected in G418 for 14 days and then plated at high density. Cells were stained 7 days after plating. Multiple foci are visible on plates with cells expressing full-length Clg and clone 1e. A, expression from pCMVTag2A and pcDNA3.1 is CMV-driven. B, expression from pMSCVneo is LTR-driven.

Full-length and truncated forms of Clg were also overexpressed in interleukin-3-dependent 32Dcl3 myeloblasts to determine whether Clg affects the growth of these cells. There was no significant difference in the growth of these cells in interleukin-3 or in the rate of apoptosis after interleukin-3 withdrawal between cells transfected with Clg and empty vectors (data not shown).

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Here we show that Evi24, a common viral integration site in BXH-2 and AKXD leukemias, contains four genes, but that only one of these genes (Clg) is up-regulated by proviral integration at Evi24. Clg is a new member of a large family of DH/PH domain-containing transforming oncogenes (10) that can morphologically transform NIH 3T3 cells in vitro. We also show that Clg can catalyze the exchange of GDP for GTP on Cdc42. This exchange activity likely accounts for the transforming activity of Clg because a truncated form of Clg containing only the DH/PH domains also transforms NIH 3T3 cells in vitro. Formal proof that Clg overexpression is causally associated with leukemogenesis will, however, require the production of Clg transgenic mice that have an increased frequency of leukemia induction or studies showing that CLG mutations also occur in human leukemias.

The location and orientation of proviral integrations at Evi24 suggest that Clg activation occurs through an enhancer mechanism. It is presently unclear why proviral integrations occur in this region rather than immediately upstream or downstream of Clg. No insertions were detected in the promoter region or 3' of Clg in a Southern blot screen of >200 BXH-2 tumors (data not shown). Perhaps this region affords a more favorable site for proviral integration than other sequences near Clg (i.e. its chromatin is more open). It is also possible that this region encodes a long-range negative regulatory element for Clg expression that is disrupted by proviral integration. However, overexpression of Zfp36 in NIH 3T3 cells did not lead to focus formation (data not shown).

The Dbl family contains many transforming oncogenes (10). Most of these genes require amino-terminal truncation (for example, see Ref. 31) or cotransfection with activated c-Raf kinase (for example, see Ref. 32) to show potent transforming effects in NIH 3T3 cells. In this report, however, we have shown that stable overexpression of Clg can transform NIH 3T3 cells without truncation. Thus, it seems entirely possible that simple overexpression of Clg, caused by proviral insertion at Evi24, contributes to leukemia development. Indeed, we have sequenced multiple Clg cDNA clones from BXH-2 tumor 134 and did not observe any nucleotide changes from the wild-type sequence. A combination of provirally mediated overexpression and gene point mutation is a plausible mode of oncogene activation in MuLV-induced cancer. However, this has never been observed in MuLV-induced models of leukemia.

The amino-terminal portion of Clg, upstream of the DH/PH domains, is 59 amino acids in length, and carboxyl-terminal truncation did not substantially increase its transforming activity. Nevertheless, transient Clg expression caused very few foci in primary focus forming assays compared with activated Dbl and did not result in appreciable colony formation in soft agar compared with activated Dbl (data not shown). It is likely that Clg transformation would be enhanced by other factors such as Ras signaling and could require additional oncogenic changes to cause myeloid leukemia or B cell lymphoma. Indeed, each cell in BXH-2 and AKXD tumors harbors three to four proviral insertions on average. We have cloned another proviral insertion from BXH-2 tumor 134, which is in the 5'-untranslated region of the gene encoding the short-chain dehydrogenase/reductase that reduces all-trans-retinal (data not shown). Perhaps changes in retinoid signaling, which is known to play a role in myeloid cell differentiation, can cooperate with Clg overexpression.

Several other genes involved in Ras signaling have also been implicated in BXH-2 and AKXD leukemias. These include guanine exchange factor genes such as Cal-Dag-GEFI and Cal-Dag-GEFII and GTPase-activating protein genes such as Nf1 (3, 5, 33, 34). These genes encode proteins that regulate various Ras-like subfamily members, including the true Ras proteins (N-, Ha-, and Ki-Ras) as well as TC21, Rap1, and R-Ras (35). Cal-Dag-GEFI and Cal-Dag-GEFII have Cdc25 catalytic domains, which are found in GEFs of this subfamily. In contrast, Clg has DH and PH domains, which are found in GEFs specific for the Rho subfamily of small GTPases (12). There is a great deal of evidence to support the interconnectivity between Ras and Rho/Rac/Cdc42 signaling. For instance, Rho, Rac, and Cdc42 activities are required for transformation of fibroblasts by activated forms of Ras (18, 19, 36, 37), and many DH domain-containing oncogenes cooperate with Ras signaling for transformation of rodent fibroblasts (10, 38). It seems possible that activation of the Ras pathway is critical for Clg-mediated transformation.

The Rho subfamily of GTPases, containing Rho/Rac/Cdc42-like proteins, has been implicated in many biological processes, including cellular transformation (16, 39, 40). Most data demonstrate the ability of activated forms of these proteins or their nucleotide exchange factors to cause transformation in experimental settings. Most often, these experiments have involved transformation of NIH 3T3 mouse fibroblasts. However, data exist showing that small GTPases of this class can actively participate in the transformation of other cell types and in human cancer. For example, human breast cancer cell lines and primary tumors show high levels of activated Rac3, which apparently drives their proliferation (41).

Various physiological effects of Cdc42 activation have been observed, and a number of effector molecules have been discovered (42). Among the possible effectors for Cdc42-mediated, Clg-induced cellular transformation are the coatomer complex (43) and PAK4 kinase (44). Aside from protection from apoptosis (45), the Cdc42 effector PAK4 can induce morphological changes, actin reorganization, and colony formation in soft agar (46). It is well established that Cdc42 activation can result in JNK and MAPK activation (47-49), and our own data on Clg are consistent with SRE and AP1 activation via Cdc42/JNK/MAPK pathways. Certainly, MAPK activation could contribute to tumor growth. Hyperactivated PI3-K signaling through an undetermined DH/PH domain-containing exchange factor(s) to Rac2, to a PAK family kinase, and finally to MAPK has been implicated in myeloid cell hyperproliferation after loss of the Nf1 tumor suppressor gene (50). Cdc42 activation has also been linked to resistance to Fas-mediated apoptosis (51), which is a feature of some human B cell lymphomas (52-54). It has also been shown that inactivation of the Fas pathway, together with overexpression of Bcl-2, can result in myeloid leukemia in mice (55). Thus, loss of susceptibility to Fas-mediated apoptosis could provide a common mechanism by which CLG expression would be selected for in other malignancies as well.

An analysis of the human genome indicates that there are a fairly large number of Dbl domain-containing genes, perhaps as many as 46 (56, 57). These genes seem to play a variety of biological roles in vivo. Mutations in the FGD1 gene cause a human genetic disease called Aarskog-Scott syndrome, which is characterized by impaired growth and facial, skeletal, and urogenital abnormalities (58). The Vav1 and Vav2 genes are required for aspects of B and T cell development (59). The Trio gene is required for proper secondary myogenesis and aspects of neuronal development (60). Clg is expressed at high levels in heart and lung, where it may have a role in cardiovascular development and/or function.

The human CLG gene maps to chromosome 19q13. This region is associated with recurrent translocations involving chromosome 11q23 (61), which is the site of the MLL gene. MLL is fused to a large number of different genes in human leukemia, including LARG, a Dbl domain-containing gene (21). LARG ties G-protein-coupled receptors to RhoA activation (62) and can cooperate with activated Raf to transform NIH 3T3 cells (32). It is unclear whether the GEF activity of LARG plays a role in transformation by the MLL-LARG fusion oncoprotein. Loss-of-function mutations or fusion to the MLL gene has also been observed for a Rho GTPase-activating protein gene, GRAF, in human leukemia with deletions of 5q or the t(5;11)(q31;q23) translocation, respectively (63). Our studies showing a possible role for Clg in mouse AML provide further evidence for a role for DH domain catalytic activity and Rho family activation in leukemogenesis.

The 19q13 region is also frequently amplified in human cancer, especially in diffuse large cell lymphoma, pancreatic adenocarcinoma, and breast cancer (61). These results, combined with our results demonstrating an apparent role for Clg in MuLV-induced leukemia/lymphoma development, suggest that CLG may be causally associated with other types of cancers as well. The CLG gene copy number and mRNA expression levels in pancreatic adenocarcinoma cells are currently under investigation. It seems possible that changes in adhesion, apoptosis sensitivity, or cell migration caused by inappropriate Cdc42 activation after CLG amplification could be selected for during adenocarcinoma progression. If this hypothesis is true, it will be important to determine how the GEF activity of Clg is regulated and why its overexpression can overcome such regulation.

	ACKNOWLEDGEMENTS

We thank Adam J. Dupuy and Scott J. Dylla for helpful discussions and technical assistance.

	FOOTNOTES

^* This work was supported in part by NCI Grant CA81051 from the National Institutes of Health.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBank^TM/EBI Data Bank with accession number(s) AF465238.

^§ Supported by NCI Cancer Biology Training Grant CA09138 from the National Institutes of Health.

^** To whom correspondence should be addressed. Tel.: 612-626-4979; Fax: 612-626-3941; E-mail: larga002@tc.umn.edu.

Published, JBC Papers in Press, February 11, 2002, DOI 10.1074/jbc.M110981200

² Available at www.ncbi.nlm.nih.gov/BLAST.

³ Available at compbio.ornl.gov/Grail-1.3/.

⁴ Available at dot.imgen.bcm.tmc.edu:9331/gene-finder/gf.html.

	ABBREVIATIONS

The abbreviations used are: LTRs, long terminal repeats; AML, acute myeloid leukemia; MuLV, murine leukemia virus; GEF, guanine nucleotide exchange factor; DH, Dbl homology; PH, pleckstrin homology; RT-PCR, reverse transcription-polymerase chain reaction; MSCV, murine stem cell virus; CMV, cytomegalovirus; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; HEK, human embryonic kidney; GST, glutathione S-transferase; SRE, serum response element; MAPK, mitogen-activated protein kinase; JNK, c-Jun N-terminal kinase; PAK, p21-activated kinase; PI3-K, phosphatidylinositol 3-kinase.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES