Oncogene (2003) 22, 1817–1821 & 2003 Nature Publishing Group All rights reserved 0950-9232/03 $25.00 www.nature.com/onc fau and its -like domain (FUBI) transforms human osteogenic sarcoma (HOS) cells to anchorage-independence

Toby G Rossman1,2, Melissa A Visalli1 and Elena V Komissarova1

1The Nelson Institute of Environmental Medicine, New York University School of Medicine, 57 Old Forge Road, Tuxedo, NY 10987, USA; 2The NYU Cancer Institute, New York University School of Medicine, 57 Old Forge Road, Tuxedo, NY 10987, USA

Arsenite is the most likelycarcinogenic form of arsenic in sarcoma virus (FBR-MuSV), originally isolated from a the environment. Previously, expression cloning for radiation-induced mouse osteosarcoma (Finkel et al., cDNAs whose overexpression confers arsenite-resistance 1976).The complete mouse fau cDNA sequence is in Chinese hamster V79 cells identified two : fau and inversely inserted as the fox sequence in FBR-MuSV. a novel , asr2. The fau gene encodes a ubiquitin-like The expression of the fox sequence as antisense of the (here called FUBI) fused to the ribosomal S30 fau gene increases the transforming capability of FBR- protein. Since the expression of the fox sequence MuSV, presumably by inactivating fau expression (antisense to fau) increased the tumorigenicityof a mouse (Michiels et al., 1993). Because of this, it has been sarcoma virus, it was proposed that fau might be a tumor suggested that the fau gene functions as a tumor suppressor gene. We intended to test its abilityto block suppressor gene. arsenite-induced transformation of human osteogenic We recently found that extremely low concentrations sarcoma (HOS) cells to anchorage-independence. Instead, of arsenite are able to transform nontumorigenic human we found that overexpressing fau itself was able to osteogenic sarcoma (HOS) cells to anchorage-indepen- transform HOS cells. When the two domains were dence (Rossman et al., 2001). These cells exhibit a flat expressed separately, only FUBI was transforming and morphology and can be transformed to anchorage- onlythe S30 domain conferred arsenite resistance. An independence by both viral and other chemical carcino- incidental finding was the transforming activityof the gens (Rhim et al., 1975; Rani and Kumar, 1992; Lin and selectable marker, hyg. FUBI belongs to the ubiquitin-like Costa, 1994; Miller et al., 1998). If fau is a tumor protein group that is capable of forming conjugates to suppressor gene, its overexpression might be expected to other , although none have so far been identified. inhibit transformation of these cells by arsenite and Alternatively, FUBI may act as a substitute or inhibitor of other agents. ubiquitin, to which it is most closelyrelated, or to close Figure 1 shows the Northern analysis of the fau ubiquitin-like relatives UCRP, FAT10, and/or Nedd8. expression in HOS cells and the various transfectant Oncogene (2003) 22, 1817–1821. doi:10.1038/sj.onc.1206283 lines constructed.HOS cells and HOS cells transfected with pZeo vector alone show a similar level of fau Keywords: transformation; fau; ubiquitin-like protein; expression (lanes 1 and 2).HOS cells transfected with arsenic; hygromycin; pZeo-fau, containing full-length Chinese hamster fau cDNA, show increased fau expression, as expected (lane 3).HOS cells transfected with pZeo-FUBI and pZeo- S30, which are expected to express the ubiquitin-like Previously, we identified two genes whose overexpres- domain and the S30 domain of Chinese hamster fau, sion confers resistance to the human carcinogen arsenite respectively, show expression of faster migrating bands in Chinese hamster V79 cells: the novel gene asr2 and in addition to the normal fau band (lanes 4 and 5), the fau gene (Rossman and Wang, 1999).The fau gene consistent with expression of the domains of interest. encodes a 133 amino-acid protein consisting of a 74 Overexpression of fau was previously found to confer amino-acid ubiquitin-like protein (here called FUBI) arsenite resistance to Chinese hamster V79 cells (Ross- fused to a 59 amino-acid ribosomal protein S30.S30 is man and Wang, 1999).Figure 2 shows that full-length released by post-translational cleavage of Fau and is fau expression also confers arsenite-resistance to HOS then incorporated into the small ribosomal subunit.The cells (LC50 ¼ 2.9 mm for the fau transfectant compared fau gene was identified as the cellular homologue of with 1.1 mm for HOS).A more modest increase in the fox sequence in the Finkel–Biskis–Reilly murine arsenite-resistance is seen in cells expressing only the S30 domain (LC50 ¼ 1.7 mm).In contrast, expression of only *Correspondence: TG Rossman, The Nelson Institute of Environ- the FUBI domain resulted in increased sensitivity to mental Medicine, New York University School of Medicine, 57 Old arsenite (LC ¼ 0.4 mm).Cells transfected with vector Forge Road, Tuxedo, NY 10987, USA; 50 E-mail: [email protected] alone show arsenite sensitivity identical to HOS (data Received 12 August 2002; revised 2 November 2002; accepted 29 not shown).Colony size decreases with arsenite toxicity November 2002 in all cells. Transformation by fau and FUBI TG Rossman et al 1818

Figure 1 Northern analysis of fau, S30 and FUBI expression in HOS transfectants.Lane 1: HOS cells; lane 2: HOS-pZeo (HOS transfected with pZeo vector alone); lane 3: HOS-pZeo-fau; lane 4: HOS-pZeo-FUBI; lane 5: HOS-pZeo-S30.pZeo is a modification of pCEP-VP (Rossman and Wang, 1999).The hyg gene and the HSV-1 TK promoter were cut at the Nar1 and RsrII sites.A zeocin cassette, containing the SV40 promoter, EM-7 promoter, zeocin resistance gene (zeo), and SV40 polyadenylation sequences, was PCR amplified from pcDNA4/HisMaxA (Invitrogen, San Diego, CA, USA) with Nar1/RsrII sites and cloned into pCEP-VP.The fau cDNA in pCEP-VP was digested with Nar1 and RsrII and inserted into pZeo.The FUBI domain (a 217 fragment) and the S30 domain (a 205 base pair fragment) of the fau gene were separately PCR amplified using the following primers: FUBI: 50GTCGGTACCATGCAGCTCTTTGT and 30CCTACATGCGA CCAGCTACT.S30: 5 0GGTACCATGTTGGGAGGTAAAGTT and 30GAAGGACTTACGCGGCGCATTA.The PCR products were cloned into pZeo and sequenced to ensure accuracy (Seq- Wright, Houston, TX, USA).Plasmid DNA was isolated by acidic phenol extraction (Wang and Rossman, 1994) and transfected into HOS (HOS TE85, American Type Culture Collection) by the Figure 2 Sensitivity (clonal survival) to arsenite of HOS cells and calcium phosphate procedure (Sambrook et al., 1989). Transformed HOS transfectants.HOS ( J); HOS(pZeo-fau) (), HOS(pZeo- cells were selected in medium containing 100 mg/ml Zeocin (Invitro- FUBI) (&); HOS(pZeo-S30) (’).HOS cells were cultured in a- gen, CA, USA).For Northern blot, total RNA was isolated using MEM (Gibco Life Technologies) containing 10% fetal bovine TRI reagent (Molecular Research Center, Inc., Cincinnati, OH, serum (Gibco) in 5% CO2 at 371C.For clonal survival assays, cells USA).Poly A + RNA was then isolated using oligo(dT)-cellulose were seeded at a density of 500 cells/60 mm dish with arsenite columns (Molecular Research Center, Inc.). The RNA was added 18 h later.Visible colonies develop in 7 days.Colonies were electrophoresed in a 1.5% agarose gel containing 3.5% formalde- fixed with methanol and stained with 0.5% crystal violet/50% hyde, transferred onto a nylon membrane and hybridized to 32P- methanol.All assays were performed in triplicate.Error bars are labelled fau and human G3PDH (housekeeping gene) probes (High for standard error of the mean Prime kit, Roche).The mRNA was detected using Genestorm phosphoimager (Molecular Dynamics) Table 1 Transformation of HOS cells by the hygromycin-resistance gene in pCEP-VP, and by fau and its FUBI domain À5 HOS cells have a low (1.2  10 ) spontaneous Cells Transformation frequency  10À4 transformation frequency.The pCEP-VP vector, which a contains the hygromycin-resistance hyg gene as select- HOS 0.12 HOS-pCEP-VP 263.8a able marker, was found to transform HOS cells to HOS-pZeo 0a anchorage-independence (Table 1).Aminoglycoside HOS-pZeo-fau 23.6a phosphoryltransferases such as hyg have structural HOS-pZeo-FUBI 284.8b similarities to eucaryotic protein kinases, are inhibited HOS-pZeo-S30 0.12a by similar agents, and sometimes demonstrate measur- Individual clones of HOS or of the transfectants (selected in able protein kinase activity (Wright et al., 1998). In hygromycin for HOS-pCEP-VP or in zeocin for the other transfec- contrast, no transformation occurred when the hyg gene tants) were isolated, expanded, and assayed for anchorage-indepen- was replaced with zeocin-resistance as a selectable dent growth marker in pZeo.Expression of full-length fau causes aAverage of four clones b significant transformation (Table 1, Figure 3).When its Average of five clones two domains are expressed separately, it is clear that the transforming activity resides in the FUBI domain (Table 1).Four independently isolated transfectant 1998).Although fau had been assumed to act as a tumor clones expressing full-length fau and five independently suppressor gene, it appears that overexpressed fau may isolated clones expressing its FUBI domain showed itself have oncogenic activity.Transcription factor significant transforming ability, whereas none of the E2F1, which regulates cyclins A and E as well as other four independently isolated clones expressing S30 did. genes involved in cell cycle control, has also been shown In cells of the mesenchymal lineage, growth in soft to possess both oncogenic and tumor suppressor activity agar correlates better than other criteria with the ability in different situations (Johnson, 2000). of a cell to grow as a tumor when injected into a suitable Most ribosomal proteins are unprocessed primary host (Shin et al., 1975). Anchorage-independent activa- products of their mRNAs.In addition to S30, there are tion of G1 cyclin-dependent kinases mediates the two other mammalian ribosomal proteins made as anchorage-independent growth phenotype (Yang et al., ubiquitin-like fusion proteins: rpS27a and rpL40 (Finley

Oncogene Transformation by fau and FUBI TG Rossman et al 1819 antimicrobial protein, ubiquicidin, appearing in an interferon-g-treated macrophage cell line (Hiemstra et al., 1999), but it is not known which domains are required for this activity. Modification of proteins by ubiquitin is critical for targeting those proteins for degradation by the ATP- dependent 26S proteosomes.Degradation of proteins to control their abundance is of importance in many cellular events (Hochstrasser, 1996).In addition to signaling proteins for degradation, ubiquitination some- times functions independently of proteosome action in cellular regulation without degradation (Hochstrasser, 1996).Arsenite has been shown to inhibit two steps in the ubiquitin pathway to protein degradation (Klem- perer and Pickart, 1989), but the concentration needed was high. A number of ubiquitin-like (ubiL) proteins with Figure 3 Transformed clones of HOS(pZeo-fau) cells growing in soft agar.Colony formation in soft agar was carried out according diverse functions have been identified (Jentsch and to the method of Rani and Kumar (1992).All assays were carried Pyrowolakis, 2000; Yeh et al., 2000). One group, out in triplicate.Single cells and small clumps in the background do ubiquitin domain proteins, consists of proteins with N- not form large colonies terminal ubiquitin-like domains that do not bind to other proteins.Proteins in the second group, the ubiL et al., 1989; Wong et al., 1993; Baker et al., 1996; modifiers, are able to form conjugates with other Redman and Burris, 1996).The function and fate of proteins.These include small ubiquitin-related FUBI is poorly understood.Although FUBI is clearly modifier-1 (SUMO-1), neural precursor cell expressed diverged from ubiquitin, it is thought to perform a developmentally downregulated-8 (Nedd8), ubiquitin similar chaperone function, facilitating incorporation of cross-reactive protein (UCRP, also called ISG15), which the 30S protein into ribosomes (Hochstrasser, 2000).We has two ubiL domains, and FAT10, formerly called have demonstrated that overexpression of the S30 diubiquitin, which also has two ubiL domains.As domain of fau confers arsenite resistance, although full shown in Figure 4, human FUBI has the highest arsenite resistance might require expression of complete homology to ubiquitin (59%) followed by UCRP II fau (Figure 2). (C-terminal domain; 57% in the aligned region), and Besides its role in ribosome biosynthesis, Fau has FAT10 I (N-terminal domain; 52% in the aligned been identified as a subunit of the monoclonal non- region).Homology between SUMO-1 and FUBI is specific immune suppressor factor (MNSF) (Nakamura much lower (26%) and is not shown.Human FUBI is et al., 1995), and its FUBI domain is responsible for this 97% homologous to Chinese hamster FUBI (not activity (Nakamura et al., 1996). MNSF is a lymphokine shown). excreted by murine T cell hybridomas (Nakamura and UCRP is an interferon-inducible cytokine, released Tanigawa, 1999).Fau was also identified as an from a number of different cells, which induces

Figure 4 Amino-acid alignment of human FUBI with ubiquitin and closely related ubiquitin-like proteins.Identities are shown in dark gray and similarities in light gray.FUBI: N-terminal ubiquitin-like domain (1–74 aa) of human fau (Accession no.P35544).Ubiq: human ubiquitin (Accession no.P02248).FAT10-I: N-terminal ubiquitin-like domain (1–83 aa) of human FAT10 protein (Accession no.AAD52982).FAT10-II: C-terminal ubiquitin-like domain (84–165 aa) of human FAT10 protein.Nedd8: human Nedd8 protein (Accession no.Q15843).UCRP-I: N-terminal ubiquitin-like domain (1–80 aa) of human UCRP (Accession no.P05161).UCRP-II: C- terminal ubiquitin-like domain (81–165 aa) of human UCRP

Oncogene Transformation by fau and FUBI TG Rossman et al 1820 proliferation of NK cells (D’Cunha et al., 1996; Potter ligases (Tateishi et al., 2001) that ubiquinate such et al., 1999). After proteolytic processing to remove a important proteins as the cell cycle inhibitor and tumor C-terminal tail, it is released either as a monomer suppressor p27(Kip1) (Podust et al., 2000) and IkB, the or conjugated to other proteins (Loeb and Haas, inhibitor of the transcription factor NF-kB, resulting in 1992).UBE1L, the activating enzyme for UCRP, NF-kB activation (Read et al., 2000). Arsenite (at toxic was found lacking in lung tumor cell lines, suggesting concentrations) also induces NF-kB activation (Huang that decreased UCRP conjugation may contribute et al., 2001). NEDD8 modification also regulates cell to carcinogenesis (McLaughlin et al., 2000). UCRP cycle progression and the cytoskeletin (Kurz et al., was identified in a screen for p53 and radiation- 2002). inducible genes (Polyak et al., 1997). p53 was found to This second class of ubiL-containing proteins con- be essential for induction of UCRP by dsRNA but not tains (or is processed to contain) a free C-terminal gly by interferon (Hummer et al., 2000). Arsenite interferes gly needed for protein conjugation with the e-amino with p53 functioning (Vogt and Rossman, 2001).UCRP group of lysine residues on other proteins.FUBI also is also inducible by genotoxic agents (Hermeking et al., has a C-terminal gly–gly residue (Figure 4) suggesting 1997), perhaps as a result of faulty transcription on a that FUBI might also take part in protein modification, damaged template resulting in dsRNA production although this has not yet been established.FUBI-lation (Tornaletti and Hanawalt, 1999) as well as p53 of proteins, should it exist, might represent a unique activation. type of protein modification.Alternatively, FUBI may FAT10, is constitutively expressed in thymus and be able to substitute for or inhibit the ubiL proteins spleen, but can be induced by TNF a and interferon g FAT10, Nedd8 or UCRP, or ubiquitin itself, for which (but not a) in many tissues (Liu et al., 1999). Stable it has high homology (Figure 4).The possibility that expression of FAT10 and its conjugation to other FUBI might interfere with other ubiquitin-containing proteins leads to apoptosis (Raasi et al., 2001). FAT10 proteins, including ubiquitin domain proteins, cannot be also binds noncovalently to MAD2 (Liu et al., 1999), ruled out. a protein that binds to unattached kinetochores in mitotic cells and is thought to cause delay of ana phase until kinetochores are attached to the spindle (Chen et al., 1996; Li and Benezra, 1996). Arsenite treatment has been shown to perturb spindle dynamics Acknowledgements We thank Eleanor Cordisco for her expert help in document resulting in aneuploidy (Ramirez et al., 1997; Huang preparation.This work was supported by United States Public and Lee, 1998). Health Service Grants R01 ES09252 and P42 ES10344, and is NEDD8 conjugates proteins that are found mostly in part of NYU’s Nelson Institute of Environmental Medicine the nucleus, such as cullins (Hori et al., 1999). Cullins Center programs supported by Grants ES00260 from NIEHS are essential components of SCF and VEC ubiquitin and CA16087 from NCI.

References

Baker RT, Williamson NA and Wettenhall REH.(1996). J. Jentsch S and Pyrowolakis G.(2000). Cell Biol., 10, 335–342. Biol. Chem., 271, 13549–13555. Johnson DG.(2000). Mol. Carcinogenesis, 27, 151–157. Chen R-H, Waters JC, Salmon ED and Murray AW.(1996). Klemperer NS and Pickart CM.(1989). J. Biol. Chem., 264, Science, 274, 242–246. 19245–19252. D’Cunha J, Knight E, Haas AL, Truitt RL and Borden EC. Kurz T, Pintard L, Willis JH, Hamill DR, Gonczy P, Peter M (1996). Proc. Natl. Acad. Sci. USA, 93, 211–215. and Bowerman B.(2002). Science, 295, 1294–1298. Finkel MP, Reilly Jr CA and Biskis BO.(1976). Recent Results Li Y and Benezra R.(1996). Science, 274, 246–248. Cancer Res., 54, 92–103. Lin X and Costa M.(1994). Environ. Health Perspect., 102, Finley D, Bartel B and Varshavsky A.(1989). Nature, 338, 289–292. 394–401. Liu Y-C, Pan J, Zhang C, Fan W, Collinge M, Bender JR and Hermeking H, Lengauer C, Polyak K, He TC, Zhang L, Weissman SM.(1999). Proc. Natl. Acad. Sci. USA, 96, 4313– Thiagalingam S, Kinzler KW and Vogelstein B.(1997). Mol. 4318. Cell, 1, 3–11. Loeb KR and Haas AL.(1992). J. Biol. Chem., 267, 7806– Hiemstra PS, van den Barselaar MT, Roest M, Nibbering PH 7813. and van Furth R.(1999). J. Leukocyte Biol., 66, 423–428. McLaughlin PMJ, Helfrich W, Kok K, Mulder M, Hu SW, Hochstrasser M.(1996). Cell, 84, 813–815. Brinker MGL, Ruiters MHJ, de Leij LFMH and Buys Hochstrasser M.(2000). Science, 289, 563–564. CHCM.(2000). Int. J. Cancer, 85, 871–876. Hori T, Osaka F, Chiba T, Miyamoto C, Okabayashi K, Michiels L, Van der Rauwelaert E, Van Hasselt F, Kas K and Shimbara N, Kato S and Tanaka K.(1999). Oncogene, 18, Merregaert J.(1993). Oncogene, 8, 2537–2546. 6829–6834. Miller AC, Blakely WF, Livengood D, Whittaker T, Xu J, Huang S-C and Lee T-C.(1998). Carcinogenesis, 19, 889–896. Ejnik JW, Hamilton MM, Parlette E, St John T, Gersten- HuangC,LiJ,DinM,WangL,ShiX,CastranovaV,Vallyathan berg HM and Hsu H.(1998). Environ. Health Perspect., 106, V, Ju G and Costa M.(2001). Mol. Cell. Biochem., 222, 29–34. 465–471. Hummer BT, Li X-L and Hassel BA.(2001). J. Virol., 75, Nakamura M and Tanigawa Y.(1999). J. Biol. Chem., 274, 7774–7777. 18026–18032.

Oncogene Transformation by fau and FUBI TG Rossman et al 1821 Nakamura M, Xavier RM and Tanigawa Y (1996). J. athy CO and Calderon R. (eds), Elsevier Science B.V.: Immunol., 156, 532–538. Amsterdam, Chapter 27, pp.285–295. Nakamura M, Xavier RM, Tsunematsu T and Tanigawa Y. Rossman TG and Wang Z.(1999). Carcinogenesis, 20, 311– (1995). Proc. Natl. Acad. Sci. USA, 92, 3463–3467. 316. Podust VN, Brownell JE, Gladysheva TB, Luo RS, Wang C, Sambrook J, Fritsch EF and Maniatis T.(1989). Molecular Coggins MB, Pierce JW, Lightcap ES and Chau V.(2000). Cloning.Cold Spring Harbor Laboratory Press: Cold Spring Proc. Natl. Acad. Sci. USA, 97, 4579–4584. Harbor, NY. Polyak K, Xia Y, Zweier JL, Kinzler KW and Vogelstein B. Shin S-I, Freedman VH, Risser R and Pollack R.(1975). Proc. (1997). Nature, 18, 300–305. Natl. Acad. Sci. USA, 72, 4435–4439. Potter JL, Narasimhan J, Mende-Mueller L and Haas AL. Tateishi K, Omata M, Tanaka K and Chiba T.(2001). J. Cell (1999). J. Biol. Chem., 274, 25061–25068. Biol., 155, 571–579. Raasi S, Schmidtke G and Groettrup M.(2001). J. Biol. Tornaletti S and Hanawalt PC.(1999). Biochimie, 81, Chem., 276, 35334–35343. 139–146. Ramirez P, Eastmond DA, Laclette JP and Ostrosky-Wegman Vogt B and Rossman TG.(2001). Mutat. Res., 478, 59–168. P.(1997). Mutat. Res., 386, 291–298. Wang Z and Rossman TG.(1994). Bio. Tech., 16, 460–463. Rani AS and Kumar S.(1992). Carcinogenesis, 13, 2021–2027. Wright G, Albert A and Mobashery S.(1998). Aminoglycoside Read MA, Brownell JE, Gladysheva TB, Hottelet M, Parent Antibiotics.New York: Kluwer Academic/Plenum Publish- LA, Coggins MB, Pierce JW, Produst VN, Luo RS, Chau V ers, pp.27–69. and Palombella VJ.(2000). Mol. Cell. Biol., 20, 2326–2333. Wong JM, Mafune K, Yow H, Rivers EN, Ravikumar TS, Redman KL and Burris GW.(1996). Biochem. J., 315, 21. Steel Jr GD and Chen LB.(1993). Cancer Res., 15, 1916– Rhim JS, Park DK, Arnstein P, Huebner RJ, Weisburger EK 1920. and Nelson-Rees WA.(1975). Nature, 256, 751–753. Yang J-J, Kang J-S and Krauss RS.(1998). Mol. Cell. Biol., Rossman TG, Visalli MA, Uddin AN and Hu Y.(2001). 18, 2586–2595. Arsenic Exposure and Health Effects.Chappel WR, Abern- Yeh ETH, Gong L and Kamitani T.(2000). Gene, 248, 1–14.

Oncogene