Human TDE1, a TDE1/TMS Family Member, Inhibits Apoptosis in Vitro and Stimulates in Vivo Tumorigenesis

Oncogene (2006) 25, 4549–4558 & 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc ORIGINAL ARTICLE Human TDE1, a TDE1/TMS family member, inhibits apoptosis in vitro and stimulates in vivo tumorigenesis

M Bossolasco1, F Veillette1, R Bertrand1,2 and A-M Mes-Masson1,2

1Centre de recherche du Centre Hospitalier de l’Universite´ de Montreál (CR-CHUM) and Institut du cancer de Montreál, Montreal, Quebec, Canada and 2Department of Medicine, Universite´ de Montreál, Montreal, Quebec, Canada

We have previously described hTDE1, the human homo- transgenic mice (Lebel and Mes-Masson, 1994). This logue of the recently described TDE1/TMS family of gene was also identified separately as TMS-1 (Grossman proteins whose members have been identified in several et al., 2000) and AIGP1 (Aoki et al., 2002) as well as species. Although a defined biochemical activity has yet to Serinc3, described as a carrier protein that incorporates be assigned to TDE1/TMS family members, previous a polar-amino acid serine into cell membranes (Inuzuka results point to the overexpression of family members in et al., 2005). In addition, we also cloned and charac- tumor cell lines or tissues. To define whether hTDE1 may terized the human homologue of TDE1 (hTDE1) from a directly impact on neoplastic transformation, we derived human placental cDNA library (Bossolasco et al., 1999). and characterized stable Rat-1 transfectants that consti- The identification of additional family members gave tutively express hTDE1 at the plasma membrane. rise to the notion of a larger TDE1/TMS family of Expression of hTDE1 in Rat-1 transfectants had a proteins (Krueger et al., 1997; Grossman et al., 2000), significant effect on cell contact inhibition in vitro as and members of this family have been identified in judged by a focus formation assay. In addition, by several species including Caenorhabditis elegans, Droso- monitoring caspase-3 activity and Hoechst staining, we phila melanogaster, Fugu rubripes, Rattus norvegicus, determined that hTDE1 overexpression partially protects Mus musculus and Homo sapiens. The members of this cells from serum starvation- and etoposide-induced family share between 30 and 80% homology, contain apoptosis. Finally, hTDE1 Rat-1-expressing clones ac- multiple hydrophobic domains and often include an celerated the formation of tumors in a nude mouse assay. N-terminal signal peptide. In mouse, at least five Our results suggest that hTDE1 contributes directly to homologues have been identified, scattered over four oncogenesis in vivo that may in part be explained by its different chromosomes, whereas human TDE1/TMS effect on apoptosis in vitro. genes have been found on five different chromosomes Oncogene (2006) 25, 4549–4558. doi:10.1038/sj.onc.1209488; (The Wellcome Trust Sanger Institute, Ensemble). published online 20 March 2006 When comparing the sequence between human and mouse genes, it has been noted that most homologues Keywords: TDE1; apoptosis; transformation; tumorigenesis between these two species locate on syntenic regions at the chromosome level suggesting evolutionary conserva- tion (Table 1). The conserved nature of the gene structure and of the proteins sequence through evolution suggests an important biological role for TDE1/TMS Introduction family members. Although membership to the TDE1/TMS family is Transgenic mice carrying the polyomavirus large T expanding, there is still no evidence for a precise antigen under the control of the metallothionein biochemical function for these proteins. Several groups promoter develop tumors after long latencies, suggesting have conducted studies related to select family members. that secondary events contribute to oncogenesis in these Mouse TDE1 was originally described as being over- mice (Chalifour et al., 1990, 1992). Using this model expressed in testicular tumors of transgenic mice and in system, we previously identified a novel gene, which we cell lines derived thereof (Lebel and Mes-Masson, 1994). named TDE1 (for Tumor Differentially Expressed) Interestingly, Although tumor-derived transgenic cell following a screen of differentially expressed genes in lines generally overexpressed TDE1, this overexpression cell lines derived from pre- or post-adematous testes of was more apparent in vivo in transgenic tissues. Similarly, we were able to detect high expression of hTDE1 in lung cancer tissue samples as compared to Correspondence:Dr A-M Mes-Masson, Centre de recherche CHUM/ normal lung tissues, although again the degree of ICM, Hoˆ pital Notre-Dame, 1560 Sherbrooke Street East, Montreal, overexpression was more marked in tumor tissues versus Quebec, Canada H2L4M1. E-mail:[email protected] cell lines (Bossolasco et al., 1999). Overexpression of Received 27 January 2005; revised 9 January 2006; accepted 9 January TMS-1/TDE1 was also noted in a small liver tumor 2006; published online 20 March 2006 (Grossman et al., 2000). Whereas TMS-1/TDE1 is Oncogenic properties of hTDE1 M Bossolasco et al 4550 Table 1 Summary of TDE1/TMS family members and comparison of human and mouse family members Mouse-gene name Chromosome location % protein homology with TDE1 Ensemble ID

(a) Summary of TDE1/TMS family membersa TDE1/TMS-1/AIGP1 2 99–100 ENSMUSG00000017707 RIKEN MGI 2441842 2 31 ENSMUSG00000046110 RIKEN MGI 1919132 4 49 ENSMUSG00000023232 TMS-2 10 57 ENSMUSG00000019877 RIKEN MGI 2444223 13 37 ENSMUSG00000021703

Human gene name Chromosome location % protein homology with HTDE Ensemble ID HTDEl 20 100 ENSG00000132824 TDE2L 1 50 ENSG00000168528 TPO1 (c5orf12) 5 38 ENSG00000164300 TDE2 6 58 ENSG00000111897 47 kDa protein [AccIPI00376054] 15 32–34 ENSG00000184716

(b) Potential mouse/human homologuesb % protein homology Chromosome sinteny TDE1/TMS-1/AIGP1:hTDEl 77 Yes RIKEN MGI 2441842:47 kDa protein [AccIPI00376054] 71–91 Yes RIKEN MGI 1919132:TDE2L 85 Yes TMS2:TDE2 95 No RIKEN MGI 2444223:TPO1 (c5orf12) 81 Yes

aInformation on family members was obtained from The Wellcome Trust Sanger Institute, Ensemble server, July 2004. bComparison of human and mouse family members. Percent identity was evaluated using the BLAST 2 SEQUENCES tool available on National Center for Biotechnology Information (NCBI, http://www.ncbi.nlm.nih.gov/blast/bl2seq/wblast2.cgi).

generally ubiquitously expressed (Lebel and Mes- quantitative PCR on lung tumors and nonmalignant Masson, 1994; Grossman et al., 2000), a second mouse lung tissues. In both experiments, TDE2 transcript family member, TMS-2, appears to localize in the mouse expression was higher in NSCLC samples compared to brain, and is more specifically associated with areas with expression levels in nonmalignant bronchiole epithe- glutamatergic excitatory neurons such as the hippocam- lium. More recently, a biochemical function was pus and cerebral cortex (Grossman et al., 2000). suggested for several members of the family. Inuzuka Although the protein location might suggest a role in and colleagues have been able to show that some vesicular glutamate transport, in vitro experiments members of TDE1/TMS family would act as carrier aimed at demonstrating a glutamate uptake into cells proteins, allowing incorporation of serine amino acid expressing transfected TMS-2 were unsuccessful (Gross- into membranes and thus facilitating the synthesis of man et al., 2000). It has been separately reported that phosphatidylserine and sphingolipids. TMS-1/TDE1 (AIGP1) expression is increased in mouse Given that overexpression of TDE1/TMS family cortical neurons after in vivo axotomy (Aoki et al., members has been associated with oncogenic transfor- 2002). When in vitro-cultured mouse neurons were mation in mice (Lebel and Mes-Masson, 1994; Gross- subjected to several endoplasmic reticulum (ER)–Golgi man et al., 2000) and humans (Bossolasco et al., 1999; stress conditions, the same increase in AIGP1 expression Player et al., 2003), we were interested in directly was observed, suggesting that the protein might act as a accessing whether hTDE1 contributes directly to cell sensor protein for ER–Golgi stress (Aoki et al., 2002). growth and transformation. Here we report on the Two new human members of the TDE1/TMS family stable expression of hTDE1 in a Rat-1 cell model and have also been cloned. One of them, C5orf12 (Xu et al., the characterization of transfectants by several in vitro 2003), is the human homologue of the TPO1 rat gene and in vivo assays associated with oncogenic transfor- (Krueger et al., 1997). Based on its pattern of expres- mation. We further describe the construction, expression, it has been postulated that TPO1 is an essential sion and effect on transformation of two N-terminal and component in the transition from proliferative pro- one C-terminal deletion mutants of hTDE1. oligodendroblasts to postmitotic oligodendrocytes and that it plays a role in the rat myelin biogenesis (Krueger et al., 1997) and an analogous role in human brain has been suggested on the basis of sequence homology (Xu Results et al., 2003). An additional human family member, TDE2, was identified by serial analysis of gene expres- Expression and subcellular localization of hTDE1 and sion in non-small-cell lung cancers (NSCLC) (Player its mutants et al., 2003). TDE2 was selected among several other As previous attempts to generate a polyclonal antibody differentially expressed genes because of its high level of were unsuccessful, we generated a fusion protein, PGK- differential expression when comparing lung tumors and hTDE1-GFP (PGK, phosphoglycerate kinase; GFP, primary lung epithelial cells. The authors performed green fluorescent protein) consisting of full-length in situ hybridization analysis as well as real-time hTDE1 fused in-frame just before the native stop codon

Oncogene Oncogenic properties of hTDE1 M Bossolasco et al 4551 a Transmembrane segments I II III IV VI V VII VIII IX X XI Lumen

2 3 Cytoplasm

b c 4 d

E1-34

-GFP

17hTD 112hTDE1-2 126hTDE1-11 17hTDE1-15 126hTDE1-7 112hTDE1-6 112hTDE1-8

∆ ∆ ∆ ∆ ∆ ∆ ∆

Rat-1 N C PGK-hTDE1-24 N PGK-hTDE1-34 N N C C 126hTDE1-GFP 112hTDE1-GFP

17hTDE1 ∆ ∆ ∆ 28 S HEK 293-hTDE1 HEK 293 N C

N PGKhTDE1-GFP 1.4Kb hTDE1 105 18 S 75 0.5Kb βactin 50 βactine

Figure 1 Hypothetical structure of hTDE1 protein and Western and Northern blot analysis. (a) Roman numbers identify transmembrane segments of the protein. Arrows identify the approximate location of the deletion mutants ND17hTDE1 (1), ND126hTDE1 (2) and CD112hTDE1 (3). (b) Western blot analysis. Protein extracts were from transiently transfected Rat-1 cells expressing PGK-hTDE1, ND17hTDE1-GFP, ND126hTDE1-GFP and CD112hTDE1-GFP. GFP antibody was used to detect fusion proteins. Although expression levels of CD112hTDE1-GFP are low, the fusion protein migrates at the expected molecular size. (c) Northern blot analysis. RNA was purified from isolated Rat-1 independent clones transfected with either full-length or mutant hTDE1. hTDE1 sequences were revealed using a DNA probe corresponding to a 710 bp digestion fragment (BglII/AccI). Mouse b-actin mRNA was used as a loading control. (d) Reverse transcription-PCR analysis of control and hTDE1-transfected HEK-293 cells. RNA was purified from HEK-293 cells and from a mixed population of HEK cells expressing full-length hTDE1. Following cDNA generation, expression was examined by semiquantitative PCR using b-actin and hTDE1 specific primers (see Materials and methods). 20 ml of the RT-PCR product was resolved on a 1% TBE agarose gel.

with sequences encoding GFP, in order to localize the To test biological function, stable clones were isolated protein at the cellular level. Two N-terminal from transfected Rat-1 cells expressing PGK-hTDE1, (ND17hTDE1 and ND126hTDE1) and one C-terminal ND17hTDE1, ND126hTDE1 and CD112hTDE1. (CD112hTDE1) deletion mutants of hTDE1 were Northern blot analysis was used to confirm the similarly constructed and fused with GFP. The putative expression of transfected sequences (Figure 1c). Clones structure of hTDE1, as defined by Ensemble (The expressing high levels of full-length hTDE1 (PGK- Wellcome Trust Sanger Institute), as well as the location hTDE1-24 and PGK-hTDE1-34) or mutants of hTDE1, of deletion mutants, is detailed in Figure 1a. including clones expressing the ND17hTDE1 construct ND17hTDE1-GFP removes the first 17 amino acids of (ND17hTDE1-15 and ND17hTDE1-34), the ND126h the protein, corresponding to the major portion of a TDE1 construct (ND126hTDE1-7 and ND126hTDE1- putative signal peptide sequence. This mutation was 11) and the CD112hTDE1 construct (CD112hTDE1-6, generated in order to verify whether this signal peptide CD112hTDE1-8 and CD112hTDE1-24), were selected would be involved in translocation of the protein to the for further analysis. A stable pooled population of plasma membrane. ND126hTDE1-GFP lacks 126 amino HEK-293 cells expressing full-length hTDE1 was also acids corresponding to the first three transmembrane generated (Figure 1d). segments including the first 17 amino acids and thus To verify the subcellular localization of full-length eliminates the first two protein loops. In CD112hTDE1- and mutant hTDE1 protein, cells (Rat-1, HEK-293 and GFP, the last 112 amino acids were removed, resulting A-549 cells) were transiently transfected with appro- in a fusion lacking the last two protein loops of hTDE1. priate constructs. Results in all cell types were compar- Removal of this portion of the protein eliminates five of able, and for clarity A-549 cells, which present a large 12 previously identified putative phosphorylation sites cytoplasm, are presented (Figure 2). Whereas the (Bossolasco et al., 1999). Protein expression was pEGFP-N1 vector localized evenly in the nucleus and determined by Western blot analysis on transiently the cytoplasm (Figure 2a), the full-length fusion protein transfected Rat-1 cells, using the GFP antibody to PGK-hTDE1-GFP localized predominantly at the detect fusion proteins. All constructs expressed a fusion plasma membrane (Figure 2b), although some signal protein of the expected molecular size (Figure 1b). was apparent at the perinuclear level. For the

Oncogene Oncogenic properties of hTDE1 M Bossolasco et al 4552

Figure 2 hTDE1 protein subcellular localization. Fluorescence in A-549 cells transiently transfected with (a) pEGFP-N1 vector expressing the green ﬂuorescent protein, (b) PGK-hTDE1-GFP vector expressing full-length hTDE1-GFP, (c)ND17hTDE1-GFP and (d)CD112hTDE1-GFP.

ND17hTDE1-GFP fusion protein, although there was (data not shown). Finally, we could not detect a detectable signal at the plasma membrane, the majority biological effect associated with full-length hTDE1 using of fluorescence appeared to concentrate in dot-like spots either a wound closure motility assay or in a Matrigel- around the nucleus (Figure 2c), suggesting that the loss based invasion assay (data not shown). of the signal peptide affects the translocation of hTDE1 We next monitored the appearance of apoptotic cells to the plasma membrane. Similar fluorescent results using a Hoechst staining assay. A significantly greater were obtained with the ND126hTDE1-GFP fusion number of cells with condensed nuclei were observed in protein (data not shown). In contrast, the C-terminal Rat-1 cells as compared to hTDE1-expressing cells mutant of hTDE1 was virtually indistinguishable from (Figure 3b, left panel) at 24 h in both serum-starved the full-length construct, with the majority of fluores- and 40 mm etoposide treatment. Similarly, we tested cence localized to the plasma membrane (Figure 2d). caspase-3 activity, a major executor in the apoptotic signaling machinery, in cells grown in either complete The effects of hTDE1 expression on various oncogenic media (control), medium with no FBS or in complete parameters in vitro medium containing 40 mm etoposide. Rat-1 cells were A focus formation test was performed in order to compared to clones expressing full-length hTDE1 determine if hTDE1 expression affects the inhibition- (PGK-hTDE1-24 and clone PGK-hTDE1-34), and of-contact associated with normal confluent Rat-1 cells caspase-3 activity was monitored by cleavage of a (Figure 3a). Control and Rat-1 stable transfectants fluorescent substrate. Whereas a moderately lower of were left at confluence for 15 days. Full-length and caspase-3 activity was observed in hTDE1-expressing C-terminal deletion hTDE1 transfectants (Figure 3a) clones at 24 h (data not shown), a clear statistical demonstrated clear loss of contact inhibition with the difference was noted both in serum-starved and etopo- formation of robust foci (>100 per plate), in contrast to side conditions between Rat-1 cells and hTDE1- N-terminal deletion mutants (Figure 3a), which formed expressing clones (Figure 3b, right panel) at 48 h. only sparse and minute foci (o10 per plate). Controls Internucleosomal DNA fragmentation, a hallmark of failed to produce significant number of foci (o2 per apoptosis, was visualized by agarose gel electrophoresis. plate). After etoposide treatment, control Rat-1 and HEK-293 We measured the effect of hTDE1 on several cells as well as clones PGK-hTDE1-24 and PGK- parameters including cell proliferation, immortalization, hTDE1-34 or HEK-293 cells expressing hTDE1 dis- motility and cell invasion. We did not detect any played the typical DNA ladder associated with cells significant difference in the growth rate between clones undergoing apoptosis (Figure 4a). Hypodiploid expressing either the full-length or mutant forms of DNA, characteristic of cells in an advanced stage hTDE1 as compared to the parental Rat-1 cells (data of apoptosis, was also measured by detecting the not shown). Similarly, using a rat embryonic fibroblast presence of a sub-G1 peak following flow cytometry assay, we were unable to demonstrate that full-length analysis. The percentage of cells with hypodiploid hTDE1 significantly impacted on cell immortalization DNA was higher in both Rat-1 and HEK-293 control

Oncogene Oncogenic properties of hTDE1 M Bossolasco et al 4553 a

Rat-1 PGK-hTDE1-34

N∆17hTDE1-15 C∆112hTDE1-8 ** p< 0.01 b 30 700 **p< 0.01 25 600 *p<0.05 *p<0.05 500 20 400 poptotic cells. poptotic 15 ** p<0.01 300 ** p< 0.01 10 *p<0.05 200 ** p< 0.01

5 100 Percentage of a of Percentage 0 pm/mg.min. activity Caspase-3 0 24 hours 24 hours 48 hours 48 hours 48 hours 0% FBS 40µm etoposide control 0% FBS 40µm etoposide Figure 3 hTDE1 in vitro oncogenic potential. (a) Focus formation assay. Foci were photographed 15 days post-confluence in either control Rat-1 cells or stable Rat-1 transfected clones PGK-hTDE1-34, ND17hTDE1-15 and CD112hTDE1-8. Note that both the full- length and the C-terminal deletion mutant of hTDE1 are able to form foci, and that those formed by the C-terminal deletion appear highly compact. (b) Apoptosis assay. Apoptosis was assessed by Hoechst staining (left panel) and caspase-3 activity measurement (right panel). Hoechst staining experiments were repeated 3–4 times, and in each, 11–16 independent fields were analysed. Caspase-3 activity was measured as pmol/mg min and experiments were repeated twice in duplicate. For both caspase-3 and Hoechst experiments, white columns represent Rat-1 cells, stripped columns represent PGK-hTDE1-24-expressing cells and squared columns represent PGK-hTDE1-34-expressing cells. cells when compared to clone PGK-hTDE1-24, produced robust tumor formation, although with a clone PGK-hTDE1-34 or with a mixed population of longer latency. HEK-293 cells expressing hTDE1. Results obtained with both Rat-1 and HEK-293 cells transfected with empty vector were similar to control cells (Figure 4b) Discussion Together, these results suggest that hTDE1 expression protects cells from undergoing programmed cell The TDE1/TMS family includes several members, and death induced by either serum starvation or etoposide different homologues have been identified in several treatment. species. Table 1 summarizes available data for the mouse and human family members. Although a definitive biochemical function has yet to be defined The effects of hTDE1 expression in vivo: nude mice for the family members, as a group they have been xenograft assay classified as transmembrane proteins. Cellular location To determine the effect of hTDE1 expression on in vivo and protein function are closely related. Indeed, tumorigenesis, we injected either Rat-1 cells or clones aberrant subcellular localization of proteins is observed expressing either full-length hTDE1 (PGK-hTDE1-24 in cells of various diseases, such as in cancer, supporting and PGK-hTDE1-34) or the hTDE1-GFP fusion (PGK- the idea that cellular protein localization provides key hTDE1-GFP-16) subcutaneously in nude mice and information essential to assign functions to known monitored tumor formation (Figure 5). Rat-1 cells, as (Ferrigno and Silver, 1999; Faust and Montenarh, well as controls transfected with empty vector, either 2000; Pearce, 2000) and novel proteins (Simpson et al., failed to form tumors or presented with small masses. In 2000). Previous reports on mouse TMS-1, TMS-2 contrast, PGK-hTDE1-34 rapidly formed large tumors (Grossman et al., 2000) and AIGP1 (Aoki et al., 2002) and the majority of mice had to be killed at B50 days. have addressed their subcellular localization. Surpris- Clones PGK-hTDE1-24 and PGK-hTDE1-GFP-16 also ingly, despite being homologues, these proteins were

Oncogene Oncogenic properties of hTDE1 M Bossolasco et al 4554 a vector vector HEK-293-

PGK-hTDE1-24 Rat-1 PGK-hTDE1 PGK-hTDE1-34 PGK empty HEK-293 PGK empty

1 2 31 2 3 1 2 3 1 2 3 1 2 3 4 1 2 3 4 1 2 3 4

% of hypodiploid % of hypodiploid % of hypodiploid DNA content DNA content DNA content HEK-293 CTRL (lane 1) 4.79 b Rat-1 CTRL (lane 1) 0.63 PGK-hTDE1-34 CTRL (lane 1) 0.49 HEK-293 24h (lane 2) 3.37 Rat-1 CTRL 24h (lane 2) 7.13 PGK-hTDE1-34 24h (lane 2) 1.89 HEK-293 48h (lane 3) 17.44 Rat-1 CTRL 48h (lane 3) 22.11 PGK-hTDE1-34 48h (lane 3) 9.35 HEK-293 72h (lane 4) 40.38

PGK-hTDE1-24 CTRL (lane 1) 1.18 Rat-1 PGK CTRL (lane 1) 0.67 HEK-293-PGK-hTDE1 CTRL (lane 1) 3.68 PGK-hTDE1-24 24h (lane 2) 5.94 Rat-1 PGK 24h (lane 2) 4.66 HEK-293-PGK-hTDE1 24h (lane 2) 1.34 PGK-hTDE1-24 48h (lane 3) 10.7 Rat-1 PGK 48h (lane 3) 22.38 HEK-293-PGK-hTDE1 48h (lane 3) 2.41 HEK-293-PGK-hTDE1 72h (lane 4) 5.13

HEK-293-PGK CTRL (lane 1) 4.62 HEK-293-PGK 24h (lane 2) 5.62 HEK-293-PGK 48h (lane 3) 19.15 HEK-293-PGK 72h (lane 4) 31.31

Figure 4 DNA ladder and determination of hypodiploid DNA content. (a) Detection of apoptotic DNA ladder. Apoptosis was induced in Rat-1 control and transfected cells with 40 mM etoposide, and in HEK-293 control and transfected cells with 10 mM etoposide. DNA was extracted after 24 (lane 2), 48 (lane 3) and 72 (lane 4) h, and was compared to untreated cells (lane 1). A 20 mg portion of each sample was electrophoresed on a 1.6% agarose/ethidium bromide gel. (b) Flow cytometric analysis of sub-G1 DNA content. All samples previously shown to display the typical DNA ladder were subsequently used to analyse hypodiploid DNA content. Cells were collected and ﬁxed in 70% ethanol for 1 h. Fixed cells were treated with RNase and then stained with 40 mg/ml propidium iodide. Sub-G1 peaks are presented as the percentage of hypodiploid DNA to total DNA content, indicating the apoptotic percentage.

localized to different compartments. Based on results be explained by hTDE1 overexpression resulting in large obtained from an HA C-terminal epitope-tagged con- amounts of protein that may accumulate transiently struct of TMS-1, transiently transfected into HEK 293 during protein maturation. This accumulation may also cells, it was determined that this protein localizes to the generate improperly folded proteins and subsequently plasma membrane (Grossman et al., 2000). In contrast, interfere with the ultimate membrane localization of the AIGP1 was predominantly localized to the Golgi protein (Stevens and Argon, 1999). Alternatively, a large complex based on immunofluorescence staining of number of transmembrane proteins, including receptors, cultured cortical neurons using a AIGPI-specific anti- channels and transporters, are known to cycle and in body (Aoki et al., 2002). Data from Inuzuka et al. (2005) some cases this includes the presence of a retention suggest a co-localization of Serinc proteins with lipid apparatus allowing membrane proteins to be stocked in biosynthetic enzymes in endoplasmic reticulum mem- vesicles outside the constitutive exocytotic pathway branes. In the present report we have addressed the (Royle and Murrell-Lagnado, 2003). We also demon- localization of hTDE1, the human homologue of this strate that the N-terminal mutant of hTDE1, lacking family. Our results, based on a GFP C-terminal fusion, the putative signal peptide, displays increased peri- clearly demonstrate the protein localized to the plasma nuclear signal with a concomitant loss in plasma membrane of the cell. In addition, it is also possible to membrane fluorescence. An accumulation in the RER– visualize highly fluorescent dot-like structures in the Golgi complex is characteristic of proteins in which perinuclear area of the cell. This latter fluorescence may N-terminal sequences involved in the translocation of

Oncogene Oncogenic properties of hTDE1 M Bossolasco et al 4555 4 of Py-LT, a protein known to exhibit immortalizing et al 3.5 properties (Rassoulzadegan ., 1983). Results were not conclusive and therefore could not attribute an 3 ) 3 PGK-hTDE1-34 immortalizing function for hTDE1. Although cancer 2.5 PGK-hTDE1-24 cells generally are thought to cycle more rapidly, all 2 tumor cells do not necessarily proliferate faster than PGK-hTDE1-GFP-16 normal cells (Cancer Medicine, 6th edn. BC Decker Inc., 1.5 2003). Compared to the Rat-1 parental cell lines, tumor size (cm 1 expression of either the full-length hTDE1,orN-and 0.5 Rat-1 C-terminal deletion mutants of hTDE1, had no appre- PGK-GFP 0 ciable effect on proliferation rates. 20 40 60 80PBS 100 It has been suggested that the unrestrained increases Time (days post-injection) in the number of tumor cells is not only owing to an Figure 5 Nude mice xenografts. Nude mice injected with parental increase in cellular proliferation because of an acceler- Rat-1 cells and different transfected cells were monitored for tumor ated or uncontrolled cell cycle, but also may more development. Tumor size was measured on average every 10 days. Around 50 days, four of six PGK-hTDE1-34-expressing mice were closely be linked to an imbalance between the number of killed owing to large tumor size. Clones PGK-hTDE1-24 and dividing cells and the number of cells progressing to PGK-hTDE1-GFP-16 produced large tumors at around 80 days terminal differentiation or death (Corn and El-Deiry, post-injection. 2002; Green and Evan, 2002; Nowell, 2002). Indeed, we assessed hTDE1 ability to impact on apoptosis, both in response to serum starvation and etoposide exposure. In protein to the plasma membrane are missing (Stevens both cases, the number of hTDE1-expressing cells and Argon, 1999) supporting an important role of undergoing apoptosis were significantly reduced com- N-terminal sequences in the final destination of mature pared to the parental Rat-1 cells, and this was protein. In contrast, the hTDE1 C-terminal mutant, accompanied by a significant decrease in caspase-3 similar to the full-length protein, translocates to the activation in the hTDE1-expressing clones. Cellular plasma membrane, suggesting that these sequences are death was also measured by sub-G1 peak detection dispensable to proper protein subcellular localization. through flow cytometry, and apoptosis was confirmed Previous reports, including ours, suggest that aberrant through visualization of a characteristic DNA ladder on expression of TDE1/TMS family members may be agarose gel. Results obtained from Rat-1 cells were involved in the process of cellular transformation (Lebel comparable to those we observed in HEK-293 cells, and Mes-Masson, 1994; Bossolasco et al., 1999; Gross- suggesting that hTDE1 impact on apoptosis was not man et al., 2000; Player et al., 2003). In this report, we dependent on the cell type used for the analysis. These have directly assessed the transforming properties of results are in stark contrast to those associated with the hTDE1 following stable transfection in the Rat-1 cell effect of AIGP1 on apoptosis, where this TDE homo- line. Comparing parental and hTDE-1 Rat-1 cells, we logue was found to promote apoptosis in axotomized analysed both in vitro and in vivo characteristics mice neurons (Aoki et al., 2002). In addition, induction associated with oncogenic cell transformation. Although of apoptosis was also observed following transient hTDE1 expression appeared to have little effect on transfection of COS 7 cells with AIGP1 (Aoki et al., assays that measure cellular migration, invasion, im- 2002). Differences between the effect of AIGP1 and mortalization and proliferation, we did observe a clear hTDE1 may be explained by differences at the protein effect of hTDE1 expression on assays measuring level (the proteins share 77% homology) or differences response to apoptotic stimuli, contact inhibition and in the model systems used to test apoptosis, including tumor formation in nude mice. the observation that AIGP1 induces apoptosis following The fact that hTDE1 affects some, but not all, of the transfection whereas in this report stable clones were parameters associated with cell transformation is tested for their ability to resist specific inducers of perhaps not surprising as this phenomenon has been apoptosis. Differences related to protein sequence and observed for a number of viral and cellular oncogenes function are supported by the observation that the (Renan, 1993). Our experimental approach, the in vitro apoptotic rate was unchanged following transfection of wound closure assay (Ronot et al., 2000; Shigeta et al., COS-7 cells with TPO1, a rat TDE1/TMS family 2003), might reflect either a change in cell migration or a member (Aoki et al., 2002). Alternatively, the level of modification in intracellular adhesiveness. When com- expression of the transfected protein may play a role in paring Rat-1 cell wound closure with clone PGK- the final outcome although this is difficult to evaluate, hTDE1-24 and PGK-hTDE1-34, we failed to detect because in the case of AIGP1 expression was driven in any significant impact, suggesting that hTDE1 is not mouse and COS-7 cells using a CMV-based promoter, involved in cell motility or cell-to-cell adhesion. The whereas hTDE1 was under the control of the PGK Matrigel matrix used to assess cellular invasiveness promoter in Rat-1 and HEK-293 cells. Finally, a again provided no differences between Rat-1 cells and fundamental difference was also noted in subcellular clones PGK-hTDE1-24 and PGK-hTDE1-34. In order localization of these family members, with AIGP1 to address the effect of hTDE1 on senescence, the in vitro localized at the Golgi and hTDE1 at the plasma immortalizing activity of hTDE1 was compared to that membrane, which may in part account for the different

Oncogene Oncogenic properties of hTDE1 M Bossolasco et al 4556 observed activities of these proteins. Nonetheless, our Materials and methods results support an antiapoptotic function for hTDE1 that would impact positively on cellular transformation. Plasmid constructs Loss of density-dependent inhibition and contact The hTDE1 full-length sequence was amplified from a placental cDNA library plasmid (Bossolasco et al., 1999) inhibition results in the formation of disordered, multi- 0 0 layer array of cells that are observed as foci growing using the 5 sense primer 5 -CACGAGCTCAGCTGGCAGT TA-30 and the 30 antisense primer 50-CCGGATCCCACCT on a cellular monolayer (Temin and Rubin, 1958). It GAACTCCCGAC-30, thus eliminating the hTDE1 stop is noteworthy that whereas both full-length and codon. The 1495 bp PCR product and the pEGFP-NI vector C-terminal mutants of hTDE1 were able to form foci, (Clontech, Palo Alto, CA, USA) were digested with SacI and N-terminal mutants, which fail to locate to the plasma BamHI and ligated to generate the intermediate cloning vector membrane, rarely formed foci. Several membrane hTDE1-GFP. The hTDE1 sequences were extracted from molecules that are part of adhesion junctions, including hTDE1-GFP using XhoI/BamHI and re-cloned within the cadherins, are thought to play a role in density- XhoI/BamHI site of a PGK-GFP-hygro vector (Rodier et al., dependent contact inhibition (Grazia Lampugnani submitted) giving rise to PGK-hTDE1-GFP where hTDE1 is et al., 2003; Bazzoni and Dejana, 2004) and the presence fused in-frame at its C-terminus with GFP. PGK-hTDE1-GFP of hTDE1 may similarly impact on adhesion when is selectable in hygromycin and the hTDE1-GFP sequence is under the control of the PGK promoter. PGK-GFP-hygro was present at the plasma membrane. C-terminal mutant also digested with BsrG1 within GFP sequences, blunt-ended clones were also able to form foci, and these were with Klenow fragment and digested with SacI. This product generally more compact than those observed with full- was ligated to the SacI/SspI hTDE1 fragment (Bossolasco length hTDE1, and coupled to the observation that these et al., 1999) giving rise to PGK-hTDE1, which directs the mutants are less able to migrate through a Matrigel expression of full-length protein not fused to GFP sequences. matrix, hint at the possibility that hTDE1 affects cell N- and C-terminal deletions mutants were generated. adhesion. How and whether hTDE1 proteins are Following intermediate steps that allowed the introduction involved in cell-to-cell adhesion and migration is unclear of 50 cloning sites associated with a start codon, ND17hTDE1- and warrants further investigation. GFP was generated. This plasmid is similar to PGK-hTDE1- The ultimate characteristic of the oncogenic cell GFP except for the deletion of the first 17 amino acids of hTDE1. A similar strategy was used to clone ND126hTDE1- is its ability to form tumors in an in vivo environ- GFP where the first 126 amino acids of hTDE1 were deleted. ment. Therefore, the most striking observation in this Introduction of stop codon between TDE1-GFP sequences report is the ability of PGK-hTDE1-34 and PGK- also allowed the generation of ND17hTDE1 and ND126h hTDE1-24 cells to form tumors in a nude mouse TDE1 in which hTDE1 N-terminal mutants are expressed xenograft assay. Control mice either failed to develop without fusion to GFP. Following intermediate steps that tumors or developed minute masses at the site of allowed the removal of the last 112 amino acids of hTDE1, the injection. Mice injected with clone PGK-hTDE1-24 plasmid CD112hTDE1-GFP was generated in the XhoI/ and clone PGK-hTDE1-GFP-16 cells were also shown BamHI backbone of the PGK-GFP-hygro vector that allows to be able to consistently form tumors, although the the expression of a fused C-terminal deletion mutant of latency was somewhat longer. This result also further hTDE1 with GFP. Introduction of a stop codon between the hTDE1 and GFP sequences allowed the generation of demonstrates that the addition of the GFP moiety at the CD112hTDE1 in which the C-terminal deletion of hTDE1 is C-terminal end of the protein does not impact on its expressed without associated GFP sequences. tumorigenic potential, and further supports that the For cloning experiments, restriction and modifying ultimate cellular localization of the protein is at the enzymes were purchased from New England Biolabs (NEB, plasma membrane. Beverly, MA, USA) and final constructs were sequence In conclusion, although we have as yet to assign verified. a precise function to TDE1/TMS family members, the results presented here, coupled to our previous results in tumor-bearing transgenic mice and the Cell culture and transfection demonstration that hTDE1 overexpression can be All cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Wisent, Quebec, Canada) supplemented with 10% observed in tumor tissues (Bossolasco et al., 1999), fetal bovine serum (FBS; Wisent), 50 mg/ml gentamycin strongly support a role for hTDE1 in the oncogenic (Invitrogen, Carlsbad, CA, USA) and 2.5 mg/ml amphotericin process. However, the ultimate role of hTDE1 in B (Wisent). Cells were maintained at 37oC under an atmo- human cell transformation, and specifically, its possible sphere of 5% CO2. For transfection experiments, Rat-1, HEK- role in human carcinogenesis, needs to be further 293 and A-549 cells were transfected using Lipofectamine investigated. The present study also underlies the (Invitrogen). Cells were grown in 100 mm dishes and trans- ability of hTDE1 to interfere with programmed cell fected at 80% confluence with 10 mg of plasmid, 20 mlof death, allowing cells to escape control mechanisms lipofectamine and 30 ml of ReagentPlus, according to the established in multicellular organisms to eliminate manufacturer’s protocol. For Rat-1 stable transfections, cells malfunctioning cells such as transformed cells. The were passed at a 1:3 dilution 24 h after transfection and maintained in medium containing 75 mg/ml hygromycin B identification of cellular proteins able to interact with (Roche Applied Science, Basel, Switzerland). For HEK-293 TDE1/TMS family members would no doubt contribute stable transfections, cells were maintained in medium contain- to a better understanding of hTDE1 function and allow ing 75 mg/ml hygromycin B (Roche Applied Science, Basel, us to better understand how these proteins impact on Switzerland) for several months resulting in a mixed popula- transformation. tion of cells expressing hTDE1.

Oncogene Oncogenic properties of hTDE1 M Bossolasco et al 4557 Western blot analysis CACGGCATCGTCA-30 for the 50 sense primer and Cells were scraped, resuspended in lysis buffer (50 mM Tris- 50-TAGCACAGCCTGGATAGCAAC-30 for the 30 antisense HCl pH 6.8, 100 mM dithiothreitol (DTT), 10% glycerol) and primer. PCR reaction was performed in a final 50 ml containing sonicated for 30 s (Scientific Instruments, KS212 model, 5 mlof10Â PCR buffer, 2 mM MgCl2, 200 mM dNTPs, 0.2 mM Ringoes, NJ, USA). Sonication was followed by 10 min of sense and antisense primers and 2 U of FastStart Taq DNA centrifugation at 13 000 r.p.m. at 41C. The supernatant Polymerase (Roche, Applied Science, Basel, Switzerland). containing the proteins was transferred to another tube and PCR was performed for 25 cycles at 941C for 30 s, 581C for protein concentration was determined with the Bradford 30 s and 721C for 1 min. method. From 20 to 50 mg of proteins were loaded on a 9% SDS—polyacrylamide gel electrophoresis acrylamide gel. After migration, proteins were subsequently electrotransferred on a Foci formation assay nitrocellulose membrane for 1 h at 41C at 200 mA (Bio-Rad Cells were seeded in 100 mm Petri dishes (SARSTEDT, Trans-Blot Cell, Hercules, CA, USA). For protein detection, Nu¨ mbrecht, Germany) grown to confluence and re-fed with membranes were blocked overnight with 5% skin milk and fresh media twice a week. Confluence was maintained for 15 0.1% Tween 20. Primary antibodies for GFP (JL-8, Clontech) days, following which cells were washed with PBS, rinsed in or b-actin (AC-14, Abcam, Cambridge, UK) detection were 50% methanol/50% PBS and subsequently incubated for diluted 1:8000 and 1:10 000, respectively, in 5% skin milk and 10 min in fresh anhydrous methanol. Cells were stained with 0.1% Tween 20 buffer. Membranes were incubated for 45 min Giemsa Stain Stock Solution (VWR Scientific Products, West with the b-actin primary antibody or for 3 h with the GFP Chester, PA, USA), diluted 1:5 in deionized water and primary antibody and were subsequently washed with incubated overnight. In the morning, dye was removed and phosphate-buffered saline (PBS) containing 2% Tween 20. Petri dishes were rinsed with deionized water, dried, photo- Secondary antibody consisting of anti-mouse IgG-HPR graphed and focus formation was scored. conjugated (Santa Cruz Antibodies, Santa Cruz, CA, USA) was used at a 1:12 000 dilution rate. The ECL (Amersham Hoechst staining Biosciences, Buckinghamshire, UK) kit was used to detect the 3 Â 104 cells were seeded on microscope coverslip glass (no. 12, signal on the membranes. Fisherbrand, Fisher Scientific, Leicestershire, UK) in 24-well plates (Falcon, Becton Dickinson, Palo Alto, CA, USA). Cells were allowed to settle for 24 h. Apoptosis was then induced by RNA purification, Northern blots analysis and reverse replacing medium with either serum-free media or by media transcription-PCR. supplemented with 40 mm etoposide (Sigma-Aldrich, St Louis, RNA was isolated with TRIsol reagent (Gibco/BRL, Life MO, USA). Following 24 and 48 h treatments, 1 ml of a 10 mg/ Technologies Inc., Grand Island, NY, USA). Purified RNA ml Hoechst stain solution (Bisbenzimide H33258, Sigma- was quantified spectrophotometrically, and the quality was Aldrich, St Louis, MO, USA) was added to each well. Cells determined by ethidium bromide staining after agarose were incubated at 371C for 1 h. Microscope coverslips were electrophoresis. For each sample, 20 mg of total RNA was then removed and placed on a microscope slide (Corning glass electrophoresed through a 1% agarose gel containing 0.36 M Works, Corning, NY, USA). Five fields per coverslip were formaldehyde. Transfer was performed on a GeneScreen Plus examined under a fluorescent microscope (excitation wave- hybridization transfer membrane (NEN Life Science Products, length 365 nm, emission wavelength 465 nm, 40 Â 0.60 magni- Boston, MA, USA). The gel was washed two times in 0.05 M fication, NIKON eclipse TE300, Tokyo, Japan) and apoptotic NaOH in 1 Â SSC for two 10 min periods, then soaked in and living cells were counted. 10 Â SSC for two 20 min periods in 10 Â SSC. Capillary transfer was performed overnight in 10 Â SSC. The membrane was UV-crosslinked (UV-Stratalinker 2400, Stratagene, Cedar Caspase-3 activation assay Creek, TX, USA) and hybridized overnight at 641C, in 1% Caspase-3 activity was measured using the Ac-DEVD-AFC SDS, 1 M NaCl and 10% dextran sulfate supplemented with substrate and assayed according to the manufacturer’s 100 mg/ml of salmon sperm DNA and labeled probe. All protocol. Briefly, cells were seeded at 70% confluence. After probes were labeled overnight at room temperature using a 24 h, apoptosis was induced by replacing media with either random primer labeling system (Redprimet II, Amersham serum-free media, or complete media supplemented with 40 mm Biosciences, Buckinghamshire, UK). A probe corresponding etoposide. Both adherent and floating cells were harvested at to the mouse b-actin RNA was used as internal control. After 24 and 48 h post-treatment, placed in a 15 ml tube and cells overnight hybridization, membranes were rinsed twice for were pelleted at 1500 r.p.m. for 10 min. Pellets were resus- 30 min with 2 Â SSC and 0.1% SDS. Membranes were exposed pended in cold PBS, transferred to Eppendorf tubes and to Kodak BioMax MS films with an intensifying screen at re-pelletted at 1500 r.p.m. for 10 min. Supernatants were À801C overnight. discarded, and pellets were suspended in 70 ml of lysis buffer Five micrograms of total RNA was converted to single- (Invitrogen, Carlsbad, CA, USA). Cells were placed on ice for stranded cDNA by using murine reverse transcriptase in a 10 min, debris was pelletted and supernatants were stored at volume of 30 ml containing 6 mlof5Â first strand buffer À801C. Protein concentration was measured at 579 nm by the (250 mM Tris-HCl (pH 8.3), 375 mM KCl, 15 mM MgCl2), BCA protein quantification kit (Pierce, Rockford, IL, USA). 10 mM DTT, 0.6 mM dNTPs, 33 U of RNase guard, 0.5 mgof For each assay, 35–50 mg of protein was incubated for 30 min oligo-dT primers and 400 U of M-MLV Reverse Transcriptase with 50 mM Ac-DEVD-AFC at 301C. Fluorescence was (Invitrogen, Carlsbad, CA, USA). The reaction was performed analysed using a FluoStar-Optima microplate reader (BMG at 421C for 60 min. For the subsequent PCR, reaction, 5 ml lab Technologies, Offenburg, Germany) in fluorescence mode of the cDNA reaction product was amplified by using using an excitation filter of 380 nm (10 nm bandpass) and an 50-GGGGCTGTGCTGGGTGTCTT-30 for the 50 sense pri- emission filter of 505 nm (10 nm bandpass). The reaction was mer and 50-CGGATCCGTGTCCTTGGCACTCAGAGGT allowed to proceed for 30 min with a reading every minute. -30 for the 30 antisense primer. Amplification of the b-actin Caspase-3 activities were obtained by calculating the slope of as a positive control was performed using 50-CATCGAG the reaction over 30 min for each assay.

Oncogene Oncogenic properties of hTDE1 M Bossolasco et al 4558 DNA ladder assay USA) and 40 mg/ml propidium iodide (Molecular Probes, Cells were seeded at 70% confluence. After 24 h, apoptosis was Invitrogen, Carlsbad, CA, USA) and incubated at room induced by replacing media with complete media supplemen- temperature for 30 min. Samples were analysed using a Becton ted with etoposide. A final concentration of 40 mM etoposide Dickinson Fac Scan Flow Cytometer (Becton Dickinson, CA, was used on Rat-1 control and transfected cells, whereas a USA) as described by Ormerod et al. (1992). Cells with sub-G1 10 mM etoposide final concentration was used on HEK-293 (hypodiploid DNA) propidium iodide incorporation were control and transfected cells. Both adherent and floating cells considered apoptotic. were harvested at 24 and 48 h post-treatment, placed in a 15 ml tube and pelletted at 1500 r.p.m. for 15 min. Pellets were resuspended in 1 ml final volume of 1 M Tris-HCl pH 8, 0.5 M Tumorigenesis assay EDTA pH 8 and 10% SDS containing 100 mg/ml of Proteinase Four- to seven-weeks-old CD1 Br Nu/Nu mice (Charles K (Fermentas, Burlington, Canada) and samples were Rivers, St-Constant, Qc) were used for tumor growth incubated at 371C with shaking overnight. The day after, xenografts experiment. Stable clones expressing PGK-GFP samples were chilled on ice for 10 min. A 400 ml volume of 5 M (empty vector), PGK-hTDE1-GFP and PGK-hTDE1 were NaCl was added and tubes were inverted several times to mix. tested. Rat-1 (non-transfected) cells and PBS alone were also Samples were spun for 15 min at 3400 r.p.m. The supernatant used as negative controls. Mouse flanks were injected 6 was transferred to a new tube and DNA was precipitated by subcutaneously with 2.5 Â 10 cells for each clone (six mice adding 2 volumes of 100% ethanol. The pellet was resus- per clone) and tumor size was measured every 2 weeks for 4 pended in 20 ml of water and incubated at 371C for 30 min with months. Mice were killed early when tumor volumes reached 3 RNase A (usb, Cleveland, OH USA) at a final concentration >2 cm . of 20 mg/ml. Electrophoresis was performed in 1.6% agarose gel in TBE buffer, after which DNA was visualized by Statistical analysis ethidium bromide staining. DNA fragmentation analysis was All statistical analyses were performed using the InStat3 assessed by resolving 10–20 mg of DNA for each sample. GraphPad Software (GraphPad Software Inc. (c) 1990–2001). Determination of hypodiploid DNA content To measure the loss of DNA owing to cell death, cells were Acknowledgements seeded at 70% confluence. After 24 h, apoptosis was induced by replacing media with complete media supplemented with We thank Jason Madore for technical assistance, and Wissal etoposide. A final concentration of 40 mM etoposide was used El-Assaad and Abdellah Belmaaza for stimulating discussions. on Rat-1 control and transfected cells, whereas a 10 mM This research was funded in part by a Canadian Institute of etoposide final concentration was used on HEK-293 control Health Research (CIHR) grant and a Cancer Research Society and transfected cells. Both adherent and floating cells were grant to A-MM-M MB has been supported by a CIHR harvested at 24 and 48 h post-treatment, placed in a 15 ml tube studentship as well as awards from the Canderel and Marc- and pelletted at 1500 r.p.m. for 15 min. Cells were fixed by Bourgie funds of the Institut du cancer de Montre´ al. FV was resuspending pellets in ice-cold 70% ethanol for 1 h. One hour the recipient of a Canderel studentship award. A-MM-M is the before performing flow cytometry analysis, ethanol-fixed cells recipient of a Fonds de la recherche en sante´ au Que´ bec were washed twice with PBS. Cells were then resuspended in (FRSQ) Chercheur National Fellowship. RB is the recipient of PBS containing 0.5 mg/ml RNase A (usb, Cleveland, OH, an FRSQ Senior Fellowship.

References

Aoki S, Su Q, Li H, Nishikawa K, Ayukawa K, Hara Y et al. Lebel M, Mes-Masson AM. (1994). DNA Sequence 5: (2002). J Neurosci 22:10751–10760. 31–39. Bazzoni G, Dejana E. (2004). Physiol Rev 84:869–901. Nowell PC. (2002). Semin Cancer Biol 12:261–266. Bossolasco M, Lebel M, Lemieux N, Mes-Masson AM. Ormerod MG, Collins MK, Rodriguez-Tarduchy G, (1999). Mol Carcinogen 26:189–200. Robertson D. (1992). J Immunol Methods 153:57–65. Chalifour LE, Gomes ML, Wang NS, Mes-Masson AM. Pearce DA. (2000). J Neurosci Res 59:19–23. (1990). Oncogene 5:1719–1726. Player A, Gillespie J, Fujii T, Fukuoka J, Dracheva T, Chalifour LE, Mes-Masson AM, Gomes ML, Wang NS. Meerzaman D et al. (2003). Int J Cancer 107:238–243. (1992). Mol Carcinogen 5:178–189. Rassoulzadegan M, Naghashfar Z, Cowie A, Carr A, Grisoni Corn PG, El-Deiry WS. (2002). BioEssays 24:83–90. M, Kamen R et al. (1983). Proc Natl Acad Sci USA 80: Faust M, Montenarh M. (2000). Cell Tissue Res 301:329–340. 4354–4358. Ferrigno P, Silver PA. (1999). Methods Cell Biol 58:107–122. Renan MJ. (1993). Mol Carcinogen 7:139–146. Grazia Lampugnani M, Zanetti A, Corada M, Takahashi T, Ronot X, Doisy A, Tracqui P. (2000). Cytometry 41: Balconi G, Breviario F et al. (2003). J Cell Biol 161:793–804. 19–30. Green DR, Evan GI. (2002). Cancer Cell 1:19–30. Royle SJ, Murrell-Lagnado RD. (2003). BioEssays 25:39–46. Grossman TR, Luque JM, Nelson N. (2000). J Exp Biol Shigeta M, Sanzen N, Ozawa M, Gu J, Hasegawa H, 203(Part 3):447–457. Sekiguchi K. (2003). J Cell Biol 163:165–176. Inuzuka M, Hayakawa M, Ingi I. (2005). J Biol Chem 280: Simpson JC, Wellenreuther R, Poustka A, Pepperkok R, 35776–35783. Wiemann S. (2000). EMBO Rep 1:287–292. Krueger WH, Gonye GE, Madison DL, Murray KE, Kumar Stevens FJ, Argon Y. (1999). Semin Cell Dev Biol 10: M, Spoerel N et al. (1997). J Neurochem 69:1343–1355. 443–454. Kufe DW, Pollock RE, Weichselbaum RR, Bast Jr RC, Temin HM, Rubin H. (1958). Virology 6:669–688. Gansler TS, Holland JF et al. (eds). (2003). Cancer Medicine, Xu J, Ji C, Wang L, Cao Y, Dai J, Ye X et al. (2003). Mol Biol 6th edn. BC Decker Inc.:Hamilton (Canada). Rep 30:47–52.

Oncogene