A Diverse Family of Proteins Containing Tumor Necrosis FactorReceptor-associated Factor Domains*

Received for publication, January 16, 2001, and in revised form, February 21, 2001Published, JBC Papers in Press, February 21, 2001, DOI 10.1074/jbc.M100354200

Juan M. Zapata‡, Krzysztof Pawlowski§, Elvira Haas¶i, Carl F. Ware¶, Adam Godzik, andJohn C. Reed**From The Burnham Institute, La Jolla, California 92037 and the ¶La Jolla Institute for Allergy and Immunology,San Diego, California 92121

We have identified three new tumor necrosis factor-receptor associated factor (TRAF) domain-containingproteins in humans using bioinformatics approaches, in-cluding: MUL, the product of the causative gene in Muli-brey Nanism syndrome; USP7 (HAUSP), an ubiquitin pro-tease; and SPOP, a POZ domain-containing protein.Unlike classical TRAF family proteins involved in TNFfamily receptor (TNFR) signaling, the TRAF domains(TDs) of MUL, USP7, and SPOP are located near the NH2termini or central region of these proteins, rather thancarboxyl end. MUL and USP7 are capable of binding invitro via their TDs to all of the previously identified TRAFfamily proteins (TRAF1, TRAF2, TRAF3, TRAF4, TRAF5,and TRAF6), whereas the TD of SPOP interacts weaklywith TRAF1 and TRAF6 only. The TD of MUL also inter-acted with itself, whereas the TDs of USP7 and SPOP didnot self-associate. Analysis of various MUL and USP7 mu-tants by transient transfection assays indicated that theTDs of these proteins are necessary and sufficient forsuppressing NF-kB induction by TRAF2 and TRAF6 aswell as certain TRAF-binding TNF family receptors. Incontrast, the TD of SPOP did not inhibit NF-kB induction.Immunofluorescence confocal microscopy indicated thatMUL localizes to cytosolic bodies, with targeting to thesestructures mediated by a RBCC tripartite domain withinthe MUL protein. USP7 localized predominantly to thenucleus, in a TD-dependent manner. Data base searchesrevealed multiple proteins containing TDs homologous tothose found in MUL, USP7, and SPOP throughout eu-karyotes, including yeast, protists, plants, invertebrates,and mammals, suggesting that this branch of the TD fam-ily arose from an ancient gene. We propose the monikerTEFs (TD-encompassing factors) for this large family ofproteins.

TNF1 receptor-associated factors (TRAFs) constitute a fam-ily of adapter proteins first identified for their binding to TNF

family receptors (TNFRs). TRAF family proteins regulate sev-eral of the functions of the TNFR superfamily, apparently bylinking the cytosolic domain of those receptors to downstreamprotein kinases or ubiquitin ligases (1–3). Six members of theTRAF family (TRAF1 through TRAF6) have been identified todate in humans and mice (1, 2).

All mammalian TRAFs and two recently characterized Dro-sophila TRAFs (4, 5) share a distinctive region near theirCOOH terminus denominated the “TRAF domain” (TD). TheTD represents a novel protein fold of about 180 amino acidsthat is the responsible for the interaction of the TRAFs with theTNF receptors and other adaptor proteins and kinases (1). Thex-ray crystallographic structures of the TDs of TRAF2 andTRAF3 show that they form an eight-stranded anti-parallelb-sandwich structure, with each TRAF domain monomer con-taining a surface crevice responsible for binding peptidyl motifsfound in the cytosolic domains of the TNF family receptors towhich they bind (6–10). Similarities and differences in thepeptidyl specificities of individual TRAFs account for theirselective associations with particular TNFR family membersand other TRAF-binding proteins, yielding specificity, diver-sity, redundancy, and competition among TRAFs with respectto ligand-inducible recruitment to various TNFR family recep-tor complexes (10–17). Through interactions with other adap-tor proteins, TRAFs also indirectly associate with members ofthe interleukin-1 receptor/Toll family (18, 19).

Several TRAFs can also bind a variety of protein kinases,including the NF-kB-inducing kinases IRAKs, NIK, RIP1, andRIP2/CARDIAK and the c-Jun NH2 kinase pathway activatorsMEKK1, Ask1, Misshapen (Msn), and the Germinal Centerkinase-related kinase (4, 5, 20–27). Thus, TRAF proteins phys-ically and functionally connect TNFRs and interleukin-1R/Tollreceptors to intracellular protein kinases, thereby linking thesereceptors to downstream signaling pathways. However, not allTRAFs are capable of interacting with downstream kinases,and some may function therefore as antagonists of TNFR andinterleukin 1-receptor/Toll signaling (28, 29).

Furthermore, some TDs self-associate, forming trimericstructures stabilized both by complementary in the eight-stranded b-sandwich fold that constitute the COOH-terminalregion of TDs and by a NH2-terminal coiled-coil region thatstabilizes TD trimerization (6–9). Heterotypic interactionsamong TDs have been described (16, 30), suggesting the possi-bility of mixed trimers that theoretically may account for theantagonism displayed among certain TRAF family members.

Meprins are a class of dimeric extracellular metalloprotein-ases of the astacin family (31) that also contain putative TDs ofunknown function (32). Since the region in the meprins that

* This work was supported in part by NCI, National Institutes ofHealth Grant CA-69381. The costs of publication of this article weredefrayed in part by the payment of page charges. This article musttherefore be hereby marked “advertisement” in accordance with 18U.S.C. Section 1734 solely to indicate this fact.

‡ Fellow of the Lady Tata Memorial Foundation.§ Present address: AstraZeneca R&D Lund, 221 87 Lund, Sweden.i Postdoctoral fellow of the Deutsche Forschungsgemeinschaft.** To whom correspondence should be addressed: The Burnham In-

stitute, 10901 N. Torrey Pines Rd., La Jolla, CA 92037. Tel.: 858-646-3140; Fax: 858-646-3194; E-mail: jreed@burnham-inst.org.

1 The abbreviations used are: TNF, tumor necrosis factor; TRAF,tumor necrosis factor receptor-associated factor; TNFR, tumor necrosisfactor family receptor; TEF, TD encompassing factor; TD, TRAF do-main; aa, amino acid(s); PCR, polymerase chain reaction; ORF, openreading frame; GST, glutathione S-transferase; PAGE, polyacrylamide

gel electrophoresis; NLS, nuclear localization signal; USP, ubiquitin-specific protease; POD, PML oncogenic domains.

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 276, No. 26, Issue of June 29, pp. 24242–24252, 2001© 2001 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

This paper is available on line at http://www.jbc.org24242

by guest on April 5, 2019

http://ww

w.jbc.org/

Dow

nloaded from

shares homology with the TD is extracellular, it is unlikely thatit plays any role in regulating TNF family receptors or intra-cellular TRAF family proteins. The presence of TD-like se-quences in meprins, however, suggests additional and broaderroles for the TDs in cell physiology (32).

In this report, we described a new branch of the TRAFfamily, which includes in humans the proteins MUL, USP7,and SPOP. The MUL gene has recently been shown to bemutated in patients with Mulibrey Nanism, an autosomal re-cessive disorder that affects several tissues of mesoderm origin(33). USP7 is a ubiquitin-specific protease that has been re-ported to bind the Vmw110 protein of herpes simplex virus(HSV1) (34, 35). SPOP was previously identified as an autoan-tigen in a patient with scleroderma pigmentosum (36). Allthree of these new TRAF proteins have a different topologicalorganization compared with the classical TRAFs implicated inTNF-R and Toll signaling, with the TDs located internally(MUL) or in the NH2-terminal part of the molecule (USP7 andSPOP), instead of the customary COOH-terminal localizationseen in the other previously identified human TRAFs.

As supportive evidence that these proteins do indeed containTDs, we demonstrate that these new TRAFs are able to inter-act with classical TRAF family proteins in vitro and to modu-late NF-kB induction by them, at least when overexpressed incells. However, the physiological role of these novel TD-con-taining proteins may be directed to purposes other that TNFR-and interleukin 1-receptor/Toll-mediated signal transduction,as discussed herein. Finally, using bioinformatics approaches,we identified several putative TD-containing proteins in di-verse eukaryotic species, including yeast, protozoa, slimemolds, nematodes, arthropods, and plants, suggesting an earlyevolutionary origin of the TD and implying roles beyond cyto-kine receptor signal transduction. We propose the monikerTEFs (TD-encompassing factors) for this large family ofproteins.

MATERIALS AND METHODS

Computer Methods—The PSI-BLAST program (NCBI, National In-stitute of Health) was used for searches of the public data bases usingas a query the predicted amino acid sequence of the TD of hTRAF2.Multiple sequence alignments were performed using the standardPM250 parameters from ClustalX and ClustalW. Relations among thedifferent members of the TD family were calculated using the phyloge-netic algorithm implemented in ClustalX. Models of the TDs of MUL,USP7, and SPOP were constructed by threading on the structure of theTD of TRAF2 (6, 7), using FFAS and MODELER (37, 38).

cDNA Cloning—Total RNA from the Jurkat cell line was reversetranscribed using Moloney murine leukemia virus reverse tran-scriptase (Stratagene) and random hexanucleotide primers. The result-ing cDNA was used to amplify the TD-containing portions of MUL (aa265–414), USP7 (aa 1–212), and SPOP (aa 1–172) by PCR using specificforward (MUL TD-F, 59-CCAGAATTCACCAGTGAATTAGTGCC-39;USP7-TD-F, 59-GCGAATTCCAGGCCGCG-39; SPOP-TD-F: 59-CTTC-GAATTCGCGATGTCAAGGGTTCC-39) and reverse (MUL-TD-R,59-CCACTCGAGTAATGTACCAATGCTAGTCC-39; USP7-TD-R, 59-TTCCTCGAGCCGACTTACCCTGTGTGC-39; SPOP-TD-R, 59-CCAT-GCTCGAGGTATTCTAGCCAGAAATG-39) primers. All PCR fragmentswere subcloned into pcDNA3-myc and pGEX-4T plasmids.

The strategy for obtaining cDNA clones encoding MUL was devisedbased on DNA sequence information provided by the EST clonesKIAA0898 (AB020705) and gi:5592172. Briefly, a MUL-specific reverseprimer MUL-FL-R (59-TTCTCGAGATTTGGCAATTACCTTCC-39) wasused to specifically reverse transcribe MUL cDNA from total RNA(Jurkat), using SuperscriptII reverse transcriptase (Life Technologies,Inc.). Two overlapping MUL cDNA fragments were then amplified byPCR, using Pwo DNA polymerase (Roche Molecular Biochemicals) andthe forward primer 59-TTGAATTCCGCCGAGAGCCGG-39 in combina-tion with the reverse primer MUL TD-R (above) as well as the forwardprimer 59-TTGAATTCGGGTCAGAAGACATCTCTAACCC-39 togetherwith the reverse primer MUL-FL-R (above). The cDNA containing thecomplete ORF of MUL was finally obtained by fusing both fragmentsusing a common EcoRV restriction site. DNA sequence analysis re-

vealed that the predicted ORF initiates with an AUG having a favorableKozak context for translation (39), with the 5-untranslated region con-taining at least one in-frame upstream stop-codon.

A similar approach was used for isolation of USP7 full-length cDNA.USP7-specific reverse primer USP7-FL-R (59-CCGTCCTCGAGTTGAA-CACACC-39) was used to specifically reverse-transcribe USP7 cDNAfrom Jurkat total RNA. Two overlapping USP7 cDNA fragments werethen amplified by PCR, using Pwo DNA polymerase (Roche MolecularBiochemicals) and the primer set (forward) 59-CGGGATCCATGAAC-CACCAGCAGC-39 with (reverse) 59-CCTGTGTGCTTCTCTAGATC-CCACGC-39, and the primer set (forward) 59-GCGTGGGATCTA-GAGAAGCACACAGG-39 together with the USP7-FL-R primer used forcDNA priming (above). A cDNA encompassing the complete ORF ofUSP7 was finally obtained by fusing both fragments using a commonXbaI restriction site.

Cells—293T and COS7 cells were obtained from ATCC (Rockville,MA) and cultured in Dulbecco’s modified Eagle’s high glucose medium(Life Technologies Inc.) supplemented with 5% fetal bovine serum (Hy-clone, UT), 1 mM glutamine, and antibiotics.

Plasmids—Plasmids containing cDNAs encompassing the completeopen reading frames (ORFs) of hTRAF1 (pSG5-TRAF1), hTRAF2(pcDNA3-HA-TRAF2), hTRAF3 (pbluscriptKS-TRAF3), hTRAF4(pcDNA3-HA-TRAF4), hTRAF5 (pcDNA3-Flag-TRAF5), and hTRAF6(pcDNA3-myc-TRAF6) have been previously described (4, 16, 40).pGEX-TRAF2(263–501), pGEX-CD40(ct), pGEX-Fas(ct), pGEX-LTbR(ct), and pGEX-NGFR(ct), and pGEX-HVEM(ct) have been previ-ously described (8). The reporter gene plasmid pUC13-4xNFkB-luc(containing 4 tandem HIV-NFkB response elements and the minimalc-fos promoter) and pCMV-b-galactosidase have been previously de-scribed (41, 42). Various deletion mutants of USP7 and MUL weregenerated by PCR, subcloned into pcDNA3-myc, and confirmed by DNAsequencing.

Production and Purification of Recombinant TRAF Domains—ForGST fusion protein production, pGEX-plasmids were transformed intocompetent XL-1 blue bacteria cells and grown in LB medium. Whenbacteria reached A600 5 1.0, GST-protein production was induced with1 mM isopropyl-1-thio-b-D-galactopyranoside for 4 h at 25 °C. Cells werethen recovered and resuspended in phosphate-buffered saline contain-ing 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, and 100mg/ml lysozyme and lysed by sonication. The GST fusion proteins werepurified from bacterial lysates by affinity chromatography using gluta-thione-Sepharose (Amersham Pharmacia Biotech), essentially as de-scribed (9, 16). The resins were then washed with phosphate-bufferedsaline containing 1 mM dithiothreitol until the OD280 nm reached ,0.01.

Protein Binding Experiments—In vitro GST-protein binding assayswere performed as described (4, 16). Briefly, [35S]methionine-labeledTRAF proteins were synthesized by coupled in vitro transcription-translation (Promega Inc.). Equal amounts of each labeled protein (2–6ml of lysate) were diluted with 250 ml of binding buffer (142 mM KCl, 5mM MgCl2, 10 mM Hepes, pH 7.4, 0.2% Nonidet P-40, 0.5 mM dithio-threitol, 1 mM EGTA, 0.5 mM phenylmethylsulfonyl fluoride, and amixture of other protease inhibitors (Roche Molecular Biochemicals),and incubated with the GST-protein resins (0.5 mg of protein immobi-lized on 10 ml of glutathione-Sepharose) at 4 °C for 2 h. The resins werethen extensively washed with binding buffer and the GST-protein bind-ing complexes were eluted with buffer containing 50 mM Tris-HCl, pH8, 1 mM dithiothreitol, and 100 mM glutathione, followed by analysis bySDS-PAGE and fluorography.

Reporter Gene Assays—For NF-kB reporter gene assays, 293T cellswere transfected using a calcium phosphate method and a total of 12 mgof plasmid DNA (including 0.5 mg of pUC13-4xNFkB-luc plasmid and 1mg of pCMV-b-galactosidase plasmid) at '50% confluence in 6-wellplates in duplicate. After 36 h, cells were lysed with 0.5 ml of Promegalysis buffer. The luciferase activity of 10 ml of each cell lysate wasdetermined using the Luciferase assay system from Promega, followingthe manufacturer’s protocol, and measured using a luminometer(EG&G Berthold). Luciferase activity was normalized relative to b-ga-lactosidase activity (mean 6 S.E.).

Immunofluorescence Confocal Microscopy—COS7 cells were trans-fected with LipofectAMINE Plus (Life Technologies, Inc.) and a total of3 mg of DNA. At 24 h after transfection, cells were plated in completetissue culture medium onto poly-lysinated cover glasses and allowed tosettle for 24 h at 37 °C, 5% CO2. Cells were fixed with 1:1 (v/v) meth-anol/acetone. After preblocking in 10 mM Hepes, pH 7.4, 150 mM NaCl,3% bovine serum albumin, 1% goat serum, cells were incubated withanti-Myc mAb (Santa Cruz Biotechnology) followed by secondary fluo-rescein isothiocyanate-labeled rabbit anti-mouse IgG (Dako) and incu-

TEFs: A Diverse Family of TRAF-containing Proteins 24243

by guest on April 5, 2019

http://ww

w.jbc.org/

Dow

nloaded from

bated with 1 mg/ml propidium iodide. Confocal microscopy was per-formed using a two-photon system (Bio-Rad).

RESULTS

Identification of Three Human TEFs—BLAST searches ofthe public data bases, including the non-redundant data base ofprotein sequences maintained at NCBI and the HGTS database, were performed using the sequence of the TD of humanTRAF2 as a query. Three human cDNAs displaying significanthomology with the TRAF2 TD were thus identified: MUL

(3e219), USP7 (e217), and SPOP (3e224). The putative TDs ofMUL, USP7, and SPOP share ;31–38% (mean 33.6%) aminoacid sequence identity with each other, and ;9–23% (mean16.7%) sequence identity with the TDs of human (hu) TRAF1-TRAF6 (Fig. 1). In contrast, the TDs of TRAF1-TRAF6 share;31–69% (mean 47.4%) amino acid sequence identity witheach other (Fig. 1B).

To further interrogate the possibility that MUL, USP7, andSPOP contain TDs, threading approaches were used, taking

FIG. 1. Comparison of sequences and predicted structures of TRAF domain proteins. A, an amino acid sequence alignment (ClustalW,standard PAM250 parameters) is presented for the TRAF domains of human USP7, SPOP, MUL, and TRAF1–6. Black and gray boxes indicateidentical and similar (conserved) residues, respectively. B, comparative analysis of the percentage of identical amino acid residues present invarious TDs. C, computer-generated models of TRAF domains are presented for human MUL, USP7, SPOP, and TRAF2, using the coordinates ofthe x-ray crystal structure of human TRAF2 as a template for threading. The coiled-coil of TRAF2 bears no sequence similarity to the coiled-coilof MUL and therefore could not be modeled using the TRAF2 coordinates, but it was added to the figure to illustrate the similarities in topologybetween TRAF2 and MUL. D, a schematic representation of the three human TEF-family proteins, MUL, USP7, and SPOP, is presented, makingcomparisons with TRAF1 and TRAF2. Symbols are as indicated.

TEFs: A Diverse Family of TRAF-containing Proteins24244

by guest on April 5, 2019

http://ww

w.jbc.org/

Dow

nloaded from

advantage of the published structure of the TD of huTRAF2 (6,7). Indeed, the sequences corresponding to the TDs of MUL,USP7, and SPOP readily adapted to the TD fold (Fig. 1C),without evidence of internal inconsistencies. All 3 of theseproteins were predicted to contain the eight-stranded anti-parallel b-sandwich structure of the COOH-terminal portion ofthe TD (so-called “TRAF-C” domain). In addition, MUL con-tains a predicted long a-helical domain upstream of theTRAF-C domain, analogous to TRAF2, and thus also displayspredicted structural similarity to the classical TRAFs in theso-called “TRAF-N” region.

Examination of the predicted complete ORFs of MUL, USP7,and SPOP reveals that the TDs are located either internally inthe molecule (MUL) or near the NH2 terminus (USP7 andSPOP) (Fig. 1D). Therefore, these proteins have a differenttopological organization from the previously described mem-bers of the TRAF family, which all have their TRAF domainslocated near the COOH terminus.

Similar to the classical TRAFs, the MUL protein also con-tains a RING finger domain (aa 15–55) near its NH2 terminus(Fig. 1D). This is followed by a ZF-Box domain (aa 90–132) andthen a predicted a-helical coiled-coil region (132–177), forminga tripartite motif, denominated the RBCC domain for RING,B-box, and coiled-coil. RBCC domains have been previouslyrecognized in a variety of proteins (43). After the RBCC do-main, a second coiled-coil region (aa 195–231) is found in MUL.Two leucine zipper domains may also be present in this regionof the protein (aa 197–218 and 222–245) (not shown), whichresides upstream of the TD. After the TD (aa 273–403), an-other predicted coiled-coil region (aa 427–446) is found, fol-lowed by two segments which are rich in acidic amino acidresidues (aa 452–577 and 868–964). This COOH-terminal re-gion of MUL distal to the TD has weak amino acid sequencesimilarity to murine histone deacetylase 1 and Haloarculamarismortu ribosomal protein L12, as determined by PSI-BLAST searches (44) (not shown). Interestingly, MUL containstwo putative nuclear localization signals (PAVEKRR at aa 847and KRRK at aa 851).

The predicted USP7 protein is 1102 aa in length (135 kDa),

and contains two domains sharing homology with ubiquitin-specific proteases (aa 215–233 and 448–518), which residedownstream of the predicted TD (aa 58–196) (Fig. 1D). Indeed,this region of USP7 has been confirmed by biochemical assaysto possess ubiquitin-specific protease activity (34).

SPOP is a 374-amino acid protein of unknown function whichcontains a POZ domain (aa 190–289) downstream of the pre-dicted TD (aa 33–164). This protein has been previously local-ized to nuclear bodies in a speckled pattern, thus prompting theacronym SPOP for Speckle-type POZ-domain protein (36).

The TDs of MUL, USP7, and SPOP Can Bind ClassicalTRAFs—TDs from the classical TRAF family proteins are ca-pable of interacting with themselves and sometimes each other.We therefore explored whether the predicted TDs of MUL,USP7, and SPOP could bind in vitro to human TRAF1, TRAF2,TRAF3, TRAF4, TRAF5, TRAF6, and the TRAF-binding pro-tein I-TRAF (TANK) (45, 46). For these experiments, the TDs ofMUL, USP7, and SPOP were expressed as GST fusion proteinsin bacteria and purified by glutathione-Sepharose affinity chro-matography, then tested for interactions with 35S-labeled invitro translated TRAF1-TRAF6 and I-TRAF, as well as the invitro translated TDs from MUL, USP7, and SPOP (Fig. 2).Consistent with the structural predictions, the TDs of MULand USP7 interacted in vitro with all of the classical TRAFs(TRAF1–6). The TD of SPOP also interacted with certainTRAFs, specifically TRAF1 and TRAF6. The TD of MUL addi-tionally interacted with itself and with I-TRAF, but did notbind the TDs of USP7 or SPOP (Fig. 2). The TDs of USP7 andSPOP, in contrast, failed to self-associate, unlike classicalTRAFs which commonly form trimers. Of note, the region ofMUL expressed for these experiments contained only the pre-dicted 8-strand anti-parallel b-sheet region, but lacked theupstream predicted a-helical segments that forms coiled-coilsand stabilizes trimerization of classical TRAFs. Binding exper-iments using GST-control proteins as well as GST-MUL, USP7,and SPOP (TDs) in combination with a variety of irrelevantcontrol proteins confirmed the specificity of these results (notshown).

Since TRAF family proteins are known to bind the cytosolicdomains of certain members of the TNFR family, we also ex-plored the possibility that MUL, USP7, or SPOP similarlymight interact with these receptors. Accordingly, in vitro pro-tein interaction assays were performed in which the TDs ofMUL, USP7, and SPOP (as well as TRAF2, which was em-ployed as a positive control) were tested for interactions withGST fusion proteins containing the cytosolic domains of severalTNF family receptors. The TDs of MUL, USP7, and SPOP wereeither produced as 35S-labeled proteins by in vitro translationor generated by expression as epitope-tagged proteins inHEK293T cells using transient transfection methods. Amongthese proteins, only the TD of MUL displayed interactions withTNF family receptors in vitro (Fig. 3 and data not shown).These data raise the possibility that the TD of MUL containssufficient structural similarity with classical TRAF family pro-teins to recognize the motifs within the cytosolic domains ofsome TNF family receptors, at least in vitro. Again, bindingexperiments using GST control proteins as well as GST-Fas,TNFR2, CD40, LTbR, and NGFRp75 in combination with avariety of irrelevant control proteins confirmed the specificityof these results (not shown).

MUL and USP7 Modulate NF-kB Induction by TRAFs—Because transient overexpression of some TRAFs inducesNF-kB activity, we tested the effects of MUL, USP7, and SPOPon activation of a NF-kB reporter gene plasmid by transienttransfection assays in HEK293 cells. Overexpression of MUL,USP7, or SPOP failed to induce NF-kB (not shown). Instead,

FIG. 2. Interactions of USP7(TEF1), SPOP(TEF2), and MUL-(TEF3) with human TRAF family proteins. In vitro protein bindingassays were performed using in vitro translated 35S-labeled TRAF1–6,I-TRAF, and the TDs of MUL, USP7, and SPOP (2–6 ml), which wereincubated with 1 mg of GST fusion proteins containing the TDs of MUL(top), USP7 (middle), or SPOP (bottom) immobilized on glutathione-Sepharose. After extensive washing, bound proteins were analyzed bySDS-PAGE followed by fluorography.

TEFs: A Diverse Family of TRAF-containing Proteins 24245

by guest on April 5, 2019

http://ww

w.jbc.org/

Dow

nloaded from

MUL and USP7 inhibited NF-kB induction caused by transientoverexpression of TRAF2 and/or TRAF6. As shown in Fig. 4, forexample, co-transfection of plasmids encoding full-length MULor USP7 with TRAF2 at a 2:1 ratio (TEF:TRAF) suppressedNF-kB activity by approximately half, relative to cells trans-fected with plasmids encoding TRAF2 alone. MUL also inhib-ited NF-kB induction by TRAF6, although USP7 was less ef-fective against this TRAF family member.

Moreover, the TDs of MUL and USP7 were sufficient forNF-kB inhibition. In fact, expression of a TD-containing regionof USP7 in the absence of its ubiquitin-protease domain wasmore inhibitory than full-length USP7, suppressing NF-kBinduction by both TRAF2 and TRAF6 (Fig. 4). In contrast, theTD of SPOP failed to suppress NF-kB induction caused byoverexpression of either TRAF2 or TRAF6. Likewise, express-

ing deletion mutants of MUL and USP7 lacking their TDsfailed to suppress NF-kB induction. Immunoblot analysis indi-cated that MUL and USP7 (as well as their TDs) did notinterfere with TRAF2 or TRAF6 protein production (notshown). We conclude therefore that the TDs of MUL and USP7are necessary and sufficient for suppression of NF-kB inductionin this assay where overexpression of TRAFs is used as amechanism for triggering an NF-kB response.

Subcellular Localization of MUL and USP7—TRAFs aretypically found in the cytosol, reportedly localizing to cytosolicstructures of undetermined origin (47–49). We therefore ex-plored the intracellular locations of MUL and USP7, and sev-eral deletion mutants of these proteins, using immunofluores-cence confocal microscopy and COS-7 cells transientlytransfected with plasmids encoding Myc epitope-tagged MULor USP7. As shown in Fig. 5, MUL is associated with cytosolicbodies, which appear as dots throughout the cytosol (panel A).In addition to COS-7 cells, this same pattern of immunolocal-ization was also seen in a variety of other cell lines, whentransfected with plasmids producing Myc-tagged or GFP-tagged MUL, including, HT1080, HT20, HeLa, and 293T. Two-color analysis using antibodies specific for proteins of mito-chondria, Golgi, lysosomes, or megasomes failed to suggest anorganelle to which MUL targets (not shown). Moreover, MULdid not co-localize with TRAF2, TRAF6, or the TNF familyreceptors CD40, Fas, and DR5.2

To determine the regions within the MUL protein responsi-ble for this pattern of subcellular targeting, a variety of MULdeletion mutants were expressed as Myc-tagged proteins bytransient transfection. Mutants lacking the RING domain(panel B), containing only the RBCC domain (panel C), lackingthe polyacidic region (panel D), or lacking the TRAF domain(panel E) displayed a similar subcellular location comparedwith full-length MUL. Thus the RBCC domain is sufficient fortargeting to punctate cytosolic structures. Furthermore, withinthe RBCC, the RING domain is expendable for proper target-ing. In contrast, diffuse cellular immunofluorescence encom-passing both the nucleus and cytosol was obtained for cellsexpressing Myc-tagged fragments of MUL containing only theTD (panel F), the TD with adjacent coiled-coil region (panel G),or TD with the adjacent COOH-terminal domain containingthe polyacidic segments and candidate nuclear localization sig-nal sequence (NLS) (panel H). A fragment of MUL containing

2 J. M. Zapata and J. C. Reed, unpublished observations.

FIG. 4. MUL and USP7 modulate NF-kB induction by TRAFs. A,293T cells were transfected in 6-well plates with a total of 12 mg of DNA,including 0.5 mg of the NF-kB reporter gene plasmid pUC13-4xNFkB-luc and 1 mg of pCMV-b-galactosidase. Control transfected cells addi-tionally received 10.5 mg of pcDNA3 empty plasmid (2). All othertransfections included 3.5 mg of either pcDNA3-hTRAF2 orpcDNA3-myc-hTRAF6 and 7 mg of plasmids encoding Myc-tagged MUL,USP7, various deletion mutants of these proteins, as indicated in thelower figure, or the TD of SPOP. Relative NF-kB activity was assessedby luciferase assays, using 10 ml of cell lysates prepared 36 h aftertransfection, with normalization for b-galactosidase activity. The re-sults are presented as percentage of NFkB activation (mean 6 S.D; n 53–8). Immunoblotting confirmed production of all protein (not shown).B, diagrammatic representation of the various MUL and USP7 mutantsanalyzed. Symbols are as indicated on the figure.

FIG. 3. Analysis of USP7(TEF1), SPOP(TEF2), and MUL(TEF3) interactions with TNF family receptors. In vitro protein binding assayswere performed using either (A) in vitro translated 35S-labeled TD of TRAF2 or MUL (10 ml reticulocyte lysate) or (B) Myc-tagged TD of TRAF2or MUL expressed in 293T cells (50 mg of cell lysate) in conjunction with GST fusion proteins containing the cytosolic domains of Fas, TNFR2,CD40, LTbR, or NGFRp75 (1 mg) immobilized on glutathione-Sepharose. Control GST and other GST control proteins were included in all assays,although results are not always presented. After washing, bound proteins were analyzed by SDS-PAGE followed by either fluorography (A) orimmunoblotting (B) using anti-TRAF2 (top) or anti-Myc (bottom) antibodies, respectively, followed by ECL-based detection.

TEFs: A Diverse Family of TRAF-containing Proteins24246

by guest on April 5, 2019

http://ww

w.jbc.org/

Dow

nloaded from

only the COOH-terminal domain with polyacidic segment andNLS was localized primarily to nuclei, although some faintercytosolic immunofluorescence was also seen (Fig. 5I).

Similar experiments were performed for USP7. The full-length USP7 protein was located predominantly in the nucleus,although cytosolic immunofluorescence was also observed (Fig.6A). A deletion mutant of USP7 containing essentially only theTD (aa 1–212) displayed a similar pattern of immunostaining,except the cytosolic component was more pronounced than seenwith the full-length USP7 protein. In contrast, a mutant ofUSP7 lacking the TRAF domain (aa 202–1102) (panel B) wascompletely excluded from the nucleus and instead was locateddiffusely throughout the cytosol. These results therefore sug-gest that the TD of USP7 is necessary and sufficient for targetof this protein to nuclei.

TEF Proteins Have an Ancient Origin and Can Be Classifiedinto Three Groups—PSI-BLAST searches of the public databases were seeded with the sequences of the TDs of MUL,USP7, and SPOP, as well as the TDs of TRAF1–6, and run tosaturation in search of additional TEFs. Only predictedpolypeptides yielding e values , 0.001 after five iterations wereconsidered positive. These studies revealed the existence ofmultiple independent cDNAs encoding potential TEF familyproteins in nearly all eukaryotic lineages, from yeast to hu-mans. Currently, almost 260 homologous proteins can be iden-tified with high confidence as members of the extended TEFfamily. The species containing candidate TEFs include yeast(Schizosaccaromyces pombe and Saccharomyces cerevisiae),protozoa (Trypanosoma brucei), Dyctiostelium (D. discoideum),nematodes (Caenorhabditis elegans), insects (Drosophila mela-nogaster), plants (including monocots (Oryza sativa (rice),Shorgum bicolor), and dicots (Arabidopsis thaliana and Medi-cago truncatula)), amphibians (Xenopus laevis), and mammals(human, mouse, and rat). It is interesting to note that despitethe complete sequencing of several prokaryotic genomes, in-cluding examples of both bacteria and archaea, no TEFs were

found in these organisms.An alignment of the amino acid sequences of some of the TDs

of these TEF family proteins is presented in Fig. 7. To make thealignment more readable, it was simplified by representinggroups of closely homologous proteins with one representativeexample (a complete multiple sequence alignment and accom-panying phylogenetic tree are available from our WEB server).Eight blocks of high homology are found within the aligned TDsequences, corresponding to the b-strands of the TDs of TRAF2and TRAF3 (6–10). The sequence alignment also demonstratesthe remarkably high percentage of identity of the TDs of hu-man SPOP and its counterparts in Drosophila (CG9924) (96%identity) and C. elegans (YNV5) (94% identity). Although onlythe TD sequences are presented in Fig. 7, comparison of thecomplete ORFs of the SPOP proteins of humans, flies, andworms also reveals striking sequence similarity (DrosophilaSPOP has 79% identity (89% homology) and C. elegans SPOPhas 63% identity (79% homology) relative to human SPOP).This level of conservation among such distant species suggestsan important role for SPOP in animal cell physiology.

Next, the phylogenetic program was used to produce a den-drogram in which different groups of TEFs with higher homol-ogies could be identified (Fig. 8). It is important to note thatbecause of the use of orthologs, paralogs, and several proteinsfrom the same species in the tree, accurate reflection of theactual evolutionary distances among the different members ofthe TEF family is not possible. Relations among the differentmembers of the TEF family were calculated using the standardparameters of the phylogenetic programs implemented inClustalX. The use of the Fitch-Margoliash method (Phylip)produced similar results (not shown). The TEF proteins repre-sented in the tree can be subdivided in three groups. Group Icontains the previously known TRAF proteins, including hu-man TRAF1 through TRAF6, and the recently identifiedTRAFs from Drosophila and C. elegans. This group I branchalso includes the Meprins, a family of extracellular metallopro-

FIG. 5. Subcellular localization of MUL and MUL deletion mutants. COS7 cells were transfected in 6-well dishes with 3 mg of pcDNA3-mycplasmids containing the MUL(TEF3) mutants depicted at the right-side of the figure: (A) full-length MUL; (B) MUL (aa 63–964); (C) MUL (aa1–270); (D) MUL (aa 1–414); (E) MUL D254–450; (F) MUL (aa 265–414); (G) MUL (aa 206–414); (H) MUL (aa 265–964); and (I) MUL (aa412–964). At 48 h after transfection, cells were fixed in methanol-acetone and stained with anti-Myc mAb, followed by anti-mouse IgG-fluoresceinisothiocyanate and propidium iodide. Cells were imaged by confocal microscopy. Cells stained with secondary antibody and cells transfected withempty pcDNA3-myc plasmid served as negative controls (not shown). 100-mm bars are shown in the Fig.

TEFs: A Diverse Family of TRAF-containing Proteins 24247

by guest on April 5, 2019

http://ww

w.jbc.org/

Dow

nloaded from

teinases (31, 32). It is interesting to note that Group I is theonly branch of the TEF family that does not include homo-logues from plants or unicellular organisms, suggesting a moremodern origin. Group II includes MUL, SPOP, and its homo-logues from Drosophila and C. elegans, as well as additionalTEFs from C. elegans and plants (Arabidopsis (At), Oryza (Or),and Shorgum (Sb)). Group III includes USP7, several USP7-related proteins that contain TDs in combination with appar-ent ubiquitin-specific protease domains (Drosophila (CG1490),C. elegans (3878045 and 3877391), Arabidopsis (6671947), andyeast, including S. pombe and S. cerevisiae (6014652, UBPB-_SCHPO and UBPF_yeast), and additional TEFs from Dro-sophila, C. elegans, plants, Dyctiostelium, Trypanosoma, and C.elegans. We presume that the Group III TDs are the mostancient, since they are present in yeast and protozoa, in addi-tion to higher organisms. Proteins belonging to Groups II andIII typically have a domain organization differing from theGroup I proteins, with the TRAF domain located near the NH2

terminus, followed by a diversity of COOH-terminal domains.

DISCUSSION

Identification of Three TEFs in Humans—Using bioinfor-matics methods, we have identified three proteins in humans,MUL, USP7, and SPOP, that contain a variant of the TD. Theevidence that this conserved domain recognized in the MUL,USP7, and SPOP proteins represents a variant of the TDincludes (a) sequence alignments; (b) structure prediction(threading); (c) ability of variant TDs to bind the TDs of clas-sical TRAF family members; and (d) modulation of the functionof classical TRAFs in terms of NF-kB induction. In the case ofMUL, this similarity to classical TRAFs also extends to addi-tional protein interactions, including an ability to bind I-TRAF(TANK) and some TNF family receptors. Unlike the previouslydescribed members of the TRAF family, the TDs in MUL,USP7, and SPOP are located near the NH2 terminus ratherthan COOH terminus. Expression studies indicate that themRNAs encoding MUL, USP7, and SPOP are widely present inadult human tissues (Refs. 33 and 36, and data not shown),suggesting the functions of these TEF-family proteins arelikely to be applicable to multiple cell types.

USP7 (TEF1)—USP7 (HAUSP) contains a variant TD fol-lowed by two ubiquitin-specific protease (USP) domains.Among the 3 human TEF family genes described here, it wasthe first to be identified and thus we have proposed TEF1 as analternative designation for this protein. The protease domain-containing region of this USP7 has been expressed in bacteriaand confirmed by biochemical assays to be capable of cleavingpolyubiquitin chains (34). The USP7(TEF1) gene maps to chro-mosome 16p13.3 (50), representing a region commonly involvedin translocations, deletions, and other cytogenetic abnormali-ties in cancers such as ependymonas, basal cell carcinomas,and acute myeloid leukemias (51–53). Genes within the 16p13region have also been implicated in certain hereditary syn-dromes, including Rubinstein-Taybi syndrome, tuberous scle-rosis complex, adult polycystic kidney disease, and a-thalasse-mia/familial mental retardation syndrome (54–59). However, itis unknown whether USP7(TEF1) is directly involved in can-cers or genetic diseases. Homologues of USP7(TEF1) whichcontain both a TD and USP domain are present in diverseorganisms, with the closest found in Drosophila (CG1490),which shares .60% amino acid sequence identity with its hu-man counterpart. This strong conservation of amino acid se-quence homology suggests an evolutionarily conserved functionfor this protein.

USP7 was originally identified based on its ability to associ-ate with the herpes simplex virus-type 1 immediate-early pro-tein Vmw110 (34). Vmw110 localizes to nuclear structureswhere PML is found, the so-called PML oncogenic domains(PODs). The function of PODs is unclear, but they have beenimplicated in interferon responses and pathogenesis of varioushuman diseases, including acute promyelocytic leukemia andviral infections (reviewed in Ref. 60). Increasing evidence sug-gests a role for PODs as specialized sites of transcriptionalregulation, perhaps in assembling multiprotein transcriptionalcomplexes or controlling targeted degradation of nuclear pro-teins (61–64). The Vmw110 protein is a potent activator of geneexpression and is required for efficient virus reactivation fromlatency (35). Of potential relevance to its interactions with aUSP, the Vmw110 protein reportedly activates proteosome-de-pendent degradation of several substrates, and it has beensuggested that this targeted protein degradation is required tore-activate the lytic cycle of the virus (65, 66). Expression ofVmw110 in cells initially causes an apparent increase in therecruitment of USP7(TEF1) into PODs, followed by dispersionof PML and other POD constituents from these nuclear struc-tures (34). Interestingly, mutations that disrupt interactions

FIG. 6. Subcellular localization of USP7(TEF1). COS7 cells weretransfected in 6-well dishes either with 3 mg of pcDNA3-myc-USP7 (A),pcDNA3-myc-USP7DTD (aa 202–1102) (B), or pcDNA3-myc-USP7DTD(aa 1–212) (C). At 36 h after transfection, cells were fixed in methanol-acetone and stained with anti-Myc mAb, followed by anti-mouse IgG-fluorescein isothiocyanate and propidium iodide. Cells were imaged byconfocal microscopy. Small squares show the USP7 (green channel) ornuclear (red channel) staining individually, whereas large squares showoverlay. Cells stained with secondary antibody and cells transfectedwith empty pcDNA3-myc plasmid served as a negative controls (notshown). 100 mm bars are shown in the figure.

TEFs: A Diverse Family of TRAF-containing Proteins24248

by guest on April 5, 2019

http://ww

w.jbc.org/

Dow

nloaded from

between Vmw110 and USP7 also inhibit the activation of geneexpression and virus replication mediated by Vmw110, sug-gesting a causal role for USP7 as a co-factor in these virus-mediated processes (35).

Since USP7(TEF1) is normally located primarily in the nu-cleus, it is unlikely that this protein participates in early signaltransduction events triggered by TNF family receptors andTRAFs. It is conceivable, however, that USP7(TEF1) could

FIG. 7. Multiple sequence alignment of TDs of representative members of the TEF family. An amino acid sequence alignment wasprepared using Clustal X (standard parameters). Gray boxes indicate identical and similar (conserved) residues. Numbers indicate residue positionwithin each protein of the last amino acid shown.

TEFs: A Diverse Family of TRAF-containing Proteins 24249

by guest on April 5, 2019

http://ww

w.jbc.org/

Dow

nloaded from

regulate or could be regulated by TRAFs that translocate to thenucleus under certain circumstances, such as TRAF4 which isreported to reside in the nuclei of some tumors (67). Interest-ingly, our deletional analysis suggests that the TD is respon-sible for the nuclear targeting of USP7(TEF1), raising thepossibility that this domain binds other nuclear proteins. Over-expression of a truncation mutant of USP7(TEF1) containingessentially only the TD may have saturated nuclear-bindingsites, explaining the spill-over of this protein fragment into thecytosol and its potent suppression of TRAF-induced NF-kB.Although recruitment of a USP to TRAFs might be expected tointerfere with NF-kB induction by preventing I-kB degradationresulting from polyubiquitination (68), the USP domains ofUSP7(TEF1) are not required for TRAF antagonism.

SPOP (TEF2)—The SPOP (TEF2) protein was previouslyidentified as an autoantigen in a patient with sclerodermapigmentosum (36). The TD of SPOP (aa 33–164) is located nearthe NH2 terminus of the molecule, and is followed by a POZdomain (aa 190–289). POZ domains have been identified inseveral transcriptional repressors (reviewed in Ref. 69). Thesedomains bind components of the SMRTzNcoR repressor com-plex, which has histone deacetylase activity (70, 71). SPOP(TEF2) and several of the POZ family proteins localize to dis-crete nuclear structures, but whether these are the same asPODs remains controversial. Deletional analysis of the SPOP(TEF2) protein has provided evidence that both the TD andPOZ domain are required for targeting to nuclear speckles (36).Of note, the TDs of SPOP and USP7 failed to interact in our invitro protein binding assays, suggesting that these proteins donot directly associate, despite their targeting to nuclear sub-compartments. Although the TD of SPOP displayed at leastweak interactions with TRAF1 and TRAF6 in vitro, it did notdemonstrate an antagonistic effect on TRAF-mediate inductionof NF-kB. The physiological roles of SPOP (TEF2) thus remainto be defined. The striking amino acid sequence similarity ofthe human, Drosophila, and C. elegans homologues of SPOP(TEF2) suggests a conserved function within the animal king-dom. The human SPOP gene is located in chromosome 17, andis flanked by NGFR (TNFR-16) and distal-less homeobox 4genes.

MUL(TEF3)—The MUL(TEF3) protein contains a RBCC do-main, TD, and polyacidic domains. Unlike the other TEFsidentified thus far in humans, MUL(TEF3) is a cytosolic pro-tein which is localized in a distinctive punctuate pattern. Fur-

thermore, the size of these punctate structures increased, withincreasing levels of MUL achieved by transfection. Based onimmunofluorescence analysis of deletion mutants of MUL-(TEF3), we determined that the RBCC domain is necessary andsufficient for targeting to cytosolic structures. The identity ofthese foci of MUL(TEF3) accumulation, however, is unclear.Using two-color immunofluorescence techniques, we have ex-cluded mitochondria, lysosomes, Golgi, and megasomes aslikely candidates for the cytosolic bodies with which MUL-(TEF3) associates.2 Interestingly, the subcellular localizationof MUL(TEF3) is reminiscent of some other members of theRBCC family, such as BERP and Rfp (72, 73). Deletional anal-ysis indicates that the B-box of Rfp is required for its targetingto cytosolic granule-like structures (73), raising the possibilitythat elements within the RBCC tripartite domain account fortargeting of MUL, Rfp, BERP, and certain other members ofthe RBCC family to these cytosolic structures. Also, the BERPprotein binds myosin V and a-actinin, and has been speculatedto regulate organelle transport (72, 74). It remains to be deter-mined, however, whether MUL(TEF3) co-localizes with BERPor Rfp in cells. Moreover, the location within cells of someRBCC family members such as Rfp (which can function as arepressor of the HIV LTR) and Xnf (which is involved in dorsal-ventral patterning in Xenopus) can change from cytosolic tonuclear in concert with differences in cell context or proteinmodification by phosphorylation (73, 75). Thus, the location ofMUL(TEF3) may be subject to regulation, although we consis-tently observed a punctate cytosolic pattern of immunofluores-cence in 5 of 5 tumor cell lines examined here.

Based on its subcellular location, MUL(TEF3) conceivablycould participate in physiological regulation of TRAF familyproteins. Of note, previous attempts to localize TRAF familyproteins have revealed association with punctate cytosolicstructures prior to activation of TNF family receptors, followedby translocation of TRAFs to the plasma membrane after li-gand addition (47–49). However, two-color immunofluores-cence studies indicated that: (a) TRAF2 and TRAF6 do notco-localize with MUL(TEF3) and (b) the subcellular targetingof MUL did not correlate with suppression of TRAF function,inasmuch as a truncation mutant of MUL containing essen-tially only the TD was equally effective as the full-length pro-tein, despite its failure to target to cytosolic subregions. Itmight also be noted that the TD-only mutant of MUL(TEF3)was also found in the Triton X-100 soluble fraction, while

FIG. 8. TEF family dendrogram. Re-lations among the various members of theTEF family were calculated using thestandard parameters of the phylogeneticprogram implemented in ClustalX. Thetree was bootstrapped 1000 times. ThreeTEF subgroups are discerned from theanalysis.

TEFs: A Diverse Family of TRAF-containing Proteins24250

by guest on April 5, 2019

http://ww

w.jbc.org/

Dow

nloaded from

full-length MUL(TEF3) was predominantly in the Triton X-100insoluble fraction.2 Thus, the association of MUL(TEF3) withthe Triton X-100 insoluble fraction precluded attempts to as-sess association of this protein with classical TRAFs by co-immunoprecipitation assays. At this point, therefore, no com-pelling evidence exists to imply a physiological role ofMUL(TEF3) in regulating signal transduction by TRAFs, al-though additional experimentation will be required to evaluatethis possibility further.

While this paper was in preparation, the MUL(TEF3) geneon chromosome 17q22–23 was reported to be mutated in pa-tients with Mulibrey Nanism, an autosomal recessive disorderthat affects several tissues of mesodermal origin (33). MulibreyNanism is characterized by severe growth failure of prenatalonset, constrictive pericardium with consequent hepatomegaly,hypoplasia of several endocrine glands with consequent hormo-nal deficiency, triangular face with hydrocephaloid skull, andsusceptibility to develop Wilm’s tumors. A substantial portionof patients are suspected to be lost by early abortion and othersby infantile death. Thus, the MUL(TEF3) protein plays impor-tant roles in human development and possibly tumor suppres-sion, but the biochemical mechanism of the protein remainsenigmatic.

Four independent frameshift mutations in the MUL(TEF3)gene have been identified in families with Mulibrey Nanism,all of which are predicted to produce truncated versions of theMUL(TEF3) protein. Although it remains to be determinedwhether these mutant MUL proteins are stable when ex-pressed in cells, the shortest of the truncated proteins identi-fied thus far retains only the RING, B-box, and a portion of thecoiled-coil domain, thus having an incomplete RBCC tripartitedomain. Two mutants retain the RBCC domain and TD, whileanother lacks only the last 227 amino acids and thus is missingthe second of the two polyacidic domains (33). With respect tothe truncation mutants of MUL(TEF3) associated with Muli-brey Nanism, our deletion analysis suggests that at least 3 ofthe 4 mutant proteins would still target to cytosolic structures,since the RBCC domain was sufficient for localization to punc-tate cytosolic structure in immunofluorescence experiments.Thus, it is unlikely that improper subcellular targeting uni-formly accounts for the dysfunction of such mutant MUL-(TEF3) proteins. Future explorations of the protein interactionpartners of the MUL(TEF3) protein may provide insights intothe specific biochemical defect that accounts for the diversedevelopmental abnormalities seen in patients harboring muta-tions of their MUL(TEF3) genes.

TEFs Represent an Ancient Group of Proteins Found in Di-verse Unicellular and Multicellular Eukaryotes—The discoveryof multiple proteins containing candidate TDs in organisms asdiverse as yeast, protists, plants, nematodes, flies, amphibians,and mammals suggests a very ancient origin for this proteinfold, and implies strong selective pressure for maintaining itsstructural and functional characteristics during evolution. Pre-sumably, the TD is used as a protein-interaction motif thatpermits both homo- and heterotypic interactions with otherproteins carrying TDs. However, as demonstrated by the inter-actions of the TDs of classical TRAFs with TNF family recep-tors and adapter proteins such as TRADD, the TD can also beemployed as a scaffold on which other types of protein recog-nition elements are displayed (6–10). Diversity among TEFshas been achieved during evolution by combining (within asingle polypeptide chain) the TD with numerous other types ofdomains, including metalloproteinase domains (meprins),RING and zinc-finger domains (TRAFs), ubiquitin-specific pro-tease domains (USP7/TEF1), POZ domains (SPOP/TEF2), andRBCC domains (MUL/TEF3). Furthermore, some eukaryotic

species contain numerous potential TEF family genes. TheArabidopsis EST data base, for example, contains over 100cDNA sequences capable of encoding polypeptides with homol-ogy to TDs (the complete phylogenetic tree and multiple align-ments are available from our WEB server. The expansion ofTEF family proteins in Arabidopsis raises the possibility that awide diversity of beneficial uses of the TD were exploited dur-ing evolution of this dicot. It will be interesting in futureexperiments to explore whether some of the functions of theseTEF family proteins are analogous to the role played by TRAFsin animals as mediators of signal transduction pathways in-volved in the innate immune responses to pathogens, particu-larly given the prevalence of Toll-like receptors in Arabidopsis(reviewed in Refs. 76 and 77). Regardless of their specific func-tions, the discovery of large numbers of diverse TEF familyproteins in multiple eukaryotic lineages suggests that the func-tions and uses of the TD are far broader than originallysuspected.

Acknowledgments—We thank D. Bredesen and E. Leo for plasmids,E. Monosov for helpful advice with the confocal analysis, R. Cornell formanuscript preparation, S. Sovath for expert technical assistance, andF. Stenner-Liewen for helpful discussion.

REFERENCES

1. Arch, R. H., Gedrich, R. W., and Thompson, C. B. (1998) Genes Dev. 12,2821–2830

2. Wallach, D., Varfolomeev, E. E., Malinin, N. L., Goltsev, Y. V., Kovalenko,A. V., and Boldin, M. P. (1999) Annu. Rev. Immunol. 17, 331–367

3. Deng, L., Wang, C., Spencer, E., Yang, L., Braun, A., You, J., Slaughter, C.,Pickart, C., and Chen, Z. J. (2000) Cell 103, 351–361

4. Zapata, J., Matsuzawa, S., Godzik, A., Leo, E., Wasserman, S., and Reed, J.(2000) J. Biol. Chem. 275, 12102–12107

5. Liu, H., Su, Y.-C., Becker, E., Treisman, J., and Skolnik, E. (1999) Curr. Biol.9, 101–104

6. McWhirter, S., Pullen, S., Holton, J., Crute, J., Kehry, M., and Alber, T. (1999)Proc. Natl. Acad. Sci. U. S. A. 96, 8408–8413

7. Park, Y., Burkitt, V., Villa, A., Tong, L., and Wu, H. (1999) Nature 398,533–538

8. Ye, H., Park, Y., Kreishman, M., Kieff, E., and Wu, H. (1999) Mol Cell 4,321–330

9. Ni, C.-Z., Welsh, K., Leo, E., Chiou, C.-K., Wu, H., Reed, J. C., and Ely, K. R.(2000) Proc. Natl. Acad. Sci. U. S. A. 97, 10395–10399

10. Park, Y. C., Ye, H., Hsia, C., Segal, D., Rich, R. L., Liou, H.-C., Myszka, D. G.,and Wu, H. (2000) Cell 101, 777–787

11. Gedrich, R. W., Gilfillan, M. C., Duckett, C. S., Van Dongen, J. L., andThompson, C. B. (1996) J. Biol. Chem. 271, 12852–12858

12. Boucher, L.-M., Marengere, L. E. M., Lu, Y., Thukral, S., and Mak, T. W.(1997) Biochem. Biophys. Res. Commun. 233, 592–600

13. Pullen, S., Miller, H., Everdeen, D., Dang, T., Crute, J., and Kehry, M. (1998)Biochemistry 37, 11830–11845

14. Pullen, S., Labadia, M., Ingraham, R., McWhirter, S., Everdeen, D., Alber, T.,Crute, J., and Kehry, M. (1999) Biochemistry 38, 10168–10177

15. Pullen, S. S., Dang, T. T., Crute, J. J., and Kehry, M. R. (1999) J. Biol. Chem.274, 14246–54

16. Leo, E., Welsh, K., Matsuzawa, S., Zapata, J. M., Kitada, S., Mitchell, R., Ely,K. R., and Reed, J. C. (1999) J. Biol. Chem. 274, 22414–22274

17. Yamamoto, H., Kishimoto, T., and Minamoto, S. (1998) J. Immunol. 161,4753–4759

18. Muzio, M., Natoli, G., Saccani, S., Levrero, M., and Mantovani, A. (1998) J.Exp. Med. 187, 2097–2101

19. Cao, Z., Xiong, J., Takeuchi, M., Kurama, T., and Goeddel, D. V. (1996) Nature383, 443–446

20. Hsu, H., Huang, J., Shu, H., Baichwal, V., and Goeddel, D. V. (1996) Immunity4, 387–396

21. Song, H. Y., Regnier, C. H., Kirschning, C. J., Goeddel, D. V., and Rothe, M.(1997) Proc. Natl. Acad. Sci. U. S. A. 94, 9792–9796

22. Shi, C. S., Leonardi, A., Kyriakis, J., Siebenlist, U., and Kehrl, J. H. (1999)J. Immunol. 163, 3279–3285

23. Thome, M., Hofmann, K., Burns, K., Martinon, F., Bodmer, J., Mattman, C.,and Tschopp, J. (1998) Curr. Biol. 16, 885–888

24. McCarthy, J., Ni, J., and Dixit, V. (1998) J. Biol. Chem. 273, 16968–1697525. Hoeflich, K. P., Yeh, W. C., Yao, Z., Mak, T. W., and Woodgett, J. R. (1999)

Oncogene 18, 5814–582026. Nishitoh, H., Saitoh, M., Mochida, Y., Takeda, K., Nakano, H., Rothe, M.,

Miyazono, K., and Ichijo, H. (1998) Mol. Cell 2, 389–39527. Baud, V., Liu, Z.-G., Bennett, B., Suzuki, N., Xia, Y., and Karin, M. (1999)

Genes Dev. 13, 1297–130828. Rothe, M., Sarma, V., Dixit, V. M., and Goeddel, D. V. (1995) Science 269,

32767–3277029. Ye, X., Mehlen, P., Rabizadeh, S., VanArsdale, T., Zhang, H., Shin, H., Wang,

J., Leo, E., Zapata, J., Hauser, C., Reed, J., and Bredesen, D. E. (1999)J. Biol. Chem. 274, 30202–30208

30. Rothe, M., Wong, S. C., Henzel, W. J., and Goeddel, D. V. (1994) Cell 78,681–692

TEFs: A Diverse Family of TRAF-containing Proteins 24251

by guest on April 5, 2019

http://ww

w.jbc.org/

Dow

nloaded from

31. Bond, J. S., and Beynon, R. J. (1995) Protein Sci. 4, 1247–126132. Uren, A. G., and Vaux, D. L. (1996) Trends Biochem. Sci. 21, 244–24533. Avela, K., Lipsanen-Nyman, M., Idanheimo, N., Seemanova, E., Rosengren, S.,

Makela, T. P., Perheentupa, J., de la Chapelle, A., and Lehesjoki, A.-E.(2000) Nat. Genet. 25, 298–301

34. Everett, R., Meredith, M., Orr, A., Cross, A., Kathoria, M., and Parkinson, J.(1997) EMBO J. 16, 566–577

35. Everett, R. D., Meredith, M., and Orr, A. (1999) J. Virol. 73, 417–42636. Nagai, Y., Kojima, T., Muro, Y., Hachiya, T., Nishizawa, Y., Wakabayashi, T.,

and Hagiwara, M. (1997) FEBS Lett. 418, 23–2637. Jaroszewski, L., Rychlewski, B., Zhang, B., and Godzik, A. (1998) Protein Sci.

7, 1431–144038. Sali, A., and Blundell, T. L. (1993) J. Mol. Biol. 234, 779–81539. Kozak, M. (1992) Annu. Rev. Cell Biol. 8, 197–22540. Leo, E., Zapata, J. M., and Reed, J. C. (1999) Eur. J. Immunol. 29, 3908–391341. Galang, C. K., Der, C. J., and Hauser, C. A. (1994) Oncogene 9, 2913–292142. Miyashita, T., and Reed, J. C. (1995) Cell 80, 293–29943. Borden, K. L. (1998) Biochem. Cell Biol. 76, 351–35844. Altschul, S. F., Madden, T. L., Schaeffer, A. A., Zhang, J., Zhang, Z., Miller, W.,

and Lipman, D. J. (1997) Nucleic Acids Res. 25, 3389–340245. Cheng, G., and Baltimore, D. (1996) Genes Dev. 10, 963–97346. Rothe, M., Xiong, J., Shu, H.-B., Williamson, K., Goddard, A., and Goeddel,

D. V. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 8241–824647. Devergne, O., Hatzivassiliou, E., Izumi, K. M., Kaye, K. M., Kleijnen, M. F.,

Kieff, E., and Mosialos, G. (1996) Mol. Cell. Biol. 16, 7098–710848. Ansieau, S., Scheffrahn, I., Mosialos, G., Brand, H., Duyster, J., Kaye, K.,

Harada, J., Dougall, B., Hubinger, G., Kieff, E., Herrmann, F., Leutz, A.,and Gruss, H.-J. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 14053–14058

49. Krajewska, M., Krajewski, S., Zapata, J. M., Van Arsdale, T., Gascoyne, R. D.,Berern, K., McFadden, D., Shabaik, A., Hugh, J., Reynolds, C., Clevenger,C. V., and Reed, J. C. (1998) Am. J. Pathol. 152, 1549–1561

50. Robinson, P. A., Lomonte, P., Leek, Markham, A. F., and Everett, R. D. (1998)Cytogenet. Cell Genet. 83, 100

51. Zheng, P.-P., Pang, J. C.-S., Hui, A. B.-Y., and Ng, H.-K. (2000) Cancer Genet.Cytogenet. 122, 18–25

52. Jin, Y., Merterns, F., Persson, B., Gullestad, H. P., Jin, C., Warloe, T.,Salemark, L., Jonsson, N., Risberg, B., Mandahl, N., Mitelman, F., andHeim, S. (1997) Cancer Res. 57, 404–406

53. Panagopoulos, I., Isaksson, M., Lindvall, C., Bjorkholm, M., Ahlgren, T.,Fioretos, T., Heim, S., Mitelman, F., and Johansson, B. (2000) Genes Chro-mosomes Cancer 28, 415–424

54. Wallerstein, R., Anderson, C. E., Hay, B., Gupta, P., Gibas, L., Ansari, K.,Cowchock, F. S., Weinblatt, V., Reid, C., Levitas, A., and Jackson, L. (1997)

J. Med. Genet. 34, 203–20655. Petrij, F., Dorsman, J. C., Dauwerse, H. G., Giles, R. H., Peeters, T.,

Hennekam, R. C., Breuning, M. H., and Peters, D. J. (2000) Am. J. Med.Genet. 92, 47–52

56. Harris, P. C., Ward, C. J., Peral, B., and Hughes, J. (1995) J. Am. Soc. Nephrol.6, 1125–1133

57. Eussen, B. H., Bartalini, G., Bakker, L., Balestri, P., Di Lucca, C., Van Hemel,J. O., Dauwerse, H., van Den Ouweland, A. M., Ris-Stalpers, C., Verhoef, S.,Halley, D. J., and Fois, A. (2000) J. Med. Genet. 37, 287–291

58. Gibbons, R. J., and Higgs, D. R. (1996) Medicine (Baltimore) 75, 45–5259. Holinski-Feder, E., Reyniers, E., Uhrig, S., Golla, A., Wauters, J., Kroisel, P.,

Bossuyt, P., Rost, I., Jedele, K., Zierler, H., Schwab, S., Wildenauer, D.,Speicher, M. R., Willems, P. J., Meitinger, T., and Kooy, R. F. (2000) Am. J.Hum. Genet. 66, 16–25

60. Everett, R. D. (1999) Trends Biochem. Sci. 24, 293–29561. Alcalay, M., Tomassoni, L., Colombo, E., Stoldt, S., Grignani, F., Fagioli, M.,

Szekely, L., Helin, K., and Pelicci, P. G. (1998) Mol. Cell. Biol. 18,1084–1093

62. LaMorte, V. J., Dyck, J. A., Ochs, R. L., and Evans, R. M. (1998) Proc. Natl.Acad. Sci. U. S. A. 95, 4991–4996

63. Wang, Z., Delva, L., Gaboli, M., Rivi, R., Giorgio, M., Cordon-Cardo, C.,Grosveld, F. P., and Pandolfi, P. (1998) Science 279, 1547–1551

64. Bloch, D. B., Chiche, J. D., Orth, D., de la Monte, S. M., Rosenzweig, A., andBloch, K. D. (1999) Mol. Cell. Biol. 19, 4423–4430

65. Everett, R. D., Freemont, P., Saitoh, H., Dasso, M., Orr, A., Kathoria, M., andParkinson, J. (1998) J. Virol. 72, 6581–6591

66. Chelbi-Alix, M. K., and de The, H. (1999) Oncogene 18, 935–94167. Regnier, C., Tomasetto, C., Moog-Lutz, C., Chenard, M., Wendling, C., Basset,

P., and Rio, M. (1995) J. Biol. Chem. 270, 25715–2572268. Karin, M., and Ben-Neriah, Y. (2000) Annu. Rev. Immunol. 18, 621–66369. Albagli, O., Dhordain, P., Deweindt, C., Lecocq, G., and Leprince, D. (1995)

Oncogene 6, 1193–119870. Wong, C. W., and Privalsky, M. L. (1998) J. Biol. Chem. 273, 27695–2770271. Huynh, K. D., and Bardwell, V. J. (1998) Oncogene 17, 2473–248472. El-Husseini, A. E., and Vincent, S. R. (1999) J. Biol. Chem. 274, 19771–1977773. Cao, T., Borden, K. L., Freemont, P. S., and Etkin, L. D. (1997) J. Cell Sci. 110,

1563–157174. El-Husseini, A. E., Kwasnicka, D., Yamada, T., Hirohashi, S., and Vincent,

S. R. (2000) Biochem. Biophys. Res. Commun. 267, 906–91175. El-Hodiri, H. M., Che, S., Nelman-Gonzalez, M., Kuang, J., and Etkin, L. D.

(1997) J. Biol. Chem. 272, 20463–2047076. Holt, B. F. I., Mackey, D., and Dangl, J. L. (1999) Curr. Biol. 2000, R5–R777. Initiative, T. A. G. (2000) Nature 408, 796–815

TEFs: A Diverse Family of TRAF-containing Proteins24252

by guest on April 5, 2019

http://ww

w.jbc.org/

Dow

nloaded from

C. ReedJuan M. Zapata, Krzysztof Pawlowski, Elvira Haas, Carl F. Ware, Adam Godzik and John

Receptor-associated Factor DomainsA Diverse Family of Proteins Containing Tumor Necrosis Factor

doi: 10.1074/jbc.M100354200 originally published online February 21, 20012001, 276:24242-24252.J. Biol. Chem. 

  10.1074/jbc.M100354200Access the most updated version of this article at doi:

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/276/26/24242.full.html#ref-list-1

This article cites 77 references, 34 of which can be accessed free at

by guest on April 5, 2019

http://ww

w.jbc.org/

Dow

nloaded from