A novel protein (Fbf-1) that binds to CD95/APO-1/FAS and shows sequence similarity to trichohyalin and plectin

Short sequence-paper

A novel protein (Fbf-1) that binds to CD95/APO-1/FAS and showssequence similarity to trichohyalin and plectin

Thorsten Schmidt a;c, Holger Karsunky a, Beate FraM a, Wiebke Baum b,Angela Denzel d, Tarik Mo«ro«y a;*

a Institut fu«r Zellbiologie (Tumorforschung), IFZ, Universita«tsklinikum Essen, Virchowstrasse 173, D-45122 Essen, Germanyb Institut fu«r Molekularbiologie und Tumorforschung, Philipps Universita«t Marburg, Emil-Mannkop¡-Strasse 2, D-35033 Marburg, Germany

c Bayer AG, Apratherweg 18a, D-42096 Wuppertal, Germanyd Department of Immunobiology, Guy's Hospital, London SE1 9RT, UK

Received 8 March 2000; received in revised form 5 June 2000; accepted 13 June 2000

Abstract

The Fas/Apo-1/CD95 cell surface receptor belongs to the TNF receptor family of cell death inducing molecules. A number of cytosolicadapter proteins that mediate signal transduction of CD95 have been characterized, but some features of the molecular mechanisms ofCD95-induced cell death remain elusive. We describe here a novel protein that can interact with the cytosolic domain of the murine CD95receptor in a yeast two-hybrid assay. This novel protein was termed Fbf-1 for Fas binding factor and bears no sequence similarity to theknown CD95 adapter proteins. Fbf-1 is 1173 aa long and has a theoretical molecular weight of around 130 kDa. The protein is expressed ina wide variety of tissues and is localized in the cytoplasm. Fbf-1 is a very hydrophilic protein, highly conserved between mouse and humanand bears a carboxyterminal leucine heptad repeat reminiscent of leucine zipper protein interaction domains. In addition, it shows sequencesimilarity to trichohyalin and plectin pointing to a function as a structural protein. ß 2000 Elsevier Science B.V. All rights reserved.

Keywords: Apo-1/Fas/CD95; Trichohyalin

The CD95/Fas/Apo-1 receptor can e¤ciently mediateprogrammed cell death in response to various stimuli. Anumber of adapter proteins have been identi¢ed that bindto the cytoplasmic domain of CD95 and are necessary torelay the signals upon binding of CD95 ligand (reviewedin [1,2]). However, despite the discovery of these adaptermolecules, a number of aspects in CD95 signaling remainopen. For instance, most of the data in the literature havebeen gathered on the human receptor and not with themurine counterpart. The human CD95 sequence shows anumber of di¡erences compared to the mouse sequence. Inorder to gain new insights into the functional details of themouse CD95-mediated signaling, we have performed ayeast two-hybrid screen with the cytoplasmic domain ofthe murine receptor. The cDNAs from the library that wasused were expressed as fusions with the transcriptionaltransactivation domain of the herpes simplex virus VP16protein [3,4]. The library was tested with a bait construct

encoding the cytoplasmic domain of CD95 fused to theLexA DNA binding domain [4]. 17 yeast clones were re-covered that were able to overcome histidine de¢ciencyand showed L-galactosidase activity.

Upon retransfection of the isolated library plasmids,only one clone proved to be speci¢c for an interactionwith the LexA^CD95 fusion. This clone was subjected toa L-galactosidase assay with the original LexA^CD95 fu-sion and several mutants containing various parts ofCD95 (Fig. 1a). Only the entire cytoplasmic domain ofCD95 from amino acids 166^306 and one mutant compris-ing the region from aa 166^291 were able to interact withthe protein encoded by clone VP16/400 as indicated byL-galactosidase activity (Fig. 1a). Further deletions at theN- or C-terminal end of this cytoplasmic region of CD95abrogated this interaction completely (Fig. 1b). Interest-ingly, a LexA^CD95 fusion construct that bears the cyto-plasmic region with the Fas (lprcg) point mutation did notreact with the clone VP16/400. The Fas(lprcg) allele is anaturally occurring mutation in the mouse with the e¡ectthat amino acid 231 of CD95 is changed from isoleucineto asparagine [5]. The Fas (lprcg) mutant mouse has the

0167-4781 / 00 / $ ^ see front matter ß 2000 Elsevier Science B.V. All rights reserved.PII: S 0 1 6 7 - 4 7 8 1 ( 0 0 ) 0 0 1 6 3 - 9

* Corresponding author. Fax: +49-201-723-5904;E-mail : [email protected]

BBAEXP 91447 28-8-00 Cyaan Magenta Geel Zwart

Biochimica et Biophysica Acta 1493 (2000) 249^254

www.elsevier.com/locate/bba

same phenotype as the Fas (lpr, lpr) and is de¢cient forCD95 signaling. Mice bearing this mutation show abnor-mal accumulation of activated peripheral T-cells and lym-phadenopathy and an accompanying lupus-like autoim-mune disease with varying severity according to strainbackground [5]. The lack of interaction with the Fas(lprcg)mutant suggests a speci¢c interaction between the proteinencoded by the clone VP16/400 and the cytoplasmic do-main of the murine CD95.

The insert of clone VP 16/400 was used as a probe to

screen a murine thymus cDNA library and a cDNA cloneof 4087 bp was obtained. The largest open reading framehad the potential to encode a protein of 934 amino acids(not shown), but did not contain an in phase upstreamstop codon. Two longer cDNA clones with di¡erent 5P-ends were generated by extending the obtained cDNAfrom the library using a 5P-RACE protocol [6]. Both ofthese longer cDNA clones contained the same open read-ing frame now with an in phase 5P-stop codon and thepotential to encode a protein of 1173 aa extending the

Fig. 1. Identi¢cation of a yeast clone VP16/400 bearing a protein that interacts with mouse CD95. (a) Schematic representation of the CD95 coding re-gion (TM, transmembrane domain) and the di¡erent LexA^bait fusion constructs used in this analysis. Lprcg denotes a point mutation in the cytoplas-mic domain of CD95 that abrogates CD95 function. The area comprising the death domain of CD95 is also denoted. (b) L-galactosidase assay with theindicated bait constructs in yeast cells bearing the clone VP16/400. Only the full-length cytoplasmic domain and a C-terminally deleted derivative areable to interact with Fbf-1 in clone VP16/400.


T. Schmidt et al. / Biochimica et Biophysica Acta 1493 (2000) 249^254250

Fig. 2. Sequence and structural domains of Fbf-1. (a) Amino acid sequence of Fbf-1 according to the full-length cDNA. The sequence part containedin the original yeast clone VP 16/400 is shaded. The leucine zipper region is boxed. (b) Scheme of the Fbf-1 protein with the indication of the regionthat is required for an interaction with CD95, the amino acid stretch that is similar to trichohyalin in primary sequence and the location of the leucinezipper (LZ) in the C-terminal region of the protein. A Garnier^Robson as well as a Chou^Fasman analysis (LASERGENE Navigator) suggests anK-helical region in the C-terminal half of Fbf-1. The Kyte^Doolittle plot indicates high hydrophilicity. (c) Garnier^Robson predictions of K-helicalregions are given for rat plectin, sheep trichohyalin and murine Fbf-1. All three proteins share a hypothetical `rod-like' domain.


T. Schmidt et al. / Biochimica et Biophysica Acta 1493 (2000) 249^254 251

previous sequence by 239 residues (Fig. 2a). As the N-terminal methionine in the longer cDNA clones was nowpreceded by in frame stop codons, it can be concluded thatthe entire sequence of the new protein with a theoreticalmolecular weight of about 130 kDa had been obtained(Fig. 2a).

The novel protein was named Fbf-1 for Fas bindingfactor 1. Fbf-1 contains a leucine heptad repeat betweenaa 963 and 1013 and is a remarkably hydrophilic proteinaccording to the Kyte^Doolittle [9] hydrophilicity plot(Fig. 2b). In addition, both Chou^Fasman [7] and Gar-nier^Robson [8] calculations predict K-helical regions foralmost the entire C-terminal half of the protein from aa600 until the end (Fig. 2b). The protein has a theoreticalmolecular mass of 130 kDa and an isolelectric point at6.17. A BLAST search with the whole protein sequencefailed to produce a match in several libraries and thus didnot reveal the identity of the protein. However, a lowdegree of sequence similarity between Fbf-1 and the pro-teins trichohyalin and plectin was found. Stretches withhighest similarity localize between aa 600 and 1000(Figs. 2b and 3a). Similar to Fbf-1, both plectin and tri-chohyalin are highly hydrophilic proteins with extendedstretches of K-helical regions (Fig. 2c) which are supposedto form a rod-like structure [10^14]. The sequence encodedby the original yeast clone VP16/400 which represents theminimally required sequence for an interaction with thecytoplasmic tail of CD95 is contained in the C-terminalhalf of Fbf-1 and falls within the large K-helical region ofFbf-1 that resembles the trichohyalin/plectin rod-like do-main (Fig. 2a,b).

It is well established that the CD95/Apo-1/Fas receptorassociates with a protein termed FADD/Mort-1 on thecytoplasmic side after stimulation with CD95 ligand [15].This molecule had originally been isolated using the samestrategy as outlined here, namely with a cytoplasmic partof the CD95 receptor as a bait in a yeast two-hybrid screen[15]. It remains unclear why under the conditions appliedin the present study a clone coding for FADD was notisolated. One reason for this could be that the bait usedto pull FADD contained the human CD95 sequence andthat the murine sequence used here which is very similar toits human counterpart, albeit not identical, may form lessstable complexes with the murine FADD in yeast.

Database searching with the cDNA sequence of Fbf-1turned up an 84% match on the sequence of an unidenti-¢ed human EST clone (accession number H30685). Theclone was acquired through the IMAGE consortium andwas entirely sequenced. The open reading frame was con-ceptually translated and showed a high degree of aminoacid sequence homology to the murine Fbf-1, suggestingthat the human EST indeed represents the human Fbf-1protein (Fig. 3b). In addition, this indicated that Fbf-1 iswell conserved between both human and mouse and thatin both species Fbf-1 may exert similar functions.

Northern blot analysis revealed that the murine Fbf-1

Fig. 3. Sequence similarities of Fbf-1. (a) The region with highest simi-larity between Fbf-1 and trichohyalin is depicted and identical aminoacid residues are shaded. (b) Sequence conservation between murine andhuman Fbf-1. The murine sequence is given underneath the human se-quence which was derived from a human EST clone. Identical aminoacids are shaded.



gene is expressed in a variety of tissues, such as heart,brain, spleen, lung, liver, skeletal muscle, kidney and testesas an mRNA of about 5.5 kb (Fig. 4a). The longest cDNAthat was isolated had a length of 5307 nucleotides andvery probably represents the transcripts found in mosttissues. A slightly longer transcript is found in heart andadditional smaller transcripts appear to be present in testes(Fig. 4). In vitro transcription/translation of a constructcontaining the Fbf-1 open reading frame produced a singleprotein with an apparent molecular weight somewhatabove the 130-kDa calculated from the primary aminoacids sequence (Fig. 4b). To detect Fbf-1 in Western blotsand to be able to identify its subcellular localization, wehave generated a fusion construct between the entire openreading frame of Fbf-1 and of the green £uorescent pro-tein (GFP). NIH 3T3 cells were transiently transfectedwith this construct and were analyzed by immunoblottingand confocal microscopy. The Fbf-1-GFP tagged proteinwas found to be expressed as a single protein with anapparent molecular weight in an SDS^PAGE larger than175 kDa. The theoretical molecular weight of a Fbf-1-

GFP fusion protein would be 156.5 kDa. The higher ap-parent molecular weight obtained in both the in vitrotranslation product and upon transient expression in eu-karyotic cells suggests posttranslational modi¢cation ofthe protein. In addition, Fbf-1 was found to be localizedubiquitously in the cytoplasm, but was particularly en-riched within a dotted structure (Fig. 4c).

More striking than the actual amino acid sequence sim-ilarity between the protein Fbf-1 and both trichohyalinand plectin is their shared domain structure. Plectin hasa long central rod domain with an K-helical coiled-coilconformation [12] £anked by globular N- and C-terminaldomains. Trichohyalin has a very similar rod structurecomprising several regions of the protein (Fig. 2c) and ispredicted to form an elongated £exible rod of at least 215nm [14]. Both trichohyalin and plectin can bind to inter-mediate ¢laments and are thought to play a role in theorganization of the part of the cytoskeleton that is respon-sible for the inner cell envelope. The Fbf-1 protein identi-¢ed here by virtue of its association with the cytoplasmictail of CD95 in a yeast two-hybrid assay shows a very

Fig. 4. Expression of Fbf-1 and subcellular localization. (a) Multiple tissue Northern (Clontech, Heidelberg) hybridized with a Fbf-1 probe. Fbf-1 is en-coded by a mRNA of about 5.5 kb which is present in all tissues tested, but in di¡erent isoforms in heart and testes. RNA loading was controlled by a hy-bridization with a GAPDH probe. Lanes: 1, heart ; 2, brain; 3, spleen; 4, lung; 5, liver ; 6, skeletal muscle; 7, kidney; 8, testes. (b) In vitro transcription/translation of the total Fbf-1 cDNA yields a single protein. (c) NIH 3T3 cells were transiently transfected with a GFP-tagged Fbf-1 full-length cDNA (inpcDNA 3) or a construct directing the expression of GFP alone. Extracts from these cells were analyzed by Western blotting using an anti-GFP antibody.Transient transfection of the Fbf-1-GFP construct leads to expression of a single protein with an apparent molecular weight larger than 175 kDa. GFPalone is detected at 32 kDa. (d) Fbf-1 protein is localized in the cytoplasm. Left and middle panel : NIH 3T3 cells transfected with GFP tagged Fbf-1.Right panel : transfection of the empty vector control. Cells were ¢xed and stained with propidium iodide for visualization of the nucleus (red).


T. Schmidt et al. / Biochimica et Biophysica Acta 1493 (2000) 249^254 253

prominent and extended K-helical region which could rep-resent a rod-like domain as it is found in plectin andtrichohyalin. Although further studies are clearly required,these structural similarities point to the possibility thatFbf-1 is a linker or sca¡old protein that provides a con-nection between the cytoplasmic tail of CD95 and the cy-toskeleton.

For the two-hybrid assay, a fragment of the murineCD95 c-DNA encoding the cytoplasmic part of the recep-tor was ampli¢ed by PCR, entirely sequenced and clonedinto the pBTM116 vector [3]. The expression library wasmade with randomly primed, size selected cDNA frag-ments derived from 9.5- to 10.5-day-old mouse embryoRNA. The cDNA fragments were ampli¢ed by PCR andinserted into the pVP16 vector [3]. For the screening ofthis c-DNA library by histidine selection in yeast, the Li-acetate transfection method was used. LexA^bait fusionconstructs and the library plasmid were introduced intothe strain L40 and clones were selected for growth onhistidine de¢cient plates containing 25^50 mM aminotria-zole and for L-galactosidase production. Interactions werecon¢rmed by retransfecting isolated library plasmids intoyeast cells containing the LexA^bait fusion vector andtesting the L-galactosidase activity in a liquid assay.

For the L-galactosidase assay, yeast cells were grown toa OD600 (OD at 600 nm) of 0.4^0.5, pelleted and resus-pended in 1.2 ml bu¡er Z (60 mM Na2HPO4, 40 mMNaH2PO4, 10 mM KCl, 1 mM KgSO4, 50 mM L-ME,ph7.0). From this suspension, 100^150 Wl (VE) were di-luted to 1 ml with bu¡er Z and cells were lysed withchloroform/SDS. The mixture was equilibrated at 30³Cfor a few minutes and the reaction was started by addingof 200 Wl ONPG (4 mg/ml O-nitrophenol galactose in 0.1M K2PO4, pH 7.0). After a de¢ned reaction period (tR,e.g. 2^3 h) the reaction was stopped by the addition of 500Wl 1 M Na2CO3. The reaction were cleared of insolublematerial by centrifugation and OD at 420 nm and at 550nm was measured. The L-galactosidase activity (A) wascalculated in units by the following equation:

A � 1000UOD42031750UOD550=OD600UVEUtR:

NIH 3T3 cells were electroporated in DMEM with upto 20 Wg plasmid DNA at 240 V and 500 WF (Bio-Radelectroporator) and harvested 1^2 days later. For the de-tection of GFP protein, 105 cells were seeded in 6-cmdishes on cover slips and 24 h later transfected accordingto the calcium phosphate method. The next day, cells werewashed with PBS, ¢xed with paraformaldehyde, againwashed with PBS and stained with propidium iodide.

In vitro translation was performed in a single-step tran-scription/translation system using the `TNT coupled retic-ulocyte lysate' (Promega) according to the supplier's infor-mation. The reaction contained 1 Wg plasmid DNA and 40WCi 35S-labeled methionine in a total volume of 50 Wl.

Detection of the mRNA for Fbf-1 has been performedusing a multiple tissue Northern blot from Clontech. The

blot contains approximately 2 Wg of poly A� RNA perlane from several di¡erent murine tissues. RNA was sep-arated on a denaturating formaldehyde 1.2% agarose gel,transferred to a charge-modi¢ed nylon membrane bycapillary transfer and was ¢xed by UV irradiation. Hy-bridization and washing was performed as suggested bythe manufacturer. The Fbf-1 probe was labelled as previ-ously described [16].

All sequence analysis was performed using the softwarefrom LASERGENE navigator.

We thank H.P. Elsa«sser, Marburg, for help with confo-cal microscopy. This work was supported by a grant fromthe Deutsche Forschungsgemeinschaft (Mo 435/5-1) andthe `Fond der chemischen Industrie'.

References

[1] P.H. Krammer, CD95(APO-1/Fas)-mediated apoptosis : live and letdie, Adv. Immunol. 71 (1999) 163^210.

[2] M.E. Peter, P.H. Krammer, Mechanisms of CD95 (APO-1/Fas)-mediated apoptosis, Curr. Opin. Immunol. 5 (1998) 545^551.

[3] S.M. Hollenberg, R. Sternglanz, P.F. Cheng, H. Weintraub, Identi-¢cation of a new family of tissue-speci¢c basic helix-loop-helix pro-teins with a two-hybrid system, Mol. Cell. Biol. 7 (1995) 3813^3822.

[4] S. Fields, O. Song, A novel genetic system to detect protein^proteininteractions, Nature 340 (1989) 245^246.

[5] M. Kimura, A. Matsuzawa, Autoimmunity in mice bearing lprcg : anovel mutant gene, Int. Rev. Immunol. 3 (1994) 193^210.

[6] M.A. Frohmann, M.K. Dush, G.R. Martin, Rapid production offull-length cDNAs from rare transcripts: ampli¢cation using a singlegene-speci¢c oligonucleotide primer, Proc. Natl. Acad. Sci. USA 85(1988) 8992^9002.

[7] P.Y. Chou, G.D. Fasman, Empirical predictions of protein confor-mation, Annu. Rev. Biochem. 47 (1978) 251^276.

[8] J. Garnier, J.F. Gibrat, B. Robson, GOR method for predictingprotein secondary structure from amino acid sequence, Methods En-zymol. 266 (1996) 540^553.

[9] J. Kyte, R.F. Doolittle, A simple method for displaying the hydro-pathic character of a protein, J. Mol. Biol. 157 (1982) 105^132.

[10] M.J. Fietz, C.J. McLaughlan, M.T. Campbell, G.E. Rogers, Analysisof the sheep trichohyalin gene: potential structural and calcium-bind-ing roles of trichohyalin in the hair follicle, J. Cell Biol. 4 (1993) 855^865.

[11] G.E. Rogers, M.J. Fietz, A. Fratini, Trichohyalin and matrix pro-teins, Ann. New York Acad. Sci. 642 (1991) 64^80.

[12] G. Wiche, B. Becker, K. Luber, G. Weitzer, M.J. Castanon, R.Hauptmann, C. Stratowa, M. Stewart, Cloning and sequencing ofrat plectin indicates a 466-kD polypeptide chain with a three-domainstructure based on a central alpha-helical coiled coil, J. Cell Biol. 114(1991) 83^99.

[13] F.A. Steinbock, G. Wiche, Plectin: a cytolinker of design, Biol.Chem. 380 (1999) 151^158.

[14] S.C. Lee, I.G. Kim, L.N. Marekov, E.J. O'Keefe, D.A. Parry, P.M.Steinert, The structure of human trichihyalin. Potential multiple rolesas a functiona EF-hand-like calcium binding protein, a corni¢ed cellenvelope precursor and an intermediate ¢lament-associated (cross-linking) protein, J. Biol. Chem. 268 (16) (1993) 12164^12167.

[15] A.M. Chinnaiyan, K. O'Rourke, M. Tewari, V.M. Dixit, FADD, anovel death domain-containing protein, interacts with the death do-main of Fas and initiates apoptosis, Cell 81 (1995) 505^512.

[16] J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular Cloning: A Lab-oratory Manual, 2nd edn., Cold Spring Harbor Laboratory Press,Cold Spring Harbor, NY, 1998.



Documents

A novel protein (Fbf-1) that binds to CD95/APO-1/FAS and shows sequence similarity to trichohyalin and plectin