Frameshift Mutation in the Dok1 Gene in Chronic Lymphocytic Leukemia

Oncogene (2004) 23, 2287–2297 & 2004 Nature Publishing Group All rights reserved 0950-9232/04 $25.00 www.nature.com/onc ORIGINAL PAPERS Frameshift mutation in the Dok1 gene in chronic lymphocytic leukemia

Sanghoon Lee1, Franc¸ois Roy1, Carlos M Galmarini2, Rosita Accardi1, Jocelyne Michelon1, Alexandra Viller1, Emeline Cros2, Charles Dumontet2 and Bakary S Sylla*,1

1International Agency for Research on Cancer, 150 Cours Albert-Thomas, Lyon 69008, France; and 2INSERM U 590 and Hospices Civils de Lyon, Lyon 69008, France

B-cell chronic lymphocytic leukemia (B-CLL) is a mulation of monoclonal CD5 þ , CD23 þ mature B malignant disease characterized by an accumulation of cells, with a high percentage of cells arrested in the G0/ monoclonal CD5 þ mature Bcells, with a high percentage G1 phase of the cell cycle. These cells have a severe of cells arrested in the G0/G1 phase of the cell cycle, and a defect in programed cell death due to high expression of particular resistance toward apoptosis-inducing agents. antiapoptotic proteins, like bcl-2 family proteins bcl-2, Dok1 (downstream of tyrosine kinases) is an abundant bcl-xL, bax and mcl-1 or survival factors such as Ras-GTPase-activating protein (Ras-GAP)-associated TRAF1, and TRAF2 (Schroeder and Dighiero, 1994; tyrosine kinase substrate, which negatively regulates cell Kitada et al., 1998; Caligaris-Cappio and Hamblin, proliferation, downregulates MAP kinase activation and 1999; Munzert et al., 2002). Signaling pathways such as promotes cell migration. The gene encoding Dok1 maps to signal transducer and activator of transcription 1 (STAT human chromosome 2p13, a region previously found to be 1) and (STAT 3), phosphatidylinositol-3 kinase (PI-3K) rearranged in B-CLL. We have screened the Dok1 gene or NF-kB have been found constitutively activated in B- for mutations from 46 individuals with B-CLL using CLL cells (Frank et al., 1997; Furman et al., 2000; heteroduplex analysis. A four-nucleotide GGCC deletion Munzert et al., 2002; Ringshausen et al., 2002). in the coding region was found in the leukemia cells from Activation of such pathways may contribute to the one patient. This mutation causes a frameshift leading to disease. protein truncation at the carboxyl-terminus, with the So far, no recurrent gene mutation has been found acquisition of a novel amino-acid sequence. In contrast to in B-CLL. Instead, numerous chromosomal abnor- the wild-type Dok1 protein, which has cytoplasmic/ malities involving chromosomes 6q21–q23, 11q22.3– membrane localization, the mutant Dok1 is a nuclear q23.1, 13q14, 17p13 or 2p13 have been identified, protein containing a functional bipartite nuclear localiza- some of which have clinical significance (Nelms et al., tion signal. Whereas overexpression of wild-type Dok1 1998; Wyatt et al., 2000). The p53 tumor-suppressor inhibited PDGF-induced MAP kinase activation, this gene and the ATM gene have been found mutated in inhibition was not observed with the mutant Dok1. B-CLL and p53 mutations are associated with Furthermore the mutant Dok1 forms heterodimers with resistance to chemotherapy (El Rouby et al., 1993; Dok1 wild type and the association can be enhanced by Wattel et al., 1994; Schaffner et al., 1999; Stankovic Lck-mediated tyrosine-phosphorylation. This is the first et al., 1999). example of a Dok1 mutation in B-CLL and the data Dok1 (downstream of tyrosine kinases), or p62Dok1, suggest that Dok1 might play a role in leukemogenesis. is a phosphorylated protein downstream of a wide range Oncogene (2004) 23, 2287–2297. doi:10.1038/sj.onc.1207385 of protein-tyrosine kinases (PTKs) (Carpino et al., 1997; Published online 19 January 2004 Yamanashi and Baltimore, 1997; Bhat et al., 1998). Dok1 and the related proteins Dok2, Dok3, Dok4, Keywords: Dok1; chronic lymphocytic leukemia; frame- Dok5 and the insulin receptor substrate (IRS) (Nelms shift mutation; NLS; Lck; tyrosine phosphorylation et al., 1998; White, 1998; Cong et al., 1999; Lemay et al., 2000; Di Cristofano et al., 2001; Grimm et al., 2001; Cai et al., 2003; Favre et al., 2003) contain a pleckstrin homology (PH) domain at the N-terminus, a central phosphotyrosine-binding (PTB) domain and a C-term- Introduction inal domain rich in proline and several tyrosine- phosphorylation sites (Carpino et al., 1997; Yamanashi B-cell chronic lymphocytic leukemia (B-CLL) is the and Baltimore, 1997). After tyrosine-phosphorylation, most common type of adult leukemia found in Western Dok1 can interact with other signaling molecules populations. The disease is characterized by an accu- containing SH2 domains, such as Ras-GTPase-activating protein (Ras-GAP), SHIP1, Nck, Csk and the X-linked lymphoproliferative syndrome (XLP) gene *Correspondence: BS Sylla; E-mail: [email protected] Received 23 May 2003; revised 19 November 2003; accepted 20 product SH2D1A (Neet and Hunter, 1995; Noguchi November 2003 et al., 1999; Sylla et al., 2000; Sattler et al., 2001). Dok1 mutation in B-CLL S Lee et al 2288 Dok1 downregulates several cell signaling processes: Results for example, it inhibits MAP kinase activity, downregulates cell proliferation, affects T- and B-cell signal- Frameshift mutation in the coding region of Dok1 ing and plays a role in activin-induced apoptosis in B-CLL (Tamir et al., 2000; Yamanashi et al., 2000; Di Cristofano et al., 2001; Martelli et al., 2001; Nemorin cDNAs isolated from lymphocytes of 11 healthy et al., 2001; Songyang et al., 2001; Wick et al., 2001; individuals and from tumor samples of 46 B-CLL Zhao et al., 2001; Kato et al., 2002; Yamakawa et al., patients were investigated for the presence of mutations 2002). Dok1 can also affect cell adhesion, spreading and in the Dok1 gene. The entire coding sequence was migration (Noguchi et al., 1999; Hosooka et al., 2001). amplified using the primers listed in Table 1 and the Tyrosine phosphorylation and PH domain-mediated PCR products were subjected to heteroduplex analysis. plasma membrane translocation are necessary for Dok1 No abnormality in migration was observed for any PCR function (Liang et al., 2002; Zhao et al., 2001). Dok1 fragment except for one obtained with F4/R4 primers can suppress cell transformation and leukemogenesis in (see Table 1). With this set of primers, a different mice (Di Cristofano et al., 2001; Wick et al., 2001). electrophoretic mobility was seen for B-CLL patient Dok1 maps to human chromosome 2p13 (Nelms et al., number 13, where the fragment migrated faster than in 1998), a region of chromosomal translocation associated the other samples (Figure 1a, lane 3 and data not with several forms of leukemia including B-CLL and B- shown). Sequence analysis of this fragment revealed a cell acute lymphoblastic leukemia (B-ALL) (Sonnier four-nucleotide GGCC deletion at position 683 in exon et al., 1983; Yoffe et al., 1990; Inaba et al., 1991; 5 (amino acid 221) of the coding sequence of Dok1 Richardson et al., 1992; Nelms et al., 1998). Thus, the (GenBank Dok1 cDNA Accession No. NM-001381) function of Dok1 is likely to be altered in these (Figure 1b). This mutation was confirmed by amplifying malignancies. and sequencing a longer cDNA fragment from CLL13 In the present study, we examined whether Dok1 spanning nucleotide 683 (data not shown). The GGCC is mutated in 46 B-CLL cases. We found a deletion was not detected in the nucleotide sequence of four-nucleotide GGCC deletion in the coding region cDNAs from normal individuals (Figure 1b, and data of the gene in leukemia cells from one patient. not shown). The mutation generates a frameshift and This mutation causes a frameshift leading to protein premature termination of the encoded Dok1 protein, truncation at the carboxyl-terminus, with acquisition of with the generation of a hybrid truncated Dok1 protein a novel amino-acid sequence. In contrast to the with 298 amino acids and a theoretical molecular mass wild-type Dok1 protein that has membrane/cytoplasmic of 33 kDa (Figure 1b, c). This predicted mutant protein, localization, the mutant Dok1 accumulates in the referred to as Dok1-CLL13, shares a 220 amino-acid nucleus and contains a functional nuclear localization sequence with wild-type Dok1, comprising the PH signal (NLS). Whereas the wild-type Dok1 down- domain at the N-terminus and part of the PTB domain regulates PDGF-induced MAP kinase activation, the in the middle. However, as a result of the frameshift, mutated Dok1 in CLL has no, or only a slightly Dok1-CLL13 contains a unique 78 amino-acid sequence increased effect on MAP kinase activity. These results at the C-terminal region that is not present in wild-type imply that Dok1 may play a role in the development of Dok1 (Figure 1c, d). Notably, the new sequence, which B-CLL. has no significant homology with any protein in the

Table 1 Primers and PCR conditions for ampliﬁcation of the Dok1 gene Primer Sequences Nucleotide positiona Exon Annealing T (1C) Expected size (bp)

F1A 50-CATGGACGGAGCAGTGATG-30 22–40 1, 2 60 361 R1C 50-AAAGGCGTTTCGGCACAG-30 362–379

F2 50-CCTTCCGCCTGGACACTG-30 285–302 2, 3 60 329 R3A 50-TGGCTCCAGTATCTGACTCT-30 593–613

F4 50-GCCTGCATGGCTCCTACG-30 528–545 4, 5 62 290 R4 50-GAGAACATCGTGCCCCTGT-30 798–817

F5 50-ACATCTTCCAGGCAGTTGAGA-30 741–761 5 67 317 R5 50-CTGCGCATGCTCATACAAGT-30 1038–1057

F6 50-GGAGAGGGAGTACAACGGAAG-30 1001–1021 5 60 370 R6 50-TGTACAGGGCTGAGTTGTG-30 1352–1370

F7 50-AGCCTTCCCTGAACCTGGTA-30 1306–1325 5 60 351 R7 50-CATCTGCCCCACTTCTTAGG-30 1637–1656

aDok1 cDNA sequence (GenBank Accession No. NM-001381)

Oncogene Dok1 mutation in B-CLL S Lee et al 2289 Cloning and expression of Dok1-CLL13 In order to obtain insight into the possible function of the mutant Dok1 protein, a cDNA fragment amplified from CLL13 using F1A/R5 primers (see Table 1) was cloned in a TA plasmid vector (see Materials and methods). After transformation, plasmid DNAs ex- tracted from several independent bacterial clones were purified and sequenced (data not shown). All the plasmids showed an identical GGCC deletion in the Dok1 sequence. Consistent with our previous observation (Figure 1b), this result implies that no wild-type sequence of Dok1 exists in the cDNA of CLL13. We could not determine whether the GGCC deletion was present in the germ-line, because no normal tissue samples from patient CLL13 were available. Efforts to isolate adherent cells (macrophages) by cultivating the leukemia sample of patient CLL13 for genetic analyses were unsuccessful. Next, the entire coding region of Dok1 from CLL13 was inserted into the expression vector pcDNA3-Flag in frame with the FLAG epitope to obtain pcDNA3-Flag-Dok1-CLL13 (see Materials and methods). This plasmid construct was transfected into 293T cells, and 2 days after transfection, cell lysates were subjected to Western blot analysis using the M5 anti-Flag antibody (Figure 2a). No protein was expressed from empty vector and wild-type Dok1 was expressed efficiently, as expected. The mutant Dok1- CLL13 was expressed in 293T cells as a protein with a molecular mass of about 35 kDa, close to the expected size of 33 kDa. Compared to wild-type Dok1, the level Figure 1 Frameshift mutation of the Dok1 gene in CLL. (a) of Dok1-CLL13 expression was relatively low and this cDNA samples from 11 normal and 46 CLL samples were was observed in several independent experiments (see amplified using the primer set F4/R4 (Table 1) and screened for detection of Dok1 mutation by heteroduplex analysis. A repre- below and data not shown). These results suggest that sentative gel with one normal sample (lane 1) and two CLL samples the CLL mutant protein could be stably expressed in (lanes 2 and 3) is shown. Arrow indicates the direction of gel cells, albeit at a lower level than the wild type. migration. Sample from CLL13 migrates faster than the control and CLL1. (b) The diagram illustrates the structure of Dok1 cDNA with the start codon from first amino acid (1 a.a.) and the stop Dok1-CLL13 protein accumulates in the nucleus codon after a.a. 481. E in the box diagram indicates exon. Below the diagram are the electrophoregrams from the sequencing of the In order to determine the subcellular localization of F4/R4 PCR product around the codon for a.a. 220. The coordinates of nucleotide sequence in the cDNA and the Dok1-CLL13, we first generated 293 cell lines stably corresponding amino-acid sequence are represented for normal expressing F-Dok1-wild type and F-Dok1-CLL13. A and CLL13. The bracket in normal indicates the two sets of GGCC cell line containing empty vector was also obtained. repeat. One set of GGCC (in box) was deleted in CLL13 and led to Several selected clones expressed similar levels of Dok1 a change of amino-acid sequence in CLL13 (underlined). Bracket in and Dok1-CLL13 proteins (Figure 2b and data not CLL13 indicates the remaining GGCC repeat. (c) Diagrams of Dok1 and Dok1-CLL13 proteins are represented. Dot box shown). Again the level of expression of Dok1-CLL13 indicates the PH domain and grey box indicates phosphotyro- was lower than that of the wild-type protein (Figure 2b, sine-binding sites (PTB). Black circles indicate tyrosine residues compare lanes 5–6 to lanes 3–4). No cell lines expressing that can be phosphorylated by tyrosine kinase, and P indicates Dok1-CLL13 at the level of wild-type Dok1 were proline residues in the C-terminal region. Dok1-CLL13 is truncated at its C-terminus as a result of a frameshift and the obtained. To determine the subcellular localization of hatched box indicates the new sequence of 78 amino acids. (d) The Dok1-CLL13, cells were fractionated into cytosolic/ 298 amino-acid sequence of Dok1-CLL13 is shown. The boxed membrane (C) and nuclear (N) fractions, and analysed sequence designates the PH domain, and the change of amino-acid by Western blot (Figure 2c). As previously reported sequences due to the frameshift is indicated in bold letters. Basic (Zhao et al., 2001), Dok1-wild type is predominantly a amino acid, arginine (R) and lysine (K) residues are underlined cytoplasmic protein with a small percentage found in the nucleus (Figure 2c, lanes 5 and 6). In contrast, Dok1- CLL13 seemed to be distributed between cytoplasm and GenBank, contains a stretch of 14 basic amino acids nucleus, with a significant fraction accumulating in the (Figure 1d) that may affect the subcellular localization nucleus (Figure 2c, compare lanes 9 to lanes 7 and 8). and the potential function of Dok1-CLL13 (see below). The presence of Dok1-CLL13 in the cytoplasmic Thus the Dok1 gene can be altered in CLL. fraction could be due to contamination from the nuclear

Oncogene Dok1 mutation in B-CLL S Lee et al 2290

Figure 3 Expression of GFP-Dok1, GFP-Dok1-CLL13 and its variants. (a) Schematic diagrams of full-length GFP-Dok1 and its deletion mutant Dok1 (1–150), and full-length GFP-Dok1-CLL13 and its variants, Dok1-CLL13 (1–220) (the N-terminus region) and Dok1-CLL13 (221–298) (the C-terminus region) are represented. The numbers indicate amino-acid coordinates. The PH domain and PTB domain are indicated. The hatched box in GFP-Dok1-CLL13 indicates the newly created 78 amino acids. (b) The protein expression of each GFP construct was checked by immunoblotting using rabbit anti-GFP and mouse anti-b-actin after transient transfection into 293T cells. The GFP fusion proteins and b-actin are indicated Figure 2 Expression of the Dok1-CLL13 protein and fractionation study. (a) Equal amounts of expression plasmids, pcDNA3 vector (lane 1), pcDNA3-F-Dok1 (lane 2) and pcDNA3-F-Dok1- CLL13 (lane 3) were transiently transfected into 293T cells and 48 h after transfection, cellular lysates were subjected to immunoblot- basically a nuclear protein and the localization in the ting using M5 anti-Flag. F-Dok1 and F-Dok1-CLL13 proteins are nucleus can probably be attributed to the new sequence indicated. (b) pcDNA3 vector (lanes 1 and 2), pcDNA3-F-Dok1 (lanes 3 and 4) and pcDNA3-F-Dok1-CLL13 (lanes 5 and 6) were at its carboxyl-terminus (see below). stably transfected into 293 cells, and lysates were prepared from To further study the subcellular localization of Dok1- representative clones and then immunoblotting was performed CLL13, the entire coding region of the protein was using M5 anti-Flag and mouse anti-b-actin as a loading control. F- cloned in frame with GFP in a pGFP expression vector. Dok1, F-Dok1-CLL13 and b-actin proteins are indicated. (c) Total lysates (L) (lanes 1, 4 and 7), cytoplasmic (C) (lanes 2, 5 and 8) and The GFP versions of Dok1 and its PH domain (Dok1 nuclear (N) (lanes 3, 6 and 9) fractions from 293 stable cell lines (1–150), as well as the N-terminus region, Dok1-CLL13 expressing F-Dok1, F-Dok1-CLL13 and from the control empty (1–220) and in the C-terminus region, Dok1-CLL13 vector cells were prepared. Immunoblotting was performed using (221–298) of Dok1-CLL13, were also generated M5 anti-Flag, rabbit antinucleolin and rabbit anti-b-tubulin. The (Figure 3a). These plasmid constructs were transfected cytoplasmic protein b-tubulin in the cytoplasmic fraction and the high proportion of the nuclear protein nucleolin in the nuclear into 293T cells and the protein expression was evaluated fraction provided controls for the experiment. F-Dok1, F-Dok1- by immunoblotting. As shown in Figure 3b, all the CLL13, b-tubulin and nucleolin proteins are indicated proteins were well expressed in 293T cells with the expected sizes. Again, the GFP-Dok1-CLL13 and its derivative proteins had lower expression levels com- fraction, as a proportion of nucleolin was also found in pared to GFP, GFP-Dok1 or GFP-Dok1 (1–150) the cytoplasmic fraction (Figure 2, compare lanes 3, 6 (Figure 3b, lanes 4–6). We next examined the subcellular and 9 to lanes 2, 5 and 8). Therefore, Dok1-CLL13 is localization of these GFP fusion constructs. As shown in

Oncogene Dok1 mutation in B-CLL S Lee et al 2291 Figure 4, the GFP protein had no specific localization, these results indicate that the mutant Dok1 found in being diffused all over the cell (Figure 4a–c). GFP-Dok1 CLL disease localizes in the nucleus, in contrast to wild- localized strongly in the cytoplasm and in the mem- type Dok1, and more importantly, that the accumula- brane, with no detectable signal in the nucleus (Figure tion in the nucleus is attributable to the unique sequence 4d–f). As expected, GFP-Dok1 (1–150) expressing the in the C-terminal region of the protein. This region may PH domain accumulated exclusively in the plasma contain an NLS that directs the protein to the nucleus membrane, where it associated with phospholipids (see below). Furthermore, the data indicate that the C- (Figure 4g–i). In contrast to GFP-Dok1, and consistent terminal region of Dok1-CLL13 has a dominant effect with the cell fractionation studies (Figure 2c), GFP- over the PH domain to concentrate the protein in the Dok1-CLL13 localized exclusively in the nucleus, with nucleus rather than in the cytoplasm or membrane. small dots in certain areas (Figure 4j–l). As expected, the N-terminus of Dok1-CLL13, GFP-Dok1-CLL13 (1– Dok1-CLL13 contains a functional NLS 220), composed of the PH domain of wild-type Dok1, localized in the membrane just as did GFP-Dok1 (1– The unique C-terminal region of Dok1-CLL13 contains 150) (Figure 4m–o). Interestingly, GFP-Dok1-CLL13 14 basic amino-acid residues whose spacing is reminis- (221–298), which expresses only the C-terminus of cent of NLS. Indeed, the alignment of the Dok1-CLL13 Dok1-CLL13, accumulated in the nucleus, like the amino-acid sequence from 233 to 253 with the sequence full-length Dok1-CLL13 (Figure 4p–r). Taken together, of bipartite NLSs from nucleoplasmin, human GADD34 or herpes simplex-1 or -2 nuclear protein g134.5 (Cheng et al., 2002), strongly suggests that Dok1- CLL13 contains a potential bipartite NLS at its C- terminus (Figure 5a). In order to determine whether the region between 233 and 253 contained a functional NLS sequence, and if the conserved arginine residues were important for the nuclear localization, GFP-Dok1- CLL13 mutants with R233 and R234, R251 and R252, or R233, R234, R251 and R252, replaced by alanines were generated (Figure 5b). Transient transfection in 293T cells followed by immunoblot analysis showed that GFP-Dok1-CLL13 and its R to A mutant forms were expressed at similar levels (Figure 5c). Confocal microscope studies showed that GFP-Dok1-CLL13 (R233A, R234A) and GFP-Dok1-CLL13 (R251A, R252A) be- haved like GFP-Dok1-CLL13 and localized in the nucleus (Figure 6j–o). These results indicate that neither mutation of R233 and R234 on the one hand, nor mutation of R251 and R252 on the other hand, is sufficient to abolish the nuclear localization of Dok1- CLL13. However, converting all four conserved arginine residues (R233, R234, R251 and R252) to alanines relocalized GFP-Dok1-CLL13 from the nucleus to the cytoplasm, with strong accumulation in the plasma membrane (Figure 6p–r). Similar results were obtained using U2OS cells (data not shown). The particular concentration of the protein in the membrane is the consequence of the major effect of the PH domain due to the absence of a functional NLS. These data strongly indicate that Dok1-CLL13, due to the frameshift mutation, has acquired a functional NLS and that the four conserved arginine residues R233, R234, R251 and R252 are required for efficient nuclear localization, which is consistent with the bipartite form of NLS.

Dok1-CLL13 is impaired to inhibit MAP kinase Figure 4 Dok1-CLL13 localizes in the nucleus. The GFP plasmid constructs described in Figure 3a were used to transfect 293T cells. activation At 48 h after transfection of cells grown on cover slides, cells were Dok1 acts as a negative regulator of MAP kinase ﬁxed with 4% paraformaldehyde and then permeabilized with 0.1% Triton before staining with PI. Samples were visualized by activation and inhibits cell proliferation (Di Cristofano GFP (left panel) and PI was used to stain the nucleus (middle et al., 2001; Hosooka et al., 2001). To examine the effect panel). The superimposed signals of GFP and PI are showed by of Dok1-CLL13 on MAPK activation, 293 cells stably merged images (right panel) expressing F-Dok1 and F-Dok1-CLL13, and the control

Oncogene Dok1 mutation in B-CLL S Lee et al 2292

Figure 6 Dok1-CLL13 has a functional NLS. GFP plasmid constructs described in Figure 5b were used for transfection into 293T cells. At 48 h after transfection, cells were treated as in Figure 4. GFP, PI staining and merged image is shown

Figure 5 Dok1-CLL13 has a conserved bipartite NLS. (a) The amino-acid sequence alignment of bipartite NLSs from nucleo- declining to the basal level (Figure 7a). Whereas PDGF- plasmin, g134.5 nuclear protein of HSV-1 and -2 and human GADD34, together with Dok1-CLL13 from amino acids 233–253 induced Erk1/2 phosphorylation was much attenuated are shown. The basic amino acid arginine (R) and lysine (K) are in in cells overexpressing wild-type Dok1 (Figure 7a), bold letters and conserved arginine residues are boxed. The phosphorylation of Erk1/2 was normal or slightly conserved residues R233, R234, R251 and R252 are outlined in elevated in cells that overexpressed Dok1-CLL13. Since the Dok1-CLL13 diagram. (b) Schematic diagrams of GFP-Dok1- CLL13 and its R to alanine (A) mutants: GFP-Dok1-CLL13 the level of expression of Dok1-CLL13 was always low (R233A, R234A) where R233 and R234 were mutated to A, GFP- compared to the wild-type protein, we cannot exclude Dok1-CLL13 (R251A, R252A) where R251 and R252 were the possible inhibition of Erk phosphorylation when mutated to A. GFP-Dok1-CLL13 (R233A, R234A, R251A, Dok1-CLL13 is expressed a high level (Figure 7a). R252A) where R233, R234, R251, R252 were replaced by A. (c) Protein expression of the indicated GFP constructs was checked by Nevertheless, Dok1 mutated in CLL lost its ability to immunoblotting after transient transfection into 293T cells. The inhibit MAP kinase activity due to an intrinsic loss of membrane was hybridized with rabbit anti-GFP and mouse anti-b- function, or a low level of expression. To check whether actin other signaling pathways are affected by Dok1-CLL13, tumor necrosis factor-alpha (TNF-a)-mediated phosphorylation and degradation of IkBa, the inhibitor of NF-kB, was evaluated. The phosphorylation and cell line containing empty vector, were stimulated by degradation of IkBa occurred similarly in the control PDGF at different time points. After stimulation, the cells, as well as in cells expressing Dok1 or Dok1-CLL13 lysates were subjected to Western blot analysis with after stimulation by TNF-a (Figure 7b). This indicates rabbit antiphospho Erk1/2. Activation of MAP kinase that like wild-type Dok1, Dok1-CLL13 has no effect on Erk1/2 by PDGF occurred in the control cell line 5 min the TNF-a-induced NF-kB activation pathway, as after stimulation, reaching a peak at 10 min before monitored by IkBa phosphorylation and degradation.

Oncogene Dok1 mutation in B-CLL S Lee et al 2293

Figure 7 Effects of Dok1-CLL13 on MAP kinase and NF-kB activation. (a) Stable cell lines were stimulated with PDGF (20 ng/ ml) at different time points and equal amounts of protein extracts were subjected to immunoblotting using rabbit antiphospho Erk1/ 2, goat anti-Erk1/2, M5 anti-Flag and mouse anti-b-actin. (b) Each stable cell line was stimulated with TNF-a (20 ng/ml) at different Figure 8 Lck induces Dok1-CLL13 tyrosine-phosphorylation and time points and the protein extracts were analysed as in (a) using mediates its association with wild-type Dok1 (a) 293 cells stably rabbit antiphospho IkBa, rabbit anti-IkBa, M5 anti-Flag and expressing F-Dok1 or F-Dok1-CLL13 were stimulated ( þ ) with mouse anti-b-actin pervanadate (PV) or left untreated (À) and cell lysates were immunoprecipitated (IP) with M2 beads anti-Flag. After electro- phoresis, immunoblotting (IB) was performed by probing sequen- Dok1-CLL13 is tyrosine-phosphorylated and forms tially the membrane with mouse anti-pTyr and anti-Flag. IgG designates immunoglobulin chains. (b) 293T cells were transfected heterodimers with wild-type Dok1 with F-Dok1 (left panel) or F-Dok1-CLL13 (right panel) together with empty vector or expression plasmid for p210Bcr–abl, c-Src, Csk Dok1 is a tyrosine-phosphorylated protein (Carpino and Lck as indicated. At 1 day after transfection, lysates were et al., 1997). To determine whether Dok1-CLL13, which immunoprecipitated and immunoblotting was performed as in (a). contains the PH and part of the PTB domain with five (c) 293T cells were transfected with F-Dok1-CLL13 alone or with tyrosine residues (Figure 1d), can be tyrosine phos- the indicated combinations of Myc-Dok1 and Lck. At 1 day after phorylated, 293 cells stably expressing F-Dok1 and F- transfection, cell lysates were prepared and small fractions were used to determine Dok1 levels by immunoblot analysis. The Dok1-CLL13 were stimulated with pervanadate, an remaining portion of the lysates were immunoprecipitated and inhibitor of tyrosine phosphatases. As shown in immunoblotting was performed as in (a) using mouse anti-pTyr, Figure 8a, tyrosine phosphorylation of F-Dok1-CLL13 anti-Dok1 and anti-Flag was induced after pervanadate stimulation. The extent of tyrosine phosphorylation of F-Dok1-CLL13 is lower (5 tyrosine residues) than in F-Dok1 (15 tyrosine Csk to phosphorylate Dok1-CLL13 is not dependent on residues) (van Dijk et al., 2000). To determine which the cell type, but resulted from the structure of Dok1- tyrosine kinases can phosphorylate Dok1-CLL13, CLL13 protein. In view of homodimerization of Dok1 p210Bcr–Abl, the Src family members c-Src, Lck and the (Songyang et al., 2001), heterodimerization of Dok1- Src inhibitor kinase Csk were tested. Whereas CLL13 and wild-type Dok1 was examined by transfec- p210Bcr–Abl, c-Src, Csk and Lck efficiently phosphory- tion in 293T cells. Both the endogenous Dok1 and the lated F-Dok1, F-Dok1-CLL13 was phosphorylated only overexpressed myc-Dok1 associated with F-Dok-CLL13 by Lck (Figure 8b), suggesting that one or several (Figure 8c, lanes 1, 2, 5 and 6 and data not shown). This tyrosine residues in Dok1-CLL13 are targets for Lck, association increased in the presence of Lck (Figure 8c, but not for Abl, Src or Csk in 293T cells. However, it compare lanes 5 and 6), indicating that tyrosine has been reported that Dok277 containing the first 277 phosphorylation mediated the interaction (Songyang amino acid of Dok1 can be tyrosine phosphorylated by et al., 2001). Remarkably, coexpression of myc-Dok1 Src in NIH3T3 cells (Songyang et al., 2001). Since reproducibly enhanced the tyrosine phosphorylation of Dok1-CLL13 contains the same tyrosine residues as F-Dok1-CLL13 by Lck (compare lanes 2 and 6). In Dok277, tyrosine phosphorylation of Dok1-CLL13 was addition, the phosphotyrosine content of F-Dok1- examined in NIH3T3. As in 293T cells, only Lck CLL13 was more pronounced than that of the phosphorylated Dok1-CLL13 in NIH3T3 cells (data coexpressed myc-Dok1 (Figure 8c, compare F-p-Dok1- not shown). This suggests that the failure of Abl, Src or CLL13 to Myc-p-Dok1 in lane 6). Taken together, these

Oncogene Dok1 mutation in B-CLL S Lee et al 2294 results indicate that Dok1 either stimulates the kinase in various cell types, including immortalized lympho- activity of Lck towards Dok1-CLL13, or recruits an blastoid cell lines and mononuclear cells from healthy endogenous kinase that can phosphorylate Dok1- individuals, as a result of alternative splicing (Hubert CLL13. It also suggests that F-Dok1-CLL13, with its et al., 2000). The 35 kDa protein Dok1-CLL13 found in much lower tyrosine content, is probably a preferential B-CLL is different from p22Dokdel at the amino-acid substrate for Lck compared to Dok1 in a tripartite sequence level. In addition, p22Dokdel lacks any complex Lck/Dok1/Dok1-CLL13. potential NLS that could cause its localization in the nucleus. Thus the 35 kDa protein Dok1-CLL13 is a novel mutated form of Dok1 found in B-CLL disease. Discussion The frameshift mutation identified in B-CLL is likely to affect several important functional properties of Dok1 is known to be a negative regulator of several cell Dok1 that are ascribed to its C-terminal region. In fact, signaling pathways: it inhibits cell proliferation, down- this region has been shown to be important for Dok1 to regulates MAP kinase activity and inhibits cell trans- suppress cell proliferation and cell transformation formation and leukemogenesis (Yamanashi et al., 2000; (Hosooka et al., 2001; Songyang et al., 2001; Shah and Di Cristofano et al., 2001; Songyang et al., 2001; Wick Shokat, 2002). In addition, tyrosine residues 296 and et al., 2001; Zhao et al., 2001). In addition, the 315, which are necessary for Dok1 to interact with Ras- localization of the Dok1 gene at 2p13, a chromosomal GAP in order to downregulate MAP kinase activity region rearranged in leukemia (Nelms et al., 1998) (Songyang et al., 2001; Shah and Shokat, 2002), are lost suggests that changes in its structure are expected during in Dok1-CLL13. Consistent with this observation, tumor initiation or promotion. However, a direct role of PDGF-induced MAP kinase activation is not affected, Dok1 in human malignant disease has yet to be and is even slightly increased in 293 cells expressing demonstrated. Our data provide evidence for the first Dok1-CLL13, compared to wild-type Dok1, which has time that mutation of the Dok1 gene may be involved in an inhibitory effect (Figure 7a). On the other hand, NF- human tumors like B-CLL. A four-nucleotide GGCC kB signaling is not affected by either wild-type Dok1 or deletion at position 683 in the coding sequence of the Dok1-CLL13 (Figure 7b). Also tyrosine residues 296 Dok1 gene was found in one of the 46 B-CLL patients and 362, known to be necessary for Dok1 to inhibit Src- analysed. No chromosomal alteration affecting 2p13 induced cell transformation (Songyang et al., 2001; Shah region was detected in this patient (Dumontet et al., and Shokat, 2002), are missing in Dok1-CLL13. In unpublished observation). The mutation, which was not addition, partial loss of the PTB domain, which also found in healthy individuals or in tumor samples from plays a role in the inhibition of cell transformation 52 patients affected by acute myeloid leukemia (AML) (Songyang et al., 2001), could contribute to affect Dok1 (Lee et al., unpublished results), could be a rare event function in Dok1-CLL13. restricted to a small set of B-CLL. The GGCC deletion Dok1-CLL13 is a tyrosine-phosphorylated protein results in a frameshift in the Dok1 coding sequence, and among the tyrosine kinases Bcr-Abl, Src, Csk and generating a hybrid Dok1 protein truncated at the C- Lck that phosphorylate efficiently wild-type Dok1, only terminus, with unique properties. In this mutant protein, Lck phosphorylated Dok1-CLL13 (Figure 8b). The the C-terminal region of wild-type Dok1 from amino Lck-mediated Dok1-CLL13 phosphorylation increased acid 220–481 is replaced by a novel sequence of 78 in the presence of wild-type Dok1 and enhanced the amino acids with no significant homology in GenBank heterodimerization of Dok1 and Dok1-CLL13 (Figure 1). In contrast to wild-type Dok1, which (Figure 8c). Dok1 may enhance Lck activity toward localizes in the cytoplasm and membrane, the mutant Dok1-CLL13, either by conformation change, or by protein, Dok1-CLL13, accumulates almost exclusively recruiting additional tyrosine kinases as it has been in the nucleus (Figure 4), due to the presence of a proposed for the increased tyrosine phosphorylation of bipartite NLS in the unique C-terminal region of Dok- Tec by Dok1 (van Dijk et al., 2000). The tyrosine CLL13. Since the mutant protein can be stably residue(s) of Dok1-CLL13 that can be phosphorylated expressed in 293 cells, we presume that it is expressed by Lck remain to be determined. Tyrosine 203 of Dok1- in leukemia cells as well, even though this was not CLL13 (and Dok1) (Figure 1) that is in the tyrosine verified due to a lack of suitable tumor samples and binding motif (YXX(I/L/V) for Lck SH2 domain, could appropriate antibodies. Several attempts to immortalize be a target for Lck (Nemorin and Duplay, 2000). the few available CLL-13 B-cells with Epstein–Barr However, phosphorylation of tyrosine 146 present in the (EBV), in order to identify the native CLL13 protein, PTB domain of both Dok1 and Dok1-CLL13 can also were unsuccessful. Converting the four-conserved argi- be induced by Lck, since phosphorylation of this residue nine residues R233, R234, R251 and R253 to alanines in is necessary for Dok1 homodimerization (Songyang the putative NLS reversed the Dok1-CLL13 nuclear et al., 2001). Therefore the fraction of cytoplasmic localization, bringing the protein into the cytoplasm Dok1-CLL13 might associate with wild-type Dok1 with a pronounced concentration in the plasma mem- through tyrosine 146 and act as a dominant negative brane (Figure 6). to inhibit Dok1 function, and Lck can play an A truncated Dok1 isoform referred to as p22Dokdel important role in this inhibition. Interestingly, Lck has recently been reported (Hubert et al., 2000). This expression is known to be restricted to T cells, but it is protein is coexpressed with the full-length Dok1 protein frequently activated in B-cell neoplasias including CLL

Oncogene Dok1 mutation in B-CLL S Lee et al 2295 (Abts et al., 1991; Jucker et al., 1991; Von Knethen et al., immunophenotypic analysis using flow cytometry (Matutes 1997). Therefore, specific tyrosine phosphorylation of score of 4/5 or 5/5). All samples contained more than 15 000 Dok1-CLL13 by Lck could be biologically relevant. cells/ml, and more than 80% were leukemic cells. Normal Finally, the mislocalization of Dok1 and the relative low blood samples were obtained from 11 healthy volunteers. expression of the mutant protein may have a dramatic Lymphoid cells were separated by Ficoll Hypaque gradient centrifugation. Patients and health individuals gave informed effect on the overall function of Dok1. It is also consent for this study. plausible that this mislocalization removed Dok1 from a signaling cascade important for its function. Thus, the RNA and cDNA preparation mutation identified in B-CLL appears to inactivate Dok1. Low expression of wild-type protein due to Total RNA was isolated from lymphoid cells using TRI- methylation of the promoter region, or mutation in the REAGENTt (Sigma, St Louis, MO, USA) according to the key regulatory sequence is another mechanism leading manufacturer’s instructions. RNA was then dissolved in 1 to the inactivation of Dok1. Indeed, Dok1 expression RNase-free H2O and stored at À80 C. For synthesis of first- strand cDNA, 2 mg of total RNA was mixed with 500 ng of has been found downregulated in the leukemia cells of a pD(N6) primer (Amersham, London, UK) in a total volume of small subset of B-CLL patients (Galmarini et al., 12 ml, heated at 701C for 10 min and then cooled quickly on ice. unpublished data). A volume of 8 ml of a mixture including 100 mM DTT, 5 Â In contrast to most malignant diseases that are reverse transcriptase buffer, 10 mM dNTP and SuperScriptt II characterized by uncontrolled cell proliferation, B- RNase H-Reverse Transcriptase (Invitrogen, Carlsbad, CA, CLL is believed to have a defect in the regulation of USA) were added to the tube, and the reaction was incubated apoptosis and a lack of cell cycle progression (Caligaris- at 421C for 5 min before inactivation at 701C for 10 min. The Cappio and Hamblin, 1999). Therefore, it is possible synthesized cDNA mixture was diluted 1/20 with distilled that mutation of Dok1 in B-CLL does not necessarily water before mutation screening by heteroduplex analysis. imply the acquisition of an uncontrolled cell proliferation phenotype, but rather the prolongation of the Heteroduplex analysis chronic stage of the disease. Indeed, stable 293 cell lines The sequence of the primers used for amplification of Dok1 expressing Dok1-CLL13 grow similarly to cell lines cDNA (GenBank Accession No. NM-001381) is shown in expressing wild-type Dok1, that is, at a low rate Table 1. PCR reactions were conducted in a Perkin-Elmer compared to control cells (Lee et al., unpublished 9600 thermal cycler. Briefly, about 50 ng cDNA were used as a observation). As a matter of fact, patient CLL13 was template in a 15 ml reaction containing 1.0 U of Taq DNA particularly resistant to a number of chemotherapeutic polymerase (Invitrogen), 200 mM each of dGTP, dATP, dTTP and dCTP, 1.0 mCi of [a-33P]dCTP and 1 mM of primer agents used in the routine treatment of B-CLL disease, (Amersham, London, UK). Cycles were as follows; 961C for such chlorambucil, fludarabine and CHOP chemother- 4 min, followed by 30 cycles of 961C for 30 s, adjusted apy. However the link between this observation and annealing temperature for 30 s, 721C for 2 min, followed by mutation of Dok1 in B-CLL remains to be established, 10 min at 721C. The PCR products were slowly cooled to room as no significant difference in the cytotoxicity and temperature. Samples mixed with loading buffer (95% apoptosis studies was observed between 293 cells formamide, 10 mM NaOH, 0.05% bromophenol blue, 0.05% expressing Dok1 or Dok1-CLL13 (Galmarini et al., xylene Cyanol) were run at room temperature on mutation unpublished data), probably due to the type of cells. It is detection enhancement (MDE) gel. worth noting that chromosomal rearrangements leading to a novel fusion protein with transcription activation Direct sequencing analysis properties are common in several forms of acute and The amplified DNA fragment was purified using a PCR chronic leukemia (Rabbitts, 1991). In this context, the product purification kit (Qiagen, Chatsworth, CA, USA). nuclear localization of Dok1-CLL13 might be function- Purified products were sequenced using a Bigdye terminator ally relevant. Nevertheless, since tumorigenicity is kit (ABI PRISM DNA sequencing kit, Perkin-Elmer) and ABI assumed to be a multistep process in which various 3100 DNA sequencer and the sequencing result was analysed genetic changes lead to tumor formation, the mutation with the BioEdit sequence alignment editor (Hall, 1999). of Dok1, in conjunction with other identified altera- tions, may be important in the progression and Plasmids, cloning and mutagenesis maintenance of a small fraction of B-CLL disease. The Dok1 expression plasmid, pcDNA3-F-Dok1, has been Further studies on a larger scale are required to evaluate previously described (Sylla et al., 2000). F designates the the potential role of Dok1 as a tumor suppressor in B- presence of a FLAG epitope in the N terminus. To construct CLL and other leukemias. the expression plasmid pcDNA3-F-Dok1-CLL13, cDNA from patient CLL13 (patient no. 13 with chronic lymphocytic leukemia) was first amplified using PCR primers F1A and R5 (see Table 1). The PCR product was then cloned in TA cloning Materials and methods pCR4 vector using a TOPO TA cloning Kit (Invitrogen) to obtain pCR4-Dok1-CLL13. pcDNA3-F-Dok1 has two Bam- Patients HI sites (nucleotides 476 and 1096) in the coding region and Blood samples were obtained from 46 patients with B-CLL one NotI site at 30 in the pcDNA3 vector; pCR4-Dok1-CLL13 obtained at the Department of Hematology of the Hospital contained the BamHI site (position nucleotide 476) and a NotI Edouard Herriot (Lyon, France). Diagnosis of CLL was site at 30 in the pCR4 vector. To obtain pcDNA3-F-Dok1- performed by an experienced cytologist and confirmed by CLL13, the BamHI (position 476)/NotI fragments of

Oncogene Dok1 mutation in B-CLL S Lee et al 2296 pcDNA3-F-Dok1 were substituted by the corresponding R Kobayashi (MD Anderson Cancer Center, Houston, Texas, fragment of pCR4-Dok1-CLL13 carrying the GGCC deletion USA) and J Cambier (St Jude Children’s Research Hospital, at nucleotide 683. To construct pGFP-Dok1-CLL13, Memphis, TE). pcDNA3-F-Dok1-CLL13 was amplified using forward primer CLL13 (F)-EcoR23(50-GGAATTCCATGGACGGAG- Immunoprecipitation and Western blotting CAGTGATGGA-30) corresponding to amino acid 1 in the coding region of Dok1-CLL13 (underlining indicates an added Cultured cells were washed with ice-cold PBS and were lysed in EcoRI site) and the reverse primer, CLL13(R)-Kpn919 an ice-cold lysis buffer (50 mM Tris-HCl, pH 7.4; 150 mM (50-GGGGTACCCCCTAAGGGCTCAGCATACAGG-30) NaCl; 2 mM EDTA; 1% Nonidet P-40; 3% glycerol, 1 mM corresponding to amino acid 220 in the coding region of Dok1- PMSF; 1 mM b-glycerophosphate, 5 mg/ml leupeptin, 2 mg/ml CLL13 (underlining indicates an added KpnI site; italics aprotinin; 1 mM sodium fluoride, 1 mM sodium orthovanadate) indicate stop codon). PCR products were digested by EcoRI/ for 30 min on ice. The lysates were centrifuged at 14 000 r.p.m. NotI and the fragment was inserted into EcoRI/KpnI sites in for 20 min. Lysates were quantified using a protein assay frame with GFP of the pEGFP-C1 vector (Clontech, Palo reagent (BIORAD) and analysed on SDS–polyacrylamide gel. Alto, CA, USA). Similarly, pGFP-Dok1-CLL13 (1–220) was Immunoprecipitation was performed as previously described constructed using the forward primer CLL13 (F)-EcoR 23 and (Sylla et al., 2000). Proteins on the gel were transferred to the reverse primer CLL13(R)-Kpn682 (50-GGGGTACCCC PDVF membranes and probed with antibodies. Membranes CTAGGCCTCGAAAGAGAACATGA-30 (underlining indi- were then visualized by using chemiluminescent luminol cates an added KpnI site; italics indicate stop codon) reagents (Santa Cruz). The nuclear and cytoplasmic fractions 6 corresponding to amino acid 220 of the coding region of were prepared as follows: 2 Â 10 of stably transfected cells Dok1-CLL13. To construct pGFP-Dok1-CLL13 (221–298), were washed twice with ice-cold PBS and then lysed with the forward primer, CLL13 (F)-EcoR 683 (50- cytoplasmic lysis buffer (20 mM HEPES, pH 7.6; 20% glycerol; GGAATTCCGCCGCTGCCCCTCAGGCCCT-30) (underlin- 10 mM NaCl, 1.5 mM MgCl2; 0.2 mM EDTA; 0.1% NP40; ing indicates added EcoRI site) corresponding to amino acid 1mM DTT; 1 mM PMSF; 1 mM b-glycerophosphate, 5 mg/ml 221 of the coding region of Dok1-CLL13 and the reverse leupeptin, 2 mg/ml aprotinin; 1 mM sodium fluoride, 1 mM primer CLL13(R)-Kpn919 were used. GFP-Dok1 CLL13 sodium orthovanadate). After 10 min on ice, lysed samples variants bearing specific mutations were obtained by PCR were centrifuged at 2000 r.p.m. for 4 min to recover the mutagenesis. pGFP-Dok1-CLL13 (R233A, R234A), pGFP- cytoplasmic fraction in the supernatant. The pellets were Dok1-CLL13 (R251A, R252A) and pGFP-Dok1-CLL13 mixed with a nuclear lysis buffer (20 mM HEPES, pH 7.6; 20% (R233A, R234A, R251A, R252A) are Dok1-CLL13 mutants glycerol; 500 mM NaCl, 1.5 mM MgCl2; 0.2 mM EDTA; 0.1% in which the respective arginine (R) residues at 233 and 234, or NP40; 1 mM DTT; 1 mM PMSF; 1 mM b-glycerophosphate, 251 and 252, or 233, 234, 251 and 252 were replaced by alanine 5 mg/ml leupeptin, 2 mg/ml aprotinin; 1 mM sodium fluoride, (A). All mutations and constructs were confirmed by sequen- 1mM sodium orthovanadate) incubated on ice for 30 min, and cing analysis. The expression plasmids for p210bcr–abl, c-Src, then centrifuged at 14 000 r.p.m. for 30 min to recover super- Csk and Lck were obtained respectively from A-M Pendergast natant containing the nuclear fractions. (Duke University, Durham, NC, USA), A August (Pennsyl- vania State University, Philadelphia, PA, USA), W Eckhart Confocal microscopy and immunofluorescence (Salk Institute, San Diego, CA, USA) and R Perlmutter Cells were seeded on a sterilized cover glass in six-well plates at (Merck & Co., Inc., Rahway, NJ, USA). 0.1 Â 106 cells per well. At 24–48 h after transfection with GFP expression plasmids, cells on the cover glass were washed twice Cell culture and transfection with PBS and fixed with 4% paraformaldehyde for 10 min. HEK 293 cells (adenovirus E1a- and E1b-transformed human Fixed cells were permeabilized with 0.1% Triton for 5 min, embryonic kidney) and HEK 293 cells expressing simian virus stained with propidium iodide (PI) and visualized using a Zeiss 40 large T antigen (293T) were maintained in Dulbecco’s confocal LSM 510 microscope. modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS). For transient or stable transfection, MAP kinase and NF-kB assays 1 mg of empty vector pcDNA3, pcDNA3-Flag-Dok1 or Cells (0.4 Â 106) were seeded in six-well plates and cells were pcDNA3-Flag-CLL13 was introduced into HEK 293 using a incubated in DMEM containing 10% FBS for 24–48 h. Cells Superfect kit (Qiagen) according to the manufacturer’s were then stimulated with 20 ng/ml of platelet-derived growth instructions. For stable transfection, G418 (800 mg/ml)-resis- factor (PDGF) (R&D System, Minneapolis, MN, USA) or tant colonies were selected and expanded. Protein expression 20 ng/ml of TNF-a (Sigma) at different times. After stimula- was confirmed by immunoblotting using M5 anti-Flag mouse tion, cells were lysed in 200 ml of lysis buffer on ice for 30 min, monoclonal antibody. and lysates were then subjected to immunoblotting analysis using specific antibodies for phosphorylated Erk1/2, total Antibodies Erk1/2 proteins, phosphorylated IkBa, or total IkBa protein. M5 anti-Flag mouse monoclonal antibody was obtained from Sigma. The rabbit polyclonal antibodies against phosphory- Acknowledgements lated Erk1/2, phosphorylated IkBa and total IkBa were We thank Drs A-M Pendergast, A August , W Eckhart, R purchased from New England Biolabs (Beverly, MA, USA). Perlmutter, R Kobayashi, and J Cambier for reagents, Mrs S Rabbit antinucleolin antibody and goat antibodies against Pauly for EBV transformation assay, Drs H Huang, V total Erk1/2 were obtained from Santa Cruz Biotechnology Krutovskikh, M Tommasino, S Tavtigian, J Hall, O Serova (Santa Cruz, CA, USA), and rabbit anti-b-tubulin antibody and G Lenoir for advice and critical reading of the manuscript. was obtained from Roche (Mannheim, Germany). Mono- We are also grateful to Dr Z-Q Wang for support and clonal antibody against b-actin was purchased from ICN comments, Dr J Cheney and Mrs T Perdrix-Thoma for editing Biomedicals. Rabbit anti-Dok1 antibodies were obtained from and G Mollon for illustrations.

Oncogene Dok1 mutation in B-CLL S Lee et al 2297 References

Abts H, Jucker M, Diehl V and Tesch H. (1991). Leuk. Res., Nemorin JG and Duplay P. (2000). J. Biol. Chem., 275, 14590– 15, 987–997. 14597. Bhat A, Johnson KJ, Oda T, Corbin AS and Druker BJ. Nemorin JG, Laporte P, Berube G and Duplay P. (2001). J. (1998). J. Biol. Chem., 273, 32360–32368. Immunol., 166, 4408–4415. Cai H, Dhe-Paganon S, Melendez PA and Shoelson SE. Noguchi T, Matozaki T, Inagaki K, Tsuda M, Fukunaga K, (2003). J. Biol. Chem., 278, 25323–25330. Kitamura Y, Kitamura T, Shii K, Yamanashi Y and Kasuga Caligaris-Cappio F and Hamblin TJ. (1999). J. Clin. Oncol., M. (1999). EMBO J., 18, 1748–1760. 17, 399–408. Rabbitts TH. (1991). Cell, 67, 641–644. Carpino N, Wisniewski D, Strife A, Marshak D, Kobayashi R, Richardson AL, Humphries CG and Tucker PW. (1992). Stillman B and Clarkson B. (1997). Cell, 88, 197–204. Oncogene, 7, 961–970. Cheng G, Brett ME and He B. (2002). J. Virol., 76, 9434–9445. Ringshausen I, Schneller F and Bogner C. (2002). Blood, 100, Cong F, Yuan B and Goff SP. (1999). Mol. Cell. Biol., 19, 3741–3748. 8314–8325. Sattler M, Verma S, Pride YB, Salgia R, Rohrschneider LR Di Cristofano A, Niki M, Zhao M, Karnell FG, Clarkson B, and Griffin JD. (2001). J. Biol. Chem., 276, 2451–2458. Pear WS, Van Aelst L and Pandolfi PP. (2001). J. Exp. Med., Schaffner C, Stilgenbauer S, Rappold GA, Dohner H and 194, 275–284. Lichter P. (1999). Blood, 94, 748–753. El Rouby S, Thomas A, Costin D, Rosenberg CR, Potmesil Schroeder Jr HW and Dighiero G. (1994). Immunol. Today, 15, M, Silber R and Newcomb EW. (1993). Blood, 82, 3452– 288–294. 3459. Shah K and Shokat KM. (2002). Chem. Biol., 9, 35–47. Favre C, Gerard A, Clauzier E, Pontarotti P, Olive D and Songyang Z, Yamanashi Y, LiuD and Baltimore D. (2001). J. Nunes JA. (2003). Genes Immunol., 4, 40–45. Biol. Chem., 276, 2459–2465. Frank DA, Mahajan S and Ritz J. (1997). J. Clin. Invest., 100, Sonnier JA, Buchanan GR, Howard-Peebles PN, Rutledge J 3140–3148. and Smith RG. (1983). N. Engl. J. Med., 309, 590–594. Furman RR, Asgary Z, Mascarenhas JO, Liou HC and Stankovic T, Weber P, Stewart G, Bedenham T, Murray J, Schattner EJ. (2000). J. Immunol., 164, 2200–2206. Byrd PJ, Moss PA and Taylor AM. (1999). Lancet, 353, 26– Grimm J, Sachs M, Britsch S, Di Cesare S, Schwarz-Round T, 29. Alitalo K and Birchmeier W. (2001). J. Cell. Biol., 154, 345– Sylla BS, Murphy K, Cahir-McFarland E, Lane WS, Mosialos 354. G and Kieff E. (2000). Proc. Natl. Acad. Sci. USA, 97, 7470– Hall TA. (1999). Nucleic Acids Symp. Ser., 41, 95–98. 7475. Hosooka T, Noguchi T, Nagai H, Horikawa T, Matozaki T, Tamir I, Stolpa JC, Helgason CD, Nakamura K, Bruhns P, Ichihashi M and Kasuga M. (2001). Mol. Cell. Biol., 21, Daeron M and Cambier JC. (2000). Immunity, 12, 347–358. 5437–5446. van Dijk TB, van Den Akker E, Amelsvoort MP, Mano H, Hubert P, Ferreira V, Debre P and Bismuth G. (2000). Eur. J. Lowenberg B and von Lindern M. (2000). Blood, 96, 3406– Immunogenet., 27, 145–147. 3413. Inaba T, OkuN, Gotoh H, MurakamiS, OkuN, Itoh K, Ura Von Knethen A, Abts H, Kube D, Diehl V and Tesch H. Y, Nakanishi S, Shimazaki C and Nakagawa M. (1991). (1997). Leuk. Lymphoma, 2, 551–562. Leukemia, 8, 719–722. Wattel E, Preudhomme C, Hecquet B, Vanrumbeke M, Jucker M, Abts H, Eick D, Lenoir GM and Tesch H. (1991). Quesnel B, Dervite I, Morel P and Fenaux P. (1994). Blood, Leukemia, 5, 528–530. 84, 3148–3157. Kato I, Takai T and Kudo A. (2002). J. Immunol., 168, 629– White MF. (1998). Mol. Cell. Biochem., 182, 3–11. 634. Wick MJ, Dong LQ, HuD, Langlais P and LiuF. (2001). J. Kitada S, Andersen J, Akar S, Zapata JM, Takayama S, Biol. Chem., 276, 42843–42850. Krajewski S, Wang HG, Zhang X, Bullrich F, Croce CM, Wyatt WA, Meyer R, Flynn H, Bone N, James CD, Jelinek D, Rai K, Hines J and Reed JC. (1998). Blood, 91, 3379–3389. Hanson C, Kay N, Witzig T and Dewald G. (2000). Blood, Lemay S, Davidson D, Latour S and Veillette A. (2000). Mol. 96, 367a. Cell. Biol., 20, 2743–2754. Yamakawa N, Tsuchida K and Sugino H. (2002). EMBO J., Liang X, Wisniewski D, Strife A, Shivakrupa R, Clarkson B 21, 1684–1694. and Resh MD. (2002). J. Biol. Chem., 277, 1372–137387. Yamanashi Y and Baltimore D. (1997). Cell, 88, 205–211. Martelli MP, Boomer J, BuM and Bierer BE. (2001). J. Biol. Yamanashi Y, Tamura T, Kanamori T, Yamane H, Nariuchi Chem., 276, 45654–45661. H, Yamanoto T and Baltimore D. (2000). Genes Dev., 14, Munzert G, Kirchner D, Stobbe H, Bergmann L, Schmid RM, 11–16. Dohner H and Heimpel H. (2002). Blood, 100, 3749–3756. Yoffe G, Howard-Peebles PN, Smith RG, Tucker PW and Neet K and Hunter T. (1995). Mol. Cell. Biol., 15, 4908–4920. Buchanan GR. (1990). J. Pediatr., 116, 114–117. Nelms K, Snow AJ and Noben-Trauth K. (1998). Genomics, Zhao M, Schmitz AA, Qin Y, Di Cristofano A, Pandolfi PP 53, 243–245. and Van Aelst L. (2001). J. Exp. Med., 194, 265–274.

Oncogene