Membrane-Proximal Cytoplasmic Domain of Fas Kinases That

Phosphorylation of FADD/MORT1 and Fas by Kinases That Associate with the Membrane-Proximal Cytoplasmic Domain of Fas

This information is current as Norman J. Kennedy and Ralph C. Budd of October 1, 2021. J Immunol 1998; 160:4881-4888; ; http://www.jimmunol.org/content/160/10/4881

References This article cites 49 articles, 22 of which you can access for free at: Downloaded from http://www.jimmunol.org/content/160/10/4881.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision http://www.jimmunol.org/

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: by guest on October 1, 2021 http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 1998 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Phosphorylation of FADD/MORT1 and Fas by Kinases That Associate with the Membrane-Proximal Cytoplasmic Domain of Fas1

Norman J. Kennedy and Ralph C. Budd2

Fas (Apo-1, CD95), a member of the TNFR family, is expressed on a variety of cell types and transduces an apoptotic signal. Since Fas does not possess known enzymatic activities, proteins that interact with the cytoplasmic domain of Fas regulate the death signal. Several proteins have been identified, primarily using the yeast two-hybrid system, that associate with the death domain of Fas. One of these proteins, FADD/MORT1, can be phosphorylated, although the kinase that is responsible has not been identified. Furthermore, direct signaling connections between Fas and its known activation of sphingomyelinase or NF-␬B have not been made, suggesting that other proteins may associate with Fas. In this study, a series of glutathione S-transferase fusion Downloaded from proteins was constructed that contained the cytoplasmic domain of murine Fas. These proteins were used to search for additional proteins that associate with Fas. Novel proteins, including kinases, were identified that associated specifically with the membrane- proximal, cytoplasmic tail of Fas but not with the death domain. One of these kinases phosphorylates FADD/MORT1. Moreover, the membrane-proximal region of Fas itself was phosphorylated by one of the associating kinases. These findings suggest that, similar to the Fas-related p55 TNFR, the membrane-proximal region of Fas likely participates in signaling. The Journal of Immunology, 1998, 160: 4881–4888. http://www.jimmunol.org/

as (APO-1, CD95) is a 45-kDa surface protein that is ex- (21, 22), and the activation of IL-1␤-converting enzyme (ICE)3- pressed on a wide array of cell types and is capable of like proteases (23, 24). Other recent reports of several Fas-asso- F inducing apoptosis (1–3). It is a member of an extensive ciating proteins are separate from the findings examining known family of cell surface molecules that includes the p55 and p75 signaling pathways. Most of these have been identified using the forms of TNFR, the p75 neurotrophin receptor, CD27, CD30, yeast two-hybrid system and include FADD/MORT1 (25, 26), re- CD40, OX40, death receptor 3, and death receptor 4 (4–8). The ceptor-interacting protein (RIP) (27), Fas-associated protein homology within this receptor family resides primarily in the ex- factor-1 (FAF1) (28), Fas-associated phosphatase (FAP-1) (29), tracellular portion of the molecule that contains a series of cys- and Daxx (17). Additional Fas-associating proteins were identified by guest on October 1, 2021 teine-rich repeats (4, 9). The cytoplasmic regions are more diverse, using either coimmunoprecipitation (30) or glutathione S-transferase (GST) fusion proteins containing the cytoplasmic domain of although Fas, death receptor 3, death receptor 4, and the p55 TNFR Fas (31). known as TNFR1 bear a similar 75-amino acid (aa) death domain Given the panel of Fas-associating proteins identified to date, it that is critical to the signaling of apoptosis (6–8, 10, 11). A single is difficult to make direct connections with many of the proposed base pair (bp) mutation in the Fas death domain occurs in lprcg Fas signaling pathways. Exceptions are the protease cascade, in mice which blocks the ability of Fas to initiate apoptosis (12). cg which an ICE-like protease, FADD-like ICE (FLICE)/MACH was Lpr mice manifest profound lymphadenopathy and a systemic demonstrated to associate with FADD (32, 33), and the c-Jun N- autoimmune disease, attesting to the importance of Fas in normal terminal kinase pathway, which was recently shown to be activated lymphoid homeostasis (12). by Daxx (17). Additional concerns are raised by the fact that the Fas reportedly signals via a variety of pathways that potentially five reports using the yeast two-hybrid approach have identified include Ras/Raf-1/extracellular signal-related kinase (13), c-Jun five different Fas-binding proteins. This suggests that additional N-terminal kinase (14–17), sphingomyelinase (13, 18–20), NF-␬B proteins likely associate with Fas. Given these diverse findings and the lack of a link between Fas and several nonprotease signal pathways, we decided to look for other proteins that bind to Fas. Since trimerization of Fas is required for signaling apoptosis (34) and may consequently be necessary for the binding of proteins that regulate death, we were concerned that the yeast two-hybrid sys- Department of Medicine, Immunobiology Program, The University of Vermont Col- tem might not provide the necessary oligomerization. Instead, the lege of Medicine, Burlington, VT 05405 GST fusion protein system was applied. A series of GST-murine Received for publication September 29, 1997. Accepted for publication January Fas variants was generated. Using these constructs, two kinase 26, 1998. activities and two novel proteins were identified that associate with The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 3 Abbreviations used in this paper: ICE, IL-1␤-converting enzyme; DD, death do- 1 This work was supported by Grant AR42489 from the National Institutes of Health. main; FADD, Fas-associating death domain protein; FLICE, FADD-like IL-1␤-con- 2 Address correspondence and reprint requests to Dr. Ralph C. Budd, Department of verting enzyme; RIP, receptor-interacting protein; FAP, Fas-associated phosphatase; Medicine, Immunobiology Program, Given Medical Building D301, The University FAF, Fas-associated protein factor; CAP, cytotoxicity-dependent APO-1-associated of Vermont College of Medicine, Burlington, VT 05405-0068. E-mail address: protein; aa, amino acid; bp, base pair; GST, glutathione S-transferase; MBP, myelin [email protected] basic protein; WTFas, wild-type Fas.

the cytoplasmic tail of Fas. Additionally, one of the kinases phos- pared as follows: 107 cells were washed twice in 5 ml PBS and lysed on ice phorylated FADD. Interestingly, these proteins do not bind to the for 20 min in 1.0 ml of ice-cold lysis buffer (0.5% Nonidet P-40, 20 mM ␮ death domain, but rather map to the membrane-proximal portion of Tris-Cl (pH 7.7), and 200 mM NaCl) containing 2 g/ml each of antipain, aprotinin, and leupeptin, 1 mM benzamidine, 1 mM PMSF, 10 mM NaF, the cytoplasmic tail of Fas. and1mMNa3VO4. Following lysis, samples were centrifuged at 13,000 rpm for 10 min at 4°C. The supernatants were mixed with immobilized Materials and Methods GST-Fas fusion proteins on ice for 2 h with agitation to allow EL4 proteins Cell culture to associate with Fas. Samples were washed four times with 1.0 ml of ice-cold lysis buffer containing salt concentrations that varied according to The mouse thymoma cell line EL4 and the mastocytoma cell line P815 the experiment, as described in Results. Samples were then washed once were cultured in RPMI 1640 supplemented with 25 mM HEPES, 5% bo- with 1.0 ml of ice-cold 50 mM Tris-HCl (pH 8), resuspended in SDS vine calf serum, 2 mM glutamine, 50 ␮M 2-ME, 1 mM pyruvate, 0.01 sample buffer, and separated on 10 to 12.5% polyacrylamide gels by SDS- mg/ml folate, 2.5 mg/ml glucose, 100 U/ml penicillin, and 100 ␮g/ml PAGE. Fas-associated proteins were visualized with silver stain (Bio-Rad, streptomycin. Hercules, CA). Construction of GST expression vectors In vitro kinase assays mRNA from splenocytes of MRL ϩ/ϩ or lprcg mice (courtesy of Dr. EL4 lysates were incubated with GST-Fas fusion proteins that had been Michael Seldin, Duke University, Durham, NC) was isolated by the Trizol immobilized on glutathione Sepharose as described above, except 0.1 method (Life Technologies, Gaithersburg, MD), and cDNA was prepared ␮g/ml microcystin was added to the lysis buffer. In vitro kinase assays by oligo(dT) priming and reverse transcriptase. The cytoplasmic domains were performed as previously described (36) with modiﬁcations. In brief, of wild-type Fas (WTFas), lprcg Fas, and each deletion variant were am- samples were washed three times with 1.0 ml of ice-cold lysis buffer con- cg Ј ␮ pliﬁed by PCR. WTFas and lpr Fas were obtained using common 5 taining 2 mM 2-ME, 10 mM NaF, 1 mM Na3VO4, and 0.1 g/ml micro- (GACGGATCCCGAAAGTACCGGAAAAGAAAGTGC) and 3Ј (CCG cystin and then washed twice with 1.0 ml of ice-cold kinase wash buffer Downloaded from

AATTCCCTCCAGACATTGTCCTTCATT) primers. FD5 (126 aa lack- (20 mM HEPES (pH 7.4), 2 mM 2-ME, 10 mM NaF, 1 mM Na3VO4, and ing the C-terminal negative regulatory domain) and FD1 (membrane- 0.1 ␮g/ml microcystin). Samples were resuspended in 30 ␮l of kinase Ј proximal 51 aa) were obtained using both the same 5 primer that was used assay buffer (20 mM HEPES (pH 7.4), 10 mM MgCl2, 2 mM 2-ME, 10 Ј ␮ to amplify full-length, cytoplasmic Fas and the unique 3 primers FD5 mM NaF, 1 mM Na3VO4, and 0.1 g/ml microcystin) and incubated for 30 (CCGAATTCTCATTTTCCAAGGTCCTTCTGGAC) and FD1 (CC min at 30°C with 5 ␮Ci [␥-32P]ATP (NEN/DuPont, Boston, MA). Where GAATTCTCAAGCTTCCTGGATTGTCATGTC). The death domains (75 indicated, 10 ␮g of myelin basic protein (MBP) (a general substrate for aa) from MRL ϩ/ϩ or lprcg were amplified using a common 5Ј (GACG serine/threonine kinase activity) or GST-FADD (courtesy of Dr. Astar http://www.jimmunol.org/ GATCCAAAAAATTTGCTCGAGAAAATAAC) primer, and the 3Ј Winoto, University of California at Berkeley, Berkeley, CA) was added to primer was used to amplify FD5. Primers were obtained commercially the kinase reaction. Samples were separated on a 12.5% polyacrylamide (Genosys Biotechnologies, The Woodlands, TX). PCR amplifications were gel by SDS-PAGE and exposed to film for 30 to 60 min at room performed as follows: 94°C ϫ 2 min, 50°C ϫ 1 min, 72°C ϫ 1 min (1 temperature. cycle); 94°C ϫ 45 s, 50°C ϫ 1 min, 72°C ϫ 1 min (35 cycles); and 94°C ϫ 1 min, 50°C ϫ 1 min, 72°C ϫ 5 min (1 cycle). The amplified products Thrombin cleavage assay were gel-purified and extracted by excising the specific band, crushing the The pGEX2T vector encodes a thrombin cleavage site at the 3Ј end of GST, agarose fragment in a syringe, and eluting the product in TE (10 mM enabling separation of the Fas moiety from GST. As described above, EL4 Tris-HCl, pH 8, and 1 mM EDTA) buffer. The DNA was washed and lysates were incubated with GST-Fas fusion proteins that had been immo- concentrated on a Centricon-50 column (Amicon, Beverly, MA) and then bilized on glutathione Sepharose. Before separation by SDS-PAGE, sam- digested with BamHI and EcoRI restriction endonucleases. Enzymes were by guest on October 1, 2021 ples were subjected to thrombin cleavage (0.01 U/␮l) as previously de- heat-inactivated and removed by extensive washing in a size-exclusion scribed (31). To determine whether Fas was phosphorylated, samples were Centricon column. Purified DNA was ligated into the BamHI/EcoRI re- subjected to thrombin cleavage after an in vitro kinase assay was striction sites of the expression vector pGEX2T (Pharmacia, Piscataway, conducted. NJ). Ligations were performed at 12°C for 16 h with T4 ligase (Life Tech- nologies). Ligated plasmids were transfected into the Escherichia coli In-gel kinase assays strain INV␣FЈ (One-Shot Cloning Kit, Invitrogen, San Diego, CA). E. coli colonies were selected on 2x YT/agar plates containing 100 ␮g/ml ampi- EL4 lysates were incubated with GST-Fas fusion proteins immobilized on cillin. Ampicillin-resistant colonies were selected and expanded in 2x YT glutathione Sepharose as described above, except 0.1 ␮g/ml microcystin broth with ampicillin, and the plasmid DNA was isolated (Qiagen, Chats- was added to the lysis buffer. Fusion protein pull-down complexes were worth, CA). Plasmid DNA was digested with BamHI and EcoRI, separated washed three times with 1.0 ml of ice-cold lysis buffer containing 10 mM ␮ by gel electrophoresis, and visualized with ethidium bromide. Colonies NaF,1mMNa3VO4, and 0.1 g/ml microcystin and then washed twice determined to have insert DNA of the appropriate size were further with 1.0 ml of ice-cold 50 mM Tris (pH 8). In-gel kinase assays were screened for inducible fusion protein production by SDS-PAGE and Coo- performed as previously described (37) using 0.4 mg/ml MBP, GST, or massie blue staining. In addition, plasmid DNA from the WTFas and lprcg GST-FADD as a substrate. constructs was sequenced using a primer that was just 5Ј from the GST- fusion protein splice site. Sequencing confirmed that the constructs were Phosphoamino acid analysis cg 786 in-frame and that lpr did contain the single-bp change (T3A ). The EL4 proteins associated with immobilized GST-WTFas proteins were Expression of GST fusion proteins subjected to an in vitro kinase assay as described above. Following SDS- PAGE separation, proteins were transferred to a polyvinylidene difluoride Cultures of E. coli-bearing expression plasmids were grown to confluency membrane. The radioactive band representing GST-WTFas was cut from at 37°C in 60 ml of 2x YT broth containing 100 ␮g/ml ampicillin. Cultures the membrane, hydrolyzed in 6 M HCl for 60 min at 110°C, and dried were diluted to a final volume of 200 ml with 2x YT containing 100 ␮g/ml under vacuum. Two-dimensional phosphoamino acid analysis was per- ampicillin and then incubated for7hat22°C with 0.1 mM isopropyl formed as previously described (38). ␤-D-thiogalactoside to induce fusion protein production. Fusion proteins were purified as previously described (35). Each batch of fusion protein Results supernatant was titered to determine the volume required to saturate 15 ␮l of glutathione Sepharose beads (Pharmacia) according to Coomassie blue To investigate the association of proteins with the intracellular staining of fusion proteins separated by SDS-PAGE. Titered supernatants domain of murine Fas, variants of the Fas cytoplasmic tail were were stored at Ϫ70°C until needed. constructed as GST fusion proteins. These included full-length cg Identification of Fas-associating proteins WTFas, four truncations, and the single-bp mutation in lpr (Fig. 1). WTFas consists of 141-aa residues containing a 15-aa C-ter- GST-Fas fusion protein supernatants were incubated with 15 ␮l of gluta- minal negative regulatory domain (10), a 75-aa death domain (39), thione Sepharose in a total volume of 1.0 ml for 30 min on ice with agi- and a 51-aa membrane-proximal domain. Lprcg contains a single tation. The immobilized fusion proteins were washed once with 1.0 ml of 225 ice-cold high salt STE (1 M NaCl, 10 mM Tris-HCl (pH 8), and 1 mM aa substitution (Ile3Asn ) in the death domain that prevents the EDTA) and once with 1.0 ml of ice-cold PBS. EL4 cell lysates were pre- transduction of an apoptotic signal (12). FD5 lacks the C-terminal, The Journal of Immunology 4883

FIGURE 1. Schematic representation of GST-Fas fusion proteins containing the cytoplasmic domain of murine Fas. GST-Fas fusion proteins were generated as described in Materials and Methods. The cytoplasmic tail of WTFas spans aa residues 166–306 (5, 10). The Fas variant, lprcg, contains a single point mutation in the death domain (T3A786) that results in a single aa substitution at position 225 (Ile3Asn) (12). The death domain is deﬁned as residues 217–291 of the Fas cytoplasmic domain (38) and is required for transducing the apoptotic signal. FD5 lacks the negative regulatory domain that spans residues 192–306 (10). The membrane-proximal domain (FD1) includes aa 166–216. Downloaded from

negative regulatory domain (10), while FD1 contains only the activity. Figure 3A illustrates a representative in vitro kinase assay membrane-proximal region. Also, the death domain of WTFas using MBP as a substrate. A kinase activity that phosphorylated cg cg cg

(DD) or lpr (lpr DD) was linked to GST. These constructs were MBP was associated with WTFas, lpr , FD1, and FD5 (lanes 7, http://www.jimmunol.org/ expressed and used to purify adherent proteins with glutathione 15, 19, and data not shown). However, this kinase activity was not Sepharose. observed in association with GST alone, DD (lanes 4 and 12), GST-outer surface protein A, or in the absence of lysate (data not Novel proteins associate with the membrane-proximal shown). cytoplasmic domain of Fas We also observed phosphorylation of the intracellular domain of Figure 2A shows a silver stain of Fas-associating proteins resolved Fas itself, as evidenced by phosphorylation of some of the Fas from lysates of Fasϩ murine EL4 thymoma cells by SDS-PAGE. fusion proteins. Minimally, the membrane-proximal region (Fig. In the absence of EL4 lysate, fusion proteins of the predicted mo- 3A, lanes 14–16), but not the death domain (lane 11), was re- lecular mass were observed (lanes 2–6). In addition, some break- quired for binding the kinase(s) capable of phosphorylating Fas. by guest on October 1, 2021 down products were seen, as reflected by the presence of 26-kDa GST itself was not a substrate for the kinase associating with Fas GST. In the presence of EL4 lysate, two predominant proteins with (Fig. 3A, lanes 8, 9, 16, and 20), nor did GST bind the kinase that a molecular mass of 45 kDa (p45) and 33 kDa (p33), respectively, phosphorylates Fas (data not shown). Thrombin cleavage of were observed to associate with several of the Fas constructs. p45 WTFas revealed all phosphorylation to be on the 18-kDa Fas moi- associated specifically with WTFas, lprcg, FD5, and FD1 (lanes ety and not on the 26-kDa GST component (Fig. 3B, lanes 1 and 8–11), but not with GST alone (lane 7), DD (lane 12), or lprcg DD 2). Eliminating Mg2ϩ from the reaction to block kinase activity (lane 13). p33 associated specifically with WTFas, lprcg, and FD5 abolished all phosphorylation (Fig. 3B, lane 3). Thus, phosphor- (lanes 8–10), but not with GST alone (lane 7). As p33 migrated ylation of Fas was not merely due to the nonspecific adherence of to nearly the same area as FD1 and DD, these fusion proteins were [32P]ATP. Two-dimensional phosphoamino acid analysis of cleaved with thrombin after they had bound associated proteins. WTFas primarily revealed phosphorylated serine residues with a Figure 2B shows that p33 did associate with FD1 (lane 4) but not lesser amount of threonine phosphorylation (Fig. 3C). No tyrosine with DD (lane 6)orlprcg DD (data not shown). These proteins phosphorylation was observed, even after prolonged exposures. also did not associate with two additional fusion proteins, GST-c- FADD is a Fas-binding protein that recruits the ICE-like pro- Jun and GST-outer surface protein A (of Borrelia burgdorferi) tease, FLICE, to Fas, forming a death-inducing signaling com- (data not shown). These findings were consistent with results using plex (or DISC) (30, 32, 33). Furthermore, FADD can be phos- both EL4 lysates metabolically labeled with [35S]methionine/cys- phorylated in vivo, although the kinase responsible has not been teine and the murine mastocytoma line P815 (data not shown). p45 identified (30, 40). To determine whether a FADD kinase ac- bound cytoplasmic Fas with high avidity, as revealed by its con- tivity could associate with GST-Fas, in vitro kinase assays were tinued association in high salt (1 M NaCl) (Fig. 2A, lanes 14–17). performed on Fas-associating EL4 or P815 proteins using GST- Furthermore, p45 appeared to associate most strongly with the FADD as the substrate. A kinase activity that phosphorylated membrane-proximal domain, FD1 (Fig. 2A, lane 17). Under high GST-FADD associated best with FD5 (Fig. 3D, lanes 5 and 10), salt conditions, p33 bound most avidly to lprcg (Fig. 2A, lane 15). but also associated with WTFas and lprcg (lanes 3, 6, and 8). In Other proteins were mostly dissociated from Fas after high-salt contrast to the MBP kinase activity, FD1 did not support the washes. phosphorylation of GST-FADD (lanes 4 and 9), suggesting that the MBP and FADD kinase activities are different. The FADD Kinases that associate with the membrane-proximal region of kinase activity did not associate with either GST alone or DD Fas phosphorylate MBP, Fas, and FADD (lanes 1 and 7). Quantitation of phosphorylated GST-FADD An alternate approach to examining Fas-associating proteins is the confirmed the observation that the FADD kinase activity asso- assessment of function, such as phosphorylation. Consequently, ciated best with FD5 (Fig. 3E). Normalization of GST-FADD we studied the ability of the associating protein(s) to display kinase phosphorylation to the phosphorylation of the corresponding 4884 ASSOCIATION OF FADD KINASE ACTIVITY WITH Fas

FIGURE 2. Silver stain of proteins that bind spe- ciﬁcally to the cytoplasmic domain of Fas. GST-Fas fusion proteins were prepared and incubated with lysates from EL4 cells as described in Materials and Methods. The silver stains of GST-Fas fusion proteins and Fas-associating EL4 proteins resolved by SDS-PAGE are shown. A, Lane 1 contains molecular mass standards. Lanes 2–6 show fusion proteins of the predicted molecular mass in the absence of EL4 cell lysates. Fusion proteins incubated in the presence of lysates from 107 EL4 cells are shown in lanes 7–17. Before electrophoretic separation, samples were washed four times with 1.0 ml of lysis buffer containing either low salt (0.2 M NaCl (lanes 2–13)) or high salt (1.0 M NaCl (lanes 14–17)). The arrows Downloaded from indicate p33 and p45 as EL4 proteins that associate speciﬁcally with WTFas, lprcg, FD1, and FD5, but not with GST alone, DD, or lprcg DD. B, Samples were incubated in the presence (lanes 2, 4, and 6)or absence (lanes 1, 3, and 5) of thrombin as described in Materials and Methods to cleave the fusion pro- http://www.jimmunol.org/ teins. Thrombin cleavage reveals the association of p33 with FD1 (lane 4). DD does not bind p33 (lane 6). WTFas (lane 2) was included as a positive control. by guest on October 1, 2021

fusion protein gave comparable results. Fas-associating kinase 9). This might result from a greater activity of p43 when asso- activity could not be demonstrated for the substrates histone H1 ciated with FD5 or from quantitatively more p43 binding to (Fig. 3D, lanes 11 and 12), histone H2B, or enolase (data not FD5 than FD1. Figure 4B also shows other kinase activities shown). The stimulation of cells with Fas ligand-transfected ranging from 50 to 100 kDa in size. However, these activities fibroblasts (a gift from Dr. I. N. Crispe, Yale University, New do not appear to be specific for Fas, as they also associate with Haven, CT) did not alter any of the observed kinase activities GST (Fig. 4B, lane 1). (data not shown). Using GST-FADD as the in-gel kinase substrate also identified a 43-kDa kinase. Consistent with the findings using MBP as In-gel kinase assays demonstrate high molecular mass MBP the substrate, the 43-kDa FADD kinase activity was also ob- kinases and a 43-kDa FADD kinase served to associate particularly well with FD5 (Fig. 4D, lane 7). To characterize the size of the Fas-associating kinase(s) that is This kinase specifically phosphorylated FADD, as considerably responsible for phosphorylating MBP and FADD, in-gel kinase more phosphorylation was seen using GST-FADD than GST assays were performed using MBP or GST-FADD as substrates. alone (compare lane 7 in Fig. 4, C and D). Importantly, unlike Figure 4B shows pronounced MBP kinase activities migrating MBP, no FADD kinase activity that could be attributable to at 105 and 120 kDa (p105, p120). These activities associated higher molecular mass kinases was observed. This further sup- most strongly with FD5 and FD1 (lanes 7 and 9). The kinase ports the earlier notion that different kinases phosphorylate was not observed to associate with GST alone (lane 1) or with MBP and FADD. DD (lane 5), nor was it present in the absence of EL4 lysate As the 43-kDa FADD kinase and GST-FD5 migrate to similar (lanes 2, 4, 6, and 8). Weak 105-/120-kDa kinase activities areas in SDS gels, FADD kinase activity may be obscured by were observed in the absence of MBP, suggesting that these the nonspecific adherence of [32P]ATP to FD5. To better reveal kinases can undergo autophosphorylation (Fig. 4A, lanes 3, 7, the FADD kinase activity, Fas-associating EL4 proteins were and 9). Additionally, FD5 and FD1 associated with a 43-kDa treated with thrombin to cleave fusion proteins before electro- autophosphorylating kinase that was capable of phosphorylat- phoretic separation on gels containing GST-FADD. The 43-kDa ing MBP (Fig. 4B, lanes 7 and 9). However, phosphorylation of FADD kinase activity was far more pronounced when copuri- MBP by p43 was most pronounced when it copurified with fied with FD5, compared with copurification with WTFas or FD5. In contrast, p43 manifested considerable autophosphory- FD1 (data not shown). By silver staining gels that contained no lation when copurified using FD1 (Fig. 4A, lane 9), such that no substrate, we determined that the p43 FADD kinase and the p45 further increase was seen in the presence of MBP (Fig. 4B, lane protein observed in Figure 2 were distinct proteins (data not The Journal of Immunology 4885

FIGURE 3. Kinase activity from EL4 cell lysates Downloaded from associates specifically with Fas and phosphorylates MBP, FADD, and Fas itself. A, Lysates of EL4 cells were incubated with the indicated GST-Fas constructs, purified on glutathione Sepharose, and subjected to an in vitro kinase assay. MBP (10 ␮g/lane) was provided to samples in lanes 4, 7, 12, 15, and 19 as a serine/ threonine kinase substrate. GST was provided as a sol- http://www.jimmunol.org/ uble substrate (lanes 8, 16, and 20) or immobilized on glutathione Sepharose (lane 9). Following the kinase reaction, samples were resolved by SDS-PAGE. The dried gel was analyzed by autoradiography. A representative example of six independent experiments is shown. B, Following an in vitro kinase assay of GST- Fas with EL4 lysate, samples were incubated in the absence (lane 1) or the presence (lane 2) of thrombin to separate the GST and Fas moieties. The findings show that phosphorylation is confined to the Fas com- by guest on October 1, 2021 ponent and not to GST. Lane 3 shows a kinase reaction performed in the absence of Mg2ϩ to inhibit kinase activity. C, Following an in vitro kinase assay of GST- WTFas and subsequent gel separation, samples were blotted to a polyvinylidene difluoride membrane, and a two-dimensional phosphoamino acid analysis was performed. D, GST-FADD (10 ␮g/lane) was provided as the substrate for an in vitro kinase assay in lanes 1–10. Histone H1 (10 ␮g/lane) (H1; lanes 11 and 12) served as a control whose phosphorylation was not increased by Fas-associating kinases. E, The gel shown in D was analyzed by a beta-scanner. The total cpm in the band representing phosphorylated GST-FADD are plotted. 4886 ASSOCIATION OF FADD KINASE ACTIVITY WITH Fas Downloaded from http://www.jimmunol.org/

FIGURE 4. GST-FADD and MBP are phosphorylated by different Fas-associated kinases. Fas-associating EL4 proteins were loaded onto polyacrylamide gels that had polymerized in the absence of substrate (A) or in the presence of MBP (B), GST (C), or GST-FADD (D). Following separation by SDS-PAGE, gels were subjected to an in-gel kinase assay as described in Materials and Methods. The experiment shown in A and B was performed separately from the experiment shown in C and D. Therefore, while gels processed together can be compared, the phosphorylation levels of gels processed separately cannot be compared. by guest on October 1, 2021 shown). Thus, we observed MBP and FADD kinase activities as tion of several other Fas-associating proteins. Overexpression of well as two additional proteins that associate with FADD/MORT1, RIP, and FAF1 potentiate apoptosis, while FAP-1 cytoplasmic Fas. inhibits this process (25–29). In addition, RIP contains a kinase-like motif (27), FAP-1 bears tyrosine phosphatase activity (29), and Discussion FADD/MORT1 can undergo phosphorylation in vivo (30, 40). We have identified MBP kinase activities, a FADD kinase activity, We used a GST fusion protein strategy rather than coimmuno- and two additional proteins (p33 and p45) that associate with the precipitation or the yeast two-hybrid method to screen for Fas- cytoplasmic domain of murine Fas. These proteins bind preferen- associating proteins. Coimmunoprecipitation techniques may not tially to the 51-aa membrane-proximal cytoplasmic tail. The cur- detect Fas-associating proteins if they are expressed at levels too rent findings suggest that proteins associate with regions other than low to permit detectable interactions with Fas, and this may ex- the death domain of Fas. Similar findings have been observed for plain why we were unable to coimmunoprecipitate kinase activity the Fas-related p55 TNFR (41–44). with an Ab to Fas (N.J.K., unpublished observations). The GST Several recent reports have described Fas-associating proteins. Us- fusion system provides the advantage of considerably augmenting ing the yeast two-hybrid system, five different studies have yielded the concentration of Fas as a bait. While several Fas-interacting five different proteins, FADD/MORT1 (27 kDa) (25, 26), RIP (74 proteins have been identified using the yeast two-hybrid system, kDa) (27), FAF1 (74 kDa) (28), FAP-1 (250 kDa) (29), and Daxx the biologic relevance of some of these associations remains un- (120 kDa) (17). The discovery of multiple proteins suggests that ad- clear. For example, while RIP was originally described as a Fas- ditional Fas-associating proteins likely exist. Indeed, using other ap- associating protein based on yeast two-hybrid screens, recent ev- proaches, similar as well as additional proteins have been identified. idence shows that RIP is involved in signal transduction through These include cytotoxicity-dependent APO-1-associated protein the p55 TNFR (45, 46). Additionally, since yeast two-hybrid (CAP)1/2 (FADD/MORT1), CAP3, and CAP4 by coimmunoprecipi- screens only identify proteins that interact directly with Fas, pro- tation (30), and p25, p50, and p70 proteins by GST-Fas association teins that associate indirectly through intermediates are not de- (31). Recently, CAP4 was identified as homologous to both FADD tected. A final concern was that trimerization of Fas is required for and the ICE/Ced-3 family of cysteine proteases and was named signaling of death (34), and it was not known with certainty if the FLICE/MACH (32, 33). CAP3 was also determined to be a FLICE/ yeast two-hybrid system would provide the necessary oligomer- MACH prodomain, which likely remains associated with the Fas re- ization for protein associations. Due to its high density on gluta- ceptor signaling complex after proteolytic activation (32). Collec- thione Sepharose, GST-Fas may efficiently mimic the needed oli- tively, this suggests a unique biochemical link between a surface gomerization. A potential disadvantage of the GST-Fas fusion receptor and a proteolytic cascade. Little is known regarding the func- protein approach is that such augmented oligomerization of Fas The Journal of Immunology 4887 could potentiate the activation of Fas-associating proteins even in tween these events and the known Fas-associating proteins has not the absence of Fas ligation. This may explain why the stimulation yet been made. The proteins and kinases described here are dis- of cells with Fas ligand-transfected fibroblasts had no effect on tinguished by their association with the membrane-proximal do- kinase activity. main of Fas and may link this domain to these other pathways and We observed that Fas could associate with a kinase activity other functions of Fas. capable of phosphorylating FADD. In vitro kinase assays showed that this activity associated with WTFas and lprcg. However, the Acknowledgments FADD kinase activity associated best with FD5, suggesting that We thank Dr. Sheldon Cooper, Dr. Michael S. Vincent, Jennifer Q. Russell, the C-terminal, negative regulatory domain, which FD5 lacks, in- David J. Wilson, Karen A. Fortner, Brian S. McLellan, and Michael W. hibited either the binding or the activity of the FADD kinase. The Kleeman for helpful discussions in the preparation of this manuscript. C-terminal, negative regulatory domain of Fas has been shown to bind FAP-1, a 250-kDa tyrosine phosphatase (29). However, since References kinase assays were performed in the presence of the tyrosine phosphatase inhibitor orthovanadate, any effect of the C-terminal do- 1. Yonehara, S., A. Ishii, and M. Yonehara. 1989. A cell-killing monoclonal antibody (anti-Fas) to a cell surface antigen co-downregulated with the receptor of main of Fas on FADD kinase activity may not be due to the en- tumor necrosis factor. J. Exp. Med. 169:1747. zymatic activity of the FAP-1 phosphatase. Alternatively, the large 2. Trauth, B. C., C. Klas, A. M. Peters, S. Matzku, P. Moller, W. Falk, K. M. Debatin, and P. H. Krammer. 1989. Monoclonal antibody-mediated tumor size of FAP-1 that interacts might promote steric hindrance to regression by induction of apoptosis. Science 245:301. other Fas-associating proteins. In sharp contrast to the MBP kinase 3. Itoh, N., S. Yonehara, A. Ishii, M. Yonehara, S. Mizushima, M. Sameshima, A. Hase, Y. Seto, and S. Nagata. 1991. The polypeptide encoded by the cDNA activity, no FADD kinase activity was observed that was associ- Downloaded from for human cell surface antigen Fas can mediate apoptosis. Cell 66:233. ated with FD1. In-gel kinase assays revealed that the FADD kinase 4. Smith, C. A., T. Davis, D. Anderson, L. Solam, M. P. Beckmann, R. Jerzy, was the same 43-kDa kinase that was noted earlier to be capable of S. K. Dower, D. Cosman, and R. G. Goodwin. 1990. A receptor for tumor ne- phosphorylating MBP. The observation that the p105 and p120 crosis factor defines an unusual family of cellular and viral proteins. Science 248:1019. MBP kinases did not phosphorylate FADD underscores the point 5. Watanabe-Fukunaga, R., C. I. Brannan, N. Itoh, S. Yonehara, N. G. Copeland, that the kinase activities are different. N. A. Jenkins, and S. Nagata. 1992. The cDNA structure, expression, and chro- mosomal assignment of the mouse Fas antigen. J. Immunol. 148:1274.

The current findings indicate that the enzymatic activity of pro- 6. Chinnaiyan, A. M., K. O’Rourke, G. L. Yu, R. H. Lyons, M. Garg, D. R. Duan, http://www.jimmunol.org/ teins that associate with the membrane-proximal region of Fas is L. Xing, R. Gentz, J. Ni, and V. M. Dixit. 1996. Signal transduction by DR3, a affected by other cytoplasmic Fas sequences. For example, while death domain-containing receptor related to TNFR-1 and CD95. Science 274: 990. the p43 kinase binds to the membrane-proximal domain of Fas, as 7. Bodmer, J. L., K. Burns, P. Schneider, K. Hofmann, V. Steiner, M. Thome, revealed by the autophosphorylation property of p43, this associ- T. Bornand, M. Hahne, M. Schroter, K. Becker, A. Wilson, L. E. French, ation is insufficient for FADD phosphorylation. This observation J. L. Browning, H. R. MacDonald, and J. Tschopp. 1997. TRAMP, a novel apoptosis-mediating receptor with sequence homology to tumor necrosis factor implies that full activity of the p43 kinase requires membrane- receptor 1 and Fas(Apo-1/CD95). Immunity 6:79. distal cytoplasmic Fas sequences, although not the C-terminal, 8. Pan, G., K. O’Rourke, A. M. Chinnaiyan, R. Gentz, R. Ebner, J. Ni, and V. M. Dixit. 1997. The receptor for the cytotoxic ligand TRAIL. Science 276: negative regulatory domain. 111. This study demonstrates the phosphorylation of Fas itself. The 9. Nagata, S., and P. Golstein. 1995. The Fas death factor. Science 267:1449. by guest on October 1, 2021 responsible kinase copurified even with FD1, which contains only 10. Itoh, N., and S. Nagata. 1993. A novel protein domain required for apoptosis: mutational analysis of human Fas antigen. J. Biol. Chem. 268:10932. the 51 aa of cytosolic Fas nearest the plasma membrane. More- 11. Tartaglia, L. A., T. M. Ayres, G. H. Wong, and D. V. Goeddel. 1993. A novel over, the membrane-proximal region of cytoplasmic Fas contains domain within the 55-kDa TNF receptor signals cell death. Cell 74:845. several serine and threonine residues that could be phosphorylated 12. Watanabe-Fukunaga, R., C. I. Brannan, N. G. Copeland, N. A. Jenkins, and S. Nagata. 1992. Lymphoproliferation disorder in mice explained by defects in by a Fas-specific kinase. Recent reports suggest Fas may also be Fas antigen that mediates apoptosis. Nature 356:314. phosphorylated in vivo. In one study, anti-Fas immunoprecipita- 13. Gulbins, E., R. Bissonnette, A. Mahboubi, S. Martin, W. Nishioka, T. Brunner, tion of [32P]orthophosphate-labeled murine B lymphoma cell ly- G. Baier, G. Baier-Bitterlich, C. Byrd, F. Lang, et al. 1995. FAS-induced apoptosis is mediated via a ceramide-initiated RAS signaling pathway. Immunity sates revealed a phosphoprotein of the same molecular mass as Fas 2:341. (40). In another study, GST fused to the intracellular domain of 14. Ham, J., C. Babij, J. Whitfield, C. M. Pfarr, D. Lallemand, M. Yaniv, and L. L. Rubin. 1995. A c-Jun dominant negative mutant protects sympathetic neu- human Fas was overexpressed in murine L929 fibroblasts. Purifi- rons against programmed cell death. Neuron 14:927. 32 cation of GST-Fas from [ P]orthophosphate-labeled cells re- 15. Xia, Z., M. Dickens, J. Raingeaud, R. J. Davis, and M. E. Greenberg. 1995. vealed Fas to be phosphorylated (47). Also, a member of the Fas Opposing effects of ERK and JNK-p38 MAP kinases on apoptosis. Science 270: 1326. family, TNFR1, is known to bind a 52-kDa kinase, TRAK (TNFR- 16. Wilson, D. J., K. A. Fortner, D. H. Lynch, R. R. Mattingly, I. G. Macara, associated kinase), that phosphorylates TNFR1 (36, 48). This ki- J. A. Posada, and R. C. Budd. 1996. JNK, but not MAPK, activation is associated nase was identified using a GST-TNFR1 fusion protein “pull- with Fas-mediated apoptosis in human T cells. Eur. J. Immunol. 26:989. 17. Yang, X., R. Khosravi-Far, H. Y. Chang, and D. Baltimore. 1997. Daxx, a novel down” system similar to the one used here. While similarities Fas-binding protein that activates JNK and apoptosis. Cell 89:1067. between the MBP and Fas kinase activities are evident, attempts to 18. Cifone, M. G., R. De Maria, P. Roncaioli, M. R. Rippo, M. Azuma, L. L. Lanier, A. Santoni, and R. Testi. 1994. Apoptotic signaling through CD95 (Fas/Apo-1) identify the Fas kinase by in-gel kinase assays were unsuccessful. activates an acidic sphingomyelinase. J. Exp. Med. 180:1547. Many of the Fas-binding proteins identified to date associate 19. Raines, M. A., R. N. Kolesnick, and D. W. Golde. 1993. Sphingomyelinase and with the death domain. This does not exclude the possibility that ceramide activate mitogen-activated protein kinase in myeloid HL-60 cells. J. Biol. Chem. 268:14572. additional proteins may bind to other regions of Fas and are im- 20. Gill, B. M., H. Nishikata, G. Chan, T. L. Delovitch, and A. Ochi. 1994. Fas portant for signaling. Yeast two-hybrid methods have generally antigen and sphingomyelin-ceramide turnover-mediated signaling: role in life and used as a screening strategy differential binding to the cytoplasmic death of T lymphocytes. Immunol. Rev. 142:113. cg 21. Malinin, N. L., M. P. Boldin, A. V. Kovalenko, and D. Wallach. 1997. MAP3K- domain of WTFas but not binding to lpr or C-terminally trun- related kinase involved in NF-␬B induction by TNF, CD95 and IL-1. Nature cated Fas. By this approach, proteins that could potentially bind to 385:540. 22. Ponton, A., M. V. Clement, and I. Stamenkovic. 1996. The CD95 (APO-1/Fas) the membrane-proximal region of Fas would not be selected. In receptor activates NF-␬B independently of its cytotoxic function. J. Biol. Chem. addition, it has been reported that Fas activates a variety of signal 271:8991. pathways, including sphingomyelinase (13, 18–20) and NF-␬B 23. Enari, M., H. Hug, and S. Nagata. 1995. Involvement of an ICE-like protease in Fas-mediated apoptosis. Nature 375:78. (21, 22), and also costimulates the proliferation of resting T lym- 24. Los, M., M. Van de Craen, L. C. Penning, H. Schenk, M. Westendorp, phocytes by augmenting IL-2 production (49). A direct link be- P. A. Baeuerle, W. Droge, P. H. Krammer, W. Fiers, and K. Schulze-Osthoff. 4888 ASSOCIATION OF FADD KINASE ACTIVITY WITH Fas

1995. Requirement of an ICE/CED-3 protease for Fas/APO-1-mediated apopto- 38. Boyle, W. J., P. van der Geer, and T. Hunter. 1991. Phosphopeptide mapping and sis. Nature 375:81. phosphoamino acid analysis by two-dimensional separation on thin-layer cellu- 25. Chinnaiyan, A. M., K. O’Rourke, M. Tewari, and V. M. Dixit. 1995. FADD, a lose plates. Methods Enzymol. 201:110. novel death domain-containing protein, interacts with the death domain of Fas 39. Golstein, P., D. Marguet, and V. Depraetere. 1995. Homology between reaper and initiates apoptosis. Cell 81:505. and the cell death domains of Fas and TNFR1. Cell 81:185. 26. Boldin, M. P., E. E. Varfolomeev, Z. Pancer, I. L. Mett, J. H. Camonis, and D. Wallach. 1995. A novel protein that interacts with the death domain of Fas/ 40. Zhang, J., and A. Winoto. 1996. A mouse Fas-associated protein with homology APO1 contains a sequence motif related to the death domain. J. Biol. Chem. to the human Mort1/FADD protein is essential for Fas-induced apoptosis. Mol. 270:7795. Cell. Biol. 16:2756. 27. Stanger, B. Z., P. Leder, T. H. Lee, E. Kim, and B. Seed. 1995. RIP: a novel 41. Adam-Klages, S., D. Adam, K. Wiegmann, S. Struve, W. Kolanus, protein containing a death domain that interacts with Fas/APO-1 (CD95) in yeast J. Schneider-Mergener, and M. Kronke. 1996. FAN, a novel WD-repeat protein, and causes cell death. Cell 81:513. couples the p55 TNF-receptor to neutral sphingomyelinase. Cell 86:937. 28. Chu, K., X. Niu, and L. T. Williams. 1995. A Fas-associated protein factor, 42. Castellino, A. M., G. J. Parker, I. V. Boronenkov, R. A. Anderson, and FAF1, potentiates Fas-mediated apoptosis. Proc. Natl. Acad. Sci. USA 92:11894. M. V. Chao. 1997. A novel interaction between the juxtamembrane region of the 29. Sato, T., S. Irie, S. Kitada, and J. C. Reed. 1995. FAP-1: a protein tyrosine p55 tumor necrosis factor receptor and phosphatidylinositol-4-phosphate 5-ki- phosphatase that associates with Fas. Science 268:411. nase. J. Biol. Chem. 272:5861. 30. Kischkel, F. C., S. Hellbardt, I. Behrmann, M. Germer, M. Pawlita, P. H. Krammer, and M. E. Peter. 1995. Cytotoxicity-dependent APO-1 (Fas/ 43. Dunbar, J. D., H. Y. Song, D. Guo, L. W. Wu, and D. B. Donner. 1997. Two- CD95)-associated proteins form a death-inducing signaling complex (DISC) with hybrid cloning of a gene encoding TNF receptor-associated protein 2, a protein the receptor. EMBO J. 14:5579. that interacts with the intracellular domain of the type 1 TNF receptor: identity 31. Orlinick, J. R., and M. V. Chao. 1996. Interactions of cellular polypeptides with with subunit 2 of the 26S protease. J. Immunol. 158:4252. the cytoplasmic domain of the mouse Fas antigen. J. Biol. Chem. 271:8627. 44. Boldin, M. P., I. L. Mett, and D. Wallach. 1995. A protein related to a protea- 32. Muzio, M., A. M. Chinnaiyan, F. C. Kischkel, K. O’Rourke, A. Shevchenko, somal subunit binds to the intracellular domain of the p55 TNF receptor upstream J. Ni, C. Scaffidi, J. D. Bretz, M. Zhang, R. Gentz, M. Mann, P. H. Krammer, to its ’death domain’. FEBS Lett. 367:39. M. E. Peter, and V. M. Dixit. 1996. FLICE, a novel FADD-homologous ICE/ 45. Ting, A. T., F. X. Pimentel-Muinos, and B. Seed. 1996. RIP mediates tumor CED-3-like protease, is recruited to the CD95 (Fas/APO-1) death-inducing sig- Downloaded from necrosis factor receptor 1 activation of NF-␬B but not Fas/APO-1-initiated apo- naling complex. Cell 85:817. 33. Boldin, M. P., T. M. Goncharov, Y. V. Goltsev, and D. Wallach. 1996. Involve- ptosis. EMBO J. 15:6189. ment of MACH, a novel MORT1/FADD-interacting protease, in Fas/APO-1- and 46. Hsu, H., J. Huang, H. B. Shu, V. Baichwal, and D. V. Goeddel. 1996. TNF- TNF receptor-induced cell death. Cell 85:803. dependent recruitment of the protein kinase RIP to the TNF receptor-1 signaling 34. Suda, T., T. Takahashi, P. Golstein, and S. Nagata. 1993. Molecular cloning and complex. Immunity 4:387. expression of the Fas ligand, a novel member of the tumor necrosis factor family. 47. Rudert, F., E. Visser, G. Gradl, P. Grandison, L. Shemshedini, Y. Wang, Cell 75:1169. A. Grierson, and J. Watson. 1996. pLEF, a novel vector for expression of glu- 35. Frangioni, J. V., and B. G. Neel. 1993. Solubilization and purification of enzy- tathione S-transferase fusion proteins in mammalian cells. Gene 169:281. http://www.jimmunol.org/ matically active glutathione S-transferase (pGEX) fusion proteins. Anal. Bio- 48. Darnay, B. G., S. Singh, M. M. Chaturvedi, and B. B. Aggarwal. 1995. The p60 chem. 210:179. 36. Darnay, B. G., S. A. Reddy, and B. B. Aggarwal. 1994. Identification of a protein tumor necrosis factor (TNF) receptor-associated kinase (TRAK) binds residues kinase associated with the cytoplasmic domain of the p60 tumor necrosis factor 344–397 within the cytoplasmic domain involved in TNF signaling. J. Biol. receptor. J. Biol. Chem. 269:20299. Chem. 270:14867. 37. Kameshita, I., and H. Fujisawa. 1989. A sensitive method for detection of calm- 49. Alderson, M. R., R. J. Armitage, E. Maraskovsky, T. W. Tough, E. Roux, odulin-dependent protein kinase II activity in sodium dodecyl sulfate-polyacryl- K. Schooley, F. Ramsdell, and D. H. Lynch. 1993. Fas transduces activation amide gel. Anal. Biochem. 183:139. signals in normal human T lymphocytes. J. Exp. Med. 178:2231. by guest on October 1, 2021