The Microfilament Cytoskeleton Moshe M

Home , Cytoskeleton, Microfilament

pp56Lck Mediates TCR ζ-Chain Binding to the Microfilament Cytoskeleton Moshe M. Rozdzial, Chris M. Pleiman, John C. Cambier and Terri H. Finkel3 This information is current as of September 23, 2021. J Immunol 1998; 161:5491-5499; ; http://www.jimmunol.org/content/161/10/5491 Downloaded from References This article cites 28 articles, 17 of which you can access for free at: http://www.jimmunol.org/content/161/10/5491.full#ref-list-1

Why The JI? Submit online. http://www.jimmunol.org/ • Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on September 23, 2021 Subscription Information about subscribing to The Journal of Immunology is online at: 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. pp56Lck Mediates TCR ␨-Chain Binding to the Microﬁlament Cytoskeleton1

Moshe M. Rozdzial,* Chris M. Pleiman,2* John C. Cambier,*† and Terri H. Finkel3†‡

The TCR ␨-chain (␨) on mature murine T lymphocytes binds to the microfilament cytoskeleton in response to Ag receptor ligation. Here, we report the role of Src family kinases in ␨-cytoskeletal binding, using mutant mice and a cell-free model system. Binding of ␨ to actin in the cell-free system has a specific requirement for ATP and divalent cations, with an apparent Michaelis-Menton constant for ATP in the millimolar range, and can be disrupted by either EDTA or the microfilament poison, cytochalasin D, suggesting that microfilaments provide the structural framework for an active process involving cellular kinases. Indeed, tyrosine- phosphorylated ␨ is a predominant form of the ␨-chain bound to polymerized actin, while challenge with alkaline phosphatase prevents ␨-chain association in solution and releases ␨-chain from the bound state. Phosphorylated Src-family kinase pp56Lck also

associates with membrane skeleton upon TCR engagement and is a component of the reconstituted cytoskeletal pellet. ␨-Chain Downloaded from phosphorylation and ␨-cytoskeletal binding are abrogated in cell lysates with reduced levels of pp56Lck and in activated mutant murine T cells lacking pp56Lck, implicating pp56Lck as the kinase involved in ␨-chain tyrosine phosphorylation and ␨-cytoskeletal binding. Finally, recombinant Lck Src homology 2 domain preferentially inhibits reconstituted ␨-cytoskeleton association, suggesting that ␨-microfilament binding is dependent on interactions between phosphorylated tyrosine residues in ␨-chain activation motifs and the Src homology 2 domain of the Lck protein tyrosine kinase. The Journal of Immunology, 1998, 161: 5491–5499. http://www.jimmunol.org/ ctivation of mature T lymphocytes is regulated by the association of specific PTKs with the ITAMs via Src homology 2 multichain TCR. The TCR, upon engagement, initiates a (SH2) domains (10, 11). A cascade of biochemical events that lead to receptor as- SH2 domains play a significant role in signal transduction in sociation with cytoskeleton (1) and culminate in the rearrangement many cell types by mediating the formation of specific heteromeric and reorientation of the cytoskeleton (2), production of lympho- protein complexes with phosphotyrosine-containing peptides (12). kines, and cellular proliferation and differentiation. The TCR con- The binding specificity of this protein-protein interaction is depen- sists of the Ag-binding subunits (␣␤), the CD3 complex (␥-, ␦-, dent on both the primary sequence flanking the site of tyrosine and ␧-chains), and the ␨-chains (3). Both the CD3 components and phosphorylation and the structure of the particular SH2 domain ␨-chains are capable of independently mediating TCR signaling (13). Individual SH2-containing polypeptides select unique se- by guest on September 23, 2021 (4), and each contains, respectively, one and three copies of the quences, except for the Src subfamily (Src, Fyn, Lck, and Fgr), immunoreceptor tyrosine-based activation motif (ITAM),4 com- which preferentially recognizes the sequence, P-Tyr-Glu-Glu-Ile

posed of the amino acid sequences YXX(L/I)X(7–8)YXX(L/I) (5). (13). Thus, SH2-containing polypeptides serve as adaptors to link As the TCR has neither intrinsic protein tyrosine kinase (PTK) nor speciﬁc effector proteins to phosphotyrosine-containing target phosphatase function, the ITAM sequences appear to be both nec- motifs. essary and sufﬁcient for the induction of protein tyrosine phos- In T cells, biochemical and genetic studies have implicated the phorylation in T cells (6). While there is a low level of constitutive Src (pp56Lck and p59Fyn) and the Syk/ZAP-70 families of PTKs in tyrosine phosphorylation of ␨, even in resting T cells (7, 8), with cellular activation, demonstrating a functional interaction of these activation, tyrosine phosphorylation of ␨ increases (9), leading to kinases with the TCR complex and/or the coreceptors CD4 and CD8. Both pp56Lck (Lck) and p59Fyn (Fyn) have been shown to ␨ ␧ *Division of Basic Sciences, Department of Pediatrics, National Jewish Medical and interact with ITAMs in the - and CD3 -chains, resulting in phos- Research Center, Denver, CO 80206; †Department of Immunology, University of phorylation of the ITAM tyrosines, and leading to recruitment of Colorado Health Sciences Center, Denver, CO 80262; and ‡Departments of Pediatrics other molecules involved in T cell signal transduction, such as and Biochemistry & Molecular Genetics, University of Colorado Health Sciences Center, Denver, CO 80262 ZAP-70 (10, 11). Received for publication December 15, 1997. Accepted for publication July 16, 1998. Although not well studied in T cells, the cytoskeleton also plays a direct role in the regulation and compartmentalization of the 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 activation process (reviewed in Ref. 1). Previously, we have shown with 18 U.S.C. Section 1734 solely to indicate this fact. that as a consequence of TCR ligation, the ␨-chain rapidly binds to 1 This work was supported by National Institutes of Health Grants R01 AI30575, P01 the membrane skeleton independent of receptor internalization (1). AI22295 (T.H.F.), and T32 AI00048 (M.M.R.), the University of Colorado Health Pretreatment with drugs that disrupt the actin cytoskeleton abro- Sciences Center Cancer Center (M.M.R. and T.H.F.), the Arthritis Foundation, the Bender Foundation, and the Eleanore and Michael Stobin Trust (T.H.F.). gated the association of ␨ with cytoskeleton (1, 14) and inhibited 2ϩ ␥ 2 Current address: Cadus Phamaceutical, 777 Old Saw Mill River Rd., Tarrytown, NY sustained Ca mobilization, IFN- (15) and IL-2 (M. M. Rozd- 12533. zial, unpublished observations) production. In addition, in T cell 3 Address correspondence and reprint requests to Dr. Terri H. Finkel, Division of hybridomas transfected with mutated or deleted ␨-chain chimeras, Basic Sciences, Department of Pediatrics, National Jewish Medical and Research lack of ␨-cytoskeleton association correlated with inhibition of late Center, 1400 Jackson St., Denver, CO 80206. activation events, such as IL-2 production (1). Recent studies have 4 Abbreviations used in this paper: ITAM, immunoreceptor tyrosine-based activation motif; PTK, protein tyrosine kinase; SH2, Src homology 2; GST, glutathione-S-trans- also documented the activation-dependent association of tyrosine ferase; SHIP, Src homology 2-containing inositol phosphatase. kinases with the membrane skeleton in both B (16) and T (17)

lymphocytes. These data provide direct evidence of an activation- the addition of sample buffer to the reconstituted pellets. Electrophoresis, dependent interaction between Ag receptor and the cytoskeletal immunoblotting, and densitometry were performed as previously described ␨ ␨ matrix that may support the interaction of signaling polypeptides (1). Amount of bound -chain was quantitated as associated per second relative to the baseline (assigned a value of 1) at the 0 time control. As the and their substrates. results for thymocytes were essentially identical to those obtained for Determination of the role of microﬁlaments in TCR-signaling lymph node T cells, these sets of data were pooled. cascades and molecular dissection of the ␨-cytoskeletal interaction ϩ Effect of Mg2 is, however, hampered by the formation of a complex insoluble ϩ pellet upon activation of intact cells. Therefore, we developed an T cell lysates were incubated with increasing concentrations of Mg2 with assay for analysis of ␨-cytoskeleton association under cell-free or without 1 mM ATP for 8 min at 37°C. Electrophoresis and immunoblotting were performed as previously described (1). conditions, so as to be able to modify the system before pellet formation (1). Here, we demonstrate that addition of divalent cat- Challenge with alkaline phosphatase ␨ ions and ATP to a T cell lysate leads to the phosphorylation of Cell lysates were incubated in the presence or absence of 1 U/100 ␮lof and association of ␨ with microﬁlaments. Furthermore, we show bacterial alkaline phosphatase for 30 min and incubated with MgATP, with

that, in the lysed cell system, ␨-cytoskeleton association is prefer- or without phosphatase inhibitors (0.2 mM VO3, 10 mM NaF), as de- entially inhibited by a recombinant Lck-SH2 peptide or by a re- scribed. The reconstituted MgATP pellets were then disrupted and incubated in the presence or absence of 1 U/10 ␮l alkaline phosphatase and duction in levels of Lck. Finally, as predicted by these data from centrifuged at 10,000 rpm for 10 min to sediment the precipitated material. the cell-free system, we show that ␨-cytoskeleton association is Immunoprecipitation of the detergent soluble fractions and immunoblot- abrogated in activated T cells from Lck-deﬁcient (knockout) mice. ting was performed as previously described (1). Downloaded from Materials and Methods Immunoprecipitation of Src-family kinases Animals T cell lysates were incubated in the presence or absence of polyclonal Abs to Fyn (courtesy of Dr. Terry Potter), Lck (courtesy of Dr. Terry Potter), or All mice, except the Lck knockout mice (the kind gift of Dr. Tak Mak, pp60c-Src (Upstate Biotechnology, Lake Placid, NY) for 30 min at 4°C. The Amgen, Thousand Oaks, CA) were bred in our facility or purchased from tyrosine kinases were depleted from solution by immunoprecipitation with The Jackson Laboratory (Bar Harbor, ME). protein A-Sepharose (Pharmacia, Uppsala, Sweden) for1hat4°C. De- In vitro reconstitution, immunoprecipitation, gel electrophoresis, pleted lysates were incubated with or without MgATP, and electrophoresis http://www.jimmunol.org/ and immunoblotting were performed as described. and immunoblotting Competition with exogenous SH2 peptide Thymocytes or lymph node T cells were freshly prepared and isolated as cell suspensions from normal adult mice by pressing organs through a Cell lysates were incubated for2hintheabsence or presence of glutathi- 200-␮ nylon mesh (Bally Ribbon, Bally, PA). The cell suspensions were one-S-transferase (GST) or GST fusion proteins containing the SH2 do- then washed three times in balanced salt solution and 5% FBS, solubilized main of the Src family kinases, Fyn (residues 144–255) (18), Lck (residues with 0.5% Nonidet P-40 in a Tris-buffered saline (150 mM NaCl, 10 mM 117–239), or the SHIP phosphatase (residues 0–114, kindly provided by Tris, pH 7.3) solution containing protease and phosphatase inhibitors (0.2 Kazuhiro Nakamura (National Jewish Medical and Research Center, Den-

mM VO3, 10 mM NaF, 1 mM PMSF, and 1 mg/ml each of aprotinin, ver, CO). GST fusion proteins were prepared as described (18) and isolated ␣ leupeptin, and -1-antitrypsin) and centrifuged at 10,000 rpm for 10 min to using glutathione-Sepharose beads (Pharmacia). Cell lysates were also in- by guest on September 23, 2021 pellet the preexisting detergent-insoluble material. The detergent soluble cubated for2hinthepresence or absence of increasing concentrations of fraction was then precleared of endogenously reconstituted components exogenous Fyn-SH2 peptide containing the Fyn residues 144–255 (18) following an initial warming at 37°C for 10 min, and then incubated with (our unpublished data), which span the 300 base pairs encoding the SH2 or without MgATP, or Mg2ϩ (0.2 mM, or as otherwise indicated) or ATP domain of the kinase. Preparation of other regions of the Fyn PTK has been (1 mM, or as otherwise indicated) alone at 37°C for 10 min, and centri- described previously (18). Following incubation in the absence or presence fuged at 10,000 rpm for 10 min to sediment the precipitated material. of the Fyn-SH2 peptide, lysates were incubated with MgATP, as described Immunoprecipitation of the detergent-soluble fraction was then performed above. Control lysates were incubated with a concentration of BSA equiv- with Sepharose-conjugated anti-␨ mAb (H146-968) (8) or, in series, with alent to the highest concentration of SH2 peptide used. agarose-linked anti-phosphotyrosine Ab (Ab-1; Oncogene Science, Union- dale, NY) and anti-␨ mAb (H146-968). Boiled protein samples at 1–5 ϫ Constructs 7 10 cell equivalents/lane were separated under nonreducing conditions by DNA fragments containing the portion encoding the SH2 domain of Fyn one-dimensional SDS-PAGE (10%). Except where otherwise indicated, (residues 144–255) were amplified with the PCR using the following prim- equivalent cell numbers were loaded per lane in each experiment. Electro- ers: Fyn 144–255, 5Ј-CAGTCAGAATTCGATGGAGTCAACTGGAGC phoretic transfer of protein onto 0.2-mm nitrocellulose filters was carried CA-3Ј and 5Ј-CAGTCAGAATTCTCCAGGTTTGTGGGGTAC-3Ј. The out in 48 mM Tris, 39 mM glycine, 1.3 mM SDS, and 20% methanol at PCR products were then ligated into pGEX-3X (Pharmacia) and trans- room temperature and constant current (150–200 mA) for 2 h. The filters fected into Escherichia coli DH5␣ (Life Technologies, Gaithersburg, MD). were then quenched in blotting buffer composed of 125 mM NaCl and 25 The Fyn-SH2 peptide was subsequently eluted by cleavage from the beads mM Tris, pH 7.6 (TS), and 5% skim milk, or with 5% crystallized BSA for with 30 ␮g of factor Xa (Boehringer Mannheim, Indianapolis, IN) and phosphotyrosine detection. Following electrotransfer, the nitrocellulose fil- ␨ dialyzed into PBS. For the generation of GST fusion proteins containing ters were immunoblotted with specific Abs to or phosphotyrosine (Ab-2; the SH2 domains of Lck (residues 117–239), Fyn (residues 144–255), or Oncogene Science) and washed in TS-0.05% Tween-20. Actin was de- SHIP (residues 0–114), PCR was used to amplify cDNAs encoding the tected with an anti-actin mAb (kindly provided by Dr. B. Jockusch, Braun- SH2 domains. The oligo pair used for amplifying the mouse Lck SH2 schweig, Germany). Lck was detected with polyclonal Abs raised in rabbits domain was: 5Ј oligo, 5Ј-GATATCGCGAAAGCAAACAGCCTG-3Ј, and against the C-terminal sequence of Lck and were the kind gift of Dr. Terry 3Ј oligo, 5Ј-GAATTCCCATTCGTCCTCCCACCATGG-3Ј. The amino Potter (National Jewish Medical and Research Center, Denver, CO). The acids encoded in this fragment span 117–239. The protein fragment con- 125 ϫ 5 washed filters were incubated with [ I]protein A (4 10 cpm/ml) in tains 10 amino acids on either side of the SH2 domain to stabilize folding quenching buffer for 1 h and washed as above. The blots were then dried Ϫ of the domain. The PCR product was obtained by amplifying from the and exposed to Kodak XAR-2 film at 70°C. Densitometry was done on mLck cDNA (kindly provided by Roger M. Perlmutter, University of a MacIntosh image scanner and analyzed with the Image 1.49 program Washington, Seattle, WA), which was then subcloned into pCR2.1 (In- (National Institutes of Health, Bethesda, MD) for 1-D scanning. Unless vitrogen, Carlsbad, CA). The 5Ј and 3Ј amplifying oligos contain EcoRV otherwise stated, all reagents were purchased from Sigma Chemical Co., and EcoRI restriction sites. The sequence was confirmed by dideoxynucle- St. Louis, MO. otide-sequence analysis using Sequenase (United States Biochemical, Kinetics of ␨-cytoskeleton association Cleveland, OH). The fragment containing the Lck SH2 domain was cleaved with EcoRV (blunt) and EcoRI and cloned into pGEX-3X (Phar- Thymocyte and lymph node T cell lysates were incubated with 0.2 mM macia) cut with SmaI (blunt) and EcoRI. The following primers were Mg2ϩ and varying ATP concentrations for 0, 0.5, 1, 2, 4, and 8 min at used for the SHIP SH2 0–114: 5Ј primer, 5Ј-GGAATTCAT 37°C, centrifuged at 10,000 rpm for 30 s, and the reaction was stopped with GCCTGCCATGGTCCCT-3Ј;3Ј primer, 5Ј-TTTTCCTTTTGCGGCCGC The Journal of Immunology 5493

TCATCAATAGCATCCTC-3Ј. After digesting with the restriction enzymes, BamHI and EcoRI (for GST-Lck SH2) and EcoRI and NotI (for GST-SHIP SH2), the resulting fragments were ligated into pGEX-3X (Pharmacia) and pGEX-5X (Pharmacia), respectively, and transfected into E. coli DH5␣ (Life Technologies) and puriﬁed with glutathione-Sepharose beads (Pharmacia).

Lck and Fyn-deﬁcient mice Thymocytes and lymph node T cells from wild-type and Lck-knockout (19) or Fyn-knockout (20) mice were isolated and activated by ligation of CD3␧ with anti-CD3␧ mAb (145-2C11) (21) on intact cells or by addition of MgATP to T cell lysates, as described.

Results Divalent cations and ATP are required for reconstitution of ␨-cytoskeletal binding in a cell-free system In earlier work, we found that although the coprecipitation of ␨ and actin is enhanced under activating conditions, actin polmerization alone is insufficient to induce ␨ cytoskeleton association (1). Re- constitution of ␨-cytoskeleton association required addition of FIGURE 1. ␨-Chain association with the cytoskeleton can be reconsti- Downloaded from MgATP to the cell lysates (1). In order to address the possibility tuted in a cell-free system. A, ␨-Cytoskeleton association in a cell-free ϩ that Mg2ϩ or nucleotide alone contributed to ␨-cytoskeleton asso- system in the presence of Mg2 , ATP, and EDTA. Thymocytes were lysed ciation in vitro, thymic cell lysates were first cleared of preexisting and cleared of the preexisting detergent-insoluble pellet. The detergent- 2ϩ soluble supernatants were incubated in the absence or presence of 0.2 mM detergent insoluble material and then incubated with Mg or 2ϩ 2ϩ EDTA in the presence or absence of ATP (Fig. 1, A and B). Rel- Mg (Mg ; lanes 2 and 5 or lanes 1, 3, 4, and 6, respectively), 5 mM EDTA (EDTA; lanes 1 and 4 or lanes 2, 3, 5, and 6, respectively), or 5 mM ative to the 5 Ϯ 1% (SE, n ϭ 7) of the TCR␨ associated with the http://www.jimmunol.org/ ATP (ATP; lanes 1, 2, and 3 or lanes 4, 5, and 6, respectively). The cell in vitro reconstituted pellet under control conditions, upon addition lysates were then incubated at 37°C and centrifuged to separate the poly- Ϯ ϭ of exogenous MgATP (Fig. 1A), an average of 30 4% (SE, n merized pellet from the cleared supernatant. The pellets were separated by 7) of the ␨-chain was decreased from solution and cosedimented PAGE and Western blotted with anti-␨ mAb. ␨ associated with the MgATP with a secondary pellet, representing a sixfold increase over back- pellet (lane 4) only under conditions that included both ATP and Mg2ϩ ground and similar to that observed in activated lymph node T without EDTA. Optimal association occurred with pH in the acidic range cells, in vivo (1). Ca2ϩ and Mn2ϩ (data not shown) substituted for (pH 6) (data not shown). Data are representative of at least three experi- Mg2ϩ in inducing the ␨-cytoskeletal interaction. In the absence of ments. B, EDTA disrupts actin polymerization in a lysed cell system. The 2ϩ ␨ detergent-soluble supernatants were incubated in the absence or presence Mg or in the presence of chelators of divalent cations, neither 2ϩ 2ϩ (Fig. 1A) nor polymerized actin (Fig. 1B) were found in the in vitro of 2.0 mM Mg (Mg ; lanes 1, 2 and 3), 5 mM EDTA (EDTA; lanes 1 by guest on September 23, 2021 and 2 or 3, respectively), or 2.5 mM ATP (ATP; lane 1,orlanes 2 and 3, reconstituted pellet, further implicating microfilaments in ␨-cy- 2ϩ respectively). Following incubation at 37°C and centrifugation, the pellets toskeleton association. Only the lysate containing both Mg and were separated by PAGE and Western blotted with anti-␨ and anti-actin ␨ ATP (Fig. 1, A and B) showed any prominent association of with mAbs. The disruption of ␨ association with the MgATP pellet (lane 3) was the pellet, suggesting that the association observed following TCR correlated with the disruption of actin polymerization under conditions that ligation on intact cells is an ATP-dependent interaction that re- included EDTA. Data are representative of at least three experiments. C, quires divalent cations. Reconstituted ␨-cytoskeletal binding is dependent on the regulation of both ϩ To determine the individual contribution of Mg2ϩ to in vitro Mg2 and ATP levels. T cell lysates were incubated with increasing con- 2ϩ induced ␨-cytoskeleton association, increasing concentrations of centrations of Mg (0.25, 0.5, 1.2, and 4 mM) in the absence (lanes 3–7) Mg2ϩ were added to cell lysates in the presence (Fig. 1C, lanes or presence (lanes 8–12) of 1 mM ATP. Lane 1 represents lysate incubated 8–12) or absence (Fig. 1C, lanes 3–7) of 1 mM ATP. Mg2ϩ con- in 1 mM ATP alone, and lane 2 is devoid of exogenous reagents. Data are representative of at least three experiments. centrations of 1–2 mM (Fig. 1C, lanes 5 and 6) were optimal in inducing some ␨-cytoskeleton association, but only about 6% of that stimulated in the presence of ATP (Fig. 1C, lanes 10 and 11), and were comparable to amounts induced by 1 mM ATP alone ␨-cytoskeleton association, presumably in concert with endoge- (Fig. 1C, lane 1). Concentrations above 2 mM Mg2ϩ were inhib- nous divalent cations. ␨-Chain association occurred predominantly itory to ␨-cytoskeleton association (Fig. 1C, lane 7). The level of with the addition of MgATP, suggesting that ATP is specifically Mg2ϩ used in most of the studies here, 0.2–0.25 mM, is not suf- used as a substrate for the ␨-cytoskeletal interaction, possibly as a ficient to induce ␨ to translocate to the pellet (Fig. 1C, lane 3). phosphoryl donor in a kinase reaction. Indeed, the level of tyrosine ␨-Cytoskeleton association was over an order of magnitude greater phosphorylated ␨ increased in the cell lysate and was a prominent with the inclusion of both Mg2ϩ and 1 mM ATP, at all Mg2ϩ component of the pellet after incubation with MgATP (Fig. 2B). concentrations tested. Thus, although Mg2ϩ or ATP alone are ca- Under nonphosphorylating conditions there was minimal ␨ binding pable of insolubilizing some ␨-chain (presumably as a result of an to the pellet, although actin polymerization was induced (Fig. 2B). interaction with endogenous stores of ATP or Mg2ϩ, respectively), Thus, we postulate that in this cell-free system, MgATP supports together, ATP and Mg2ϩ have a synergistic effect. The observed in vitro phosphorylation of tyrosine residues on the ␨-chain, which ␨-cytoskeleton association is dependent on regulation of both the then binds (directly or indirectly) to polymerized actin. Mg2ϩ and ATP levels, as shown below. To examine the amount of ATP required for ␨-cytoskeleton as- To determine the nucleotide specificity of ␨-cytoskeleton asso- sociation and the kinetics of this association, cell lysates were in- ciation, T cell lysates were incubated in the presence or absence of cubated with increasing concentrations of MgATP. Fig. 3A shows Mg2ϩ with ATP, GTP, ADP, or the nonhydrolyzable ATP analog, the data of a time course experiment using two ATP concentra- AMPPNP (Fig. 2A). In the absence of EDTA, ATP alone induced tions, differing by an order of magnitude, as representative of the 5494 ROLE OF pp56Lck IN ␨-CHAIN-MICROFILAMENT ASSOCIATION

Interestingly, the reaction kinetics at ATP concentrations greater than or equal to 0.5 mM (Fig. 3, B and C) showed a biphasic peak, for reasons not yet understood. In addition, the degree of ␨ association at an ATP concentration greater than 1.0 mM decreased after an initial peak (Fig. 3B; 2.5 mM ATP), as was seen previously in intact cells (1), and is correlated with a decrease in actin from the pelleted material over time (Fig. 3C). Likewise, at ATP concentrations below 1.0 mM (Fig. 3A), or at temperatures equal to or below 25°C (data not shown), the kinetics of association were two to ﬁve times that observed in intact cells, peaking between 2 and 4 min. Thus, in the presence of divalent cations and physio- logic concentrations of ATP, ␨-chain associates with the cytoskeleton in a cell-free system with rate and extent similar to that seen following TCR ligation on intact cells.

␨-Cytoskeleton association is dependent upon tyrosine phosphorylation of the ␨-chain Because tyrosine phosphorylation of the ␨-chain appeared to cor- relate with microfilament association, we asked whether agents Downloaded from FIGURE 2. Reconstituted ␨-cytoskeletal binding requires ATP and is that dephosphorylate the ␨-chain, such as alkaline phosphatase, ␨ associated with tyrosine phosphorylation of the -chain. A, ATP is a spe- would disrupt ␨-cytoskeleton association. Fig. 4A shows that al- ␨ cific substrate for reconstitution of -cytoskeleton association in vitro. Cell kaline phosphatase added directly to the in vitro reconstituted lysates were incubated in the absence (lane 1) or presence of 0.5 mM Mg2ϩ MgATP pellet (Fig. 4A, lane 2) or to the cell lysate prior to acti- (Mg, lane 2)orin5mMATP(lanes 3 and 4), AMP-PNP (lanes 5 and 6), ␨ ADP (lanes 7 and 8) or GTP (lanes 9 and 10), with or without Mg2ϩ, vation (Fig. 4A, lane 3) disrupted -cytoskeleton association rela- respectively. The cell lysates were then incubated at 37°C and centrifuged tive to the control MgATP pellet (Fig. 4A, lane 1). This disruption http://www.jimmunol.org/ to isolate the pellets, which were separated by PAGE and Western blotted by the phosphatase was dose dependent, was correlated with the with anti-␨ mAb. ␨ associated predominantly with the MgATP pellet (lane increasing release of dephosphorylated ␨-chain from the pellet, and 4), relative to untreated (lane 1), Mg2ϩ-treated (lane 2), and other nucle- was inhibited by the inclusion of phosphatase inhibitors in the otide-treated lysates (lanes 5–10). Data are representative of at least two buffer (Fig. 4B). These data suggest that ␨-chain must be phos- experiments. B, Tyrosine-phosphorylated ␨-chain is a component of the phorylated in order to bind, directly or indirectly, to the insoluble pellet reconstituted with MgATP. Thymocytes were lysed, incu- cytoskeleton. bated with 2 mM Mg2ϩ (lanes 1–4) and 2.5 mM ATP (lanes 2 and 4), and centrifuged to sediment the reconstituted pellet (lanes 1–4). The pellets Lck is the tyrosine kinase responsible for the ␨-chain ␨ were separated by PAGE and Western blotted with anti- or anti- phosphorylation that leads to ␨-cytoskeleton association by guest on September 23, 2021 phosphotyrosine mAb on duplicate blots. Data are representative of at least three experiments. Dephosphorylation by alkaline phosphatase is not specific to phosphorylated tyrosines. We therefore sought to determine the involvement of the Src family members in ␨-cytoskeleton associa- aggregate data. Little or no ␨ binding above background was ob- tion. Lck, Fyn, and c-Src were probed for association with the served below 0.1 mM and appeared optimal at 1 mM ATP and cytoskeleton in intact activated thymocytes, and with the detergent above (Fig. 3, B and D). At concentrations of 1 mM ATP and insoluble pellet following reconstitution in cell lysates. Lck (Fig. above, optimal association of ␨ and the actin cytoskeleton in this 5) and Fyn (data not shown) associate with the insoluble pellet lysed cell system was very rapid (30 s at 37°C; Fig. 3, A and B) and from intact cells activated by TCR ligation (Fig. 5A, lane 2) and was dependent upon temperature and lysate concentration (data with the reconstituted pellet in the presence of MgATP (Fig. 5B, not shown). The kinetics of association of tyrosine phosphorylated lane 2). In contrast, c-Src does not appear to cosediment with ␨ with the pellet were similar to that observed for ␨-cytoskeleton either the insoluble or reconstituted pellets (data not shown). Un- association (data not shown). The kinetics of actin polymerization der both conditions, Lck is tyrosine phosphorylated (Fig. 5, lower were also similar, although with a lower threshold of MgATP con- panels), suggesting that activated Lck and phosphorylated ␨-chain centration (Fig. 3, A and C). colocalize to the actin cytoskeleton in response to TCR ligation. The differential dependence on MgATP concentration for ␨-cy- Lck phosphorylates ␨-chain upon TCR ligation (22) and is, toskeleton association and actin polymerization was borne out by along with Fyn, a candidate kinase involved in the tyrosine phos- rate measurements of relative amount of ␨ bound to the cytoskel- phorylation of ␨-chain leading to ␨-cytoskeleton association (23). etal pellet and relative actin polymerized with increasing ATP con- To differentiate the role of various protein tyrosine kinases in ␨-cy- centration. These rate measurements showed that both ␨ associa- toskeleton association, we examined T cell lysates in which levels tion and actin polymerization were saturable and followed of specific Src family kinases were decreased by immunoprecipi- Michaelis-Menton kinetics (Fig. 3, D and E, respectively). Double tation (Fig. 6). ␨-Cytoskeleton association was abrogated in the reciprocal plots of these data (Fig. 3, F and G) permitted the cal- reconstituted pellets in which levels of Lck (Fig. 6A, lane 4, and culation of the apparent Michaelis-Menton constant for ATP of the 6B) were reduced. In contrast, in T cell lysates in which levels of ␨-cytoskeletal binding (0.9 mM) and actin polymerization (0.5 either Fyn (Fig. 6A, lane 6) or pp60c-Src (Fig. 6A, lane 8) were mM) in this lysed cell system. The line of best fit was determined reduced, ␨-chain still associated with the reconstituted pellet upon ϭ ␨ by linear regression (r 0.98, for both sets of data) and the Km MgATP addition. This inhibition of -cytoskeleton association is was calculated from the x-intercept. Thus, ␨-cytoskeleton associ- not the result of coimmunoprecipitation of ␨-chain with Lck, nor ␨ ation had an apparent Km for ATP almost twice that measured for are the levels of soluble -chain in the cell lysate affected (data not actin polymerization, suggesting that actin polymerization occurs shown). Titration of the anti-Lck Ab (Fig. 6B) showed that inhi- at a lower ATP threshold and perhaps prior to ␨-chain association. bition of ␨-cytoskeleton association (Fig. 6B, bottom panel) was The Journal of Immunology 5495

FIGURE 3. Reaction kinetics and time course of ␨-actin association in a lysed cell system. A, Representative ATP concentration dependence of relative ␨ association and actin polymerization in the insoluble pellet at 37°C. Thymocytes were lysed and incubated with 0.2 mM Mg2ϩ and 0.25 or 2.5 mM ATP for 0 (control), 0.5, 1, 2, 4, and 8 min at 37°C. The pellets were separated by PAGE and Western blotted with anti-␨ or actin mAb. The rate of association with the reconstituted pellet was dependent on the ATP concentration. Data are representative of at least four experiments. B and C, ATP concentration dependence of relative ␨-cytoskeleton association and actin polymerization over time. Thymocytes or lymph node T cells were lysed and incubated with 0.2 mM Mg2ϩ and 0.25 (ᮀ), Downloaded from 0.5 (◊), 1.0 (E), or 2.5 (‚) mM ATP for 0, 0.5, 1, 2, 4, or 8 min at 37°C. The reaction was stopped by boiling the precip- itate in SDS sample buffer after rapid centrifugation. The pellets were separated by PAGE and Western blotted with anti-␨

mAb (B) or with an anti-actin mAb (C). http://www.jimmunol.org/ The amount of ␨-cytoskeleton association (B) or actin polymerization (C) relative to baseline at the 0 time control (arbitrarily assigned a value of 1) was determined by densitometry for values below saturation. Since the data for thymocytes and lymph node T cells were similar for each ATP concentration, the data are representative of at least four pooled experiments. D and E, Rate of ␨-cytoskeleton association and by guest on September 23, 2021 actin polymerization as a function of ATP concentration in a cell-free system. Rate measurements, calculated as the relative amount of ␨ bound per second (␨ bound/ sec, D) and actin polymerized per second (actin polym/sec, E), were determined from the linear portion of the time course plots (presented in B and C, respectively), for each ATP concentration. Each data point was generated from at least four experiments. F and G. Double reciprocal plots of data presented in D and E. The lines were obtained by least squares method (r ϭ 0.98). The x-intercepts rep-

resent the apparent Km for ATP (0.5 mM) ␨ of association (F) and the apparent Km for ATP (0.9 mM) of actin polymerization (G), in this reconstituted system.

directly correlated with both the loss of Lck from the cytoskeletal Fyn is not involved in ␨-cytoskeleton association and that c-Src pellet (Fig. 6B, middle panel) and with the amount of Lck immu- may act on upstream regulators to inhibit the association. noprecipitated from solution (Fig. 6B, top panel), with a corre- To conﬁrm these results in vivo, we analyzed ␨-cytoskeleton sponding decrease in tyrosine phosphorylation of the ␨-chain in association in intact thymocytes and lymph node T cells from mice both pellet and supernatant fractions (data not shown). Thus, in lacking Lck (Lck knockouts) (19) (Fig. 7) or Fyn (Fyn knockouts) this cell-free system, Lck plays a crucial role in inducing both (20) (Fig. 8). Thymocytes and lymph node T cells from Lck-neg- ␨-chain tyrosine phosphorylation and ␨-cytoskeleton association in ative (LckϪ) mice are reduced about 10-fold in number compared response to addition of MgATP. In contrast, titration of Fyn and to wild-type (Lckϩ) mice, express CD3 (although at reduced lev- Src in the cell lysates, by immunoprecipitation with increasing Ab els), and have measurable proliferative responses (19). Crosslink- concentrations, resulted in no effect on, or potentiated, ␨-cytoskel- ing of CD3␧ on T cells from LckϪ mice did not induce ␨-chain eton association, respectively (data not shown), suggesting that association with the insoluble pellet (Fig. 7A). In contrast, 5496 ROLE OF pp56Lck IN ␨-CHAIN-MICROFILAMENT ASSOCIATION

FIGURE 4. Dephosphorylation of the ␨-chain inhibits ␨-cytoskeleton association in solution and releases ␨-chain from the bound state. A, Thy- mocytes were lysed and incubated with 0.2 mM Mg2ϩ and 5 mM ATP (MgATP, lanes 1–3), in the absence (lane 1) or presence (lanes 2 and 3) of bacterial alkaline phosphatase (AP) added at 1 U/100 ␮l to the resus- pended reconstituted pellet (lane 2) or at 10 U/100 ␮l to the cell lysate prior to incubation with MgATP (lane 3). Data are representative of at least two experiments. B, Thymocytes were lysed and incubated in the presence of 0.2 mM Mg2ϩ and 2.5 mM ATP (MgATP) with increasing concentrations Downloaded from of bacterial alkaline phosphatase (AP) at8U(lane 1), 16 U (lane 2), 32 U (lane 3). ␨-Association with the cytoskeleton decreases in the presence of increasing concentrations of AP (MgATP pellet), while ␨-chain coordi- FIGURE 6. Immunoprecipitation of pp56Lck from the cell lysate inhibits nately increases in the supernatant (released ␨). The addition of phospha- ␨ association with the cytoskeleton. A, T cell lysates from resting cells were ␨ ␮ tase inhibitors (0.2 mM VO3, 10 mM NaF), block the release of -chain incubated in the presence of 50 g normal rabbit serum (lanes 1 and 2)or from the reconstituted pellet (MgATP pellet ϩ P-ase inhibitors). Data are Abs to Lck (lanes 3 and 4), Fyn (lanes 5 and 6), or Src (lanes 7 and 8). The

representative of at least two experiments. tyrosine kinases were immunoprecipitated from solution with protein A- http://www.jimmunol.org/ Sepharose, and incubated in MgATP (even lanes) or Mg2ϩ alone (odd lanes), as described. Data are representative of at least three experiments. ␧ Ϫ B, Inhibition of ␨ association with the cytoskeleton is inversely correlated crosslinking of CD3 on T cells from both Fyn-negative (Fyn ) Lck ϩ ␨ with the amount of pp56 in the cell lysate. T cell lysates were incubated and wild-type (Fyn ) mice induced association of -chain with the ␮ ␨ in the absence (lane 1) or presence of 80 g normal rabbit serum (lane 2) insoluble pellet (Fig. 8A). Analysis of immunoprecipitated -chain or polyclonal Abs to Lck (lanes 3–6) at increasing Ab concentrations of 10 by antiphosphotyrosine immunoblotting conﬁrmed that the ␨-chain ␮ ␮ ␮ ␮ ϩ Ϫ g(lane 3), 20 g(lane 4), 40 g(lane 5), or 80 g(lane 6). Lck was from Lck , but not Lck , T cells was tyrosine phosphorylated subsequently immunoprecipitated from solution with protein A-Sepharose. following TCR ligation (Fig. 7A). These data suggest that tyrosine Lysates were then incubated with MgATP (lanes 2–6)orMg2ϩ alone (lane phosphorylation is required for binding of ␨ to the cytoskeleton, 1) and processed as described. The protein A immunoprecipitates (top by guest on September 23, 2021 and implicate Lck, rather than Fyn, as the kinase involved in this panel) and reconstituted cytoskeletal pellets (middle and lower panels) ␨-chain tyrosine phosphorylation. Interestingly, the addition of were resolved on Western blots with anti-Lck (top and middle panels)or Ϫ Ϫ ␨ MgATP to T cell lysates from either Lck (Fig. 7B)orFyn (Fig. anti- (lower panel) Abs. Complete depletion of the Src family kinases by 8B) mice induced ␨-chain phosphorylation and ␨-cytoskeleton as- immunoprecipitation was not feasible, presumably due to protection of a subset of protein in micelles (data not shown). Data are representative of at sociation, suggesting that other Src family kinases may substitute least three experiments. for Lck in this cell-free assay and that all other requisite components for ␨-cytoskeleton association are present in the LckϪ mice.

These data suggest that tyrosine phosphorylation is required for ␨-cytoskeletal binding, and implicate Lck as the kinase involved in ␨-cytoskeleton association in vivo.

The SH2 domain of Lck PTK preferentially inhibits ␨- cytoskeleton association Since SH2 domains are the specific downstream targets of tyrosine-phosphorylated proteins, including the ␨-chain, we analyzed ␨-cytoskeleton association using competitive inhibition by syn- thetic SH2 domains. T cell lysates were incubated in the absence FIGURE 5. pp56Lck compartmentalizes to the cytoskeleton upon TCR or presence of 10 ␮M of GST or GST fusion proteins (Fig. 9A) ligation of intact T cells and in a cell-free system. A, Thymocytes were containing the SH2 domain of the Src-family kinases Fyn or Lck, lysed and the detergent-insoluble pellet isolated, following cross-linking of or the SHIP phosphatase, as a non-Src-family SH2 control domain. CD3␧ with anti-CD3␧ (145-2C11) mAb and goat anti-mouse (GAM) poly- These lysates were subsequently incubated with MgATP and an- clonal Ab (lane 2), or treated with GAM alone (lane 1). B, Alternatively, alyzed for ␨-cytoskeleton association. In comparison with the other 2ϩ cell lysates were incubated in the presence of 0.5 mM Mg (lane 1)or0.5 GST-fusion proteins, GST-Fyn SH2 or GST-SHIP SH2 (Fig. 9A, mM Mg2ϩ and 5 mM ATP (lane 2). The cell lysates were then fractionated lanes 3 and 6), or with MgATP (Fig. 9A, lane 2) or GST alone into reconstituted pellet fractions (lanes 1 and 2). Subsequently, all pellets were separated by PAGE and Western blotted with Abs specific for Lck. (Fig. 9A, lane 4), the GST-Lck SH2 fusion protein preferentially ␨ Inductive association of Lck with the cytoskeleton occurs in the presence inhibited -cytoskeleton association (Fig. 9A, lane 5). Both the of MgATP (B, lane 2) or cell activation (A, lane 2), with a concomitant GST-Fyn SH2 and the GST-Lck SH2 fusion proteins also inhibited depletion from the lysate supernatant (data not shown). Data are represen- Lck association with the reconstituted pellet (Fig. 9A, lanes 3 and tative of at least five experiments. 5), yet only GST-Lck SH2 domain resulted in the concomitant The Journal of Immunology 5497

FIGURE 7. Induction of ␨-cytoskeleton association is abrogated in intact T cells from pp56Lck knockout mice, but is rescued in the reconstituted cell-free system. A, Isolated lymph node T cells from wild-type (Lckϩ) or Lck-negative (LckϪ) mice were incubated in the absence (lanes 1 and 3) or presence (lanes 2 and 4) of an anti-CD3␧ (␣-CD3␧) mAb (145-2C11) and cross-linked with a secondary goat anti-mouse (GAM) Ab or treated with GAM (lanes 1 and 3) alone. The cells were then lysed in nonionic detergent and centrifuged to separate the detergent-insoluble pellet fractions. The pellets were separated by PAGE and Western blotted with anti-␨ or anti-phosphotyrosine mAbs (1). In comparison with Lckϩ T cells (compare lane 2 with lane 1), activated T cells from LckϪ mice (lane 4) showed neither increase in ␨-cytoskeleton association nor ␨ tyrosine phosphorylation relative to nonactivated controls (lane 3). Identical results were obtained in analyses of thymocytes and splenic T cells from LckϪ mice (data not shown). B, Isolated thymocytes from Lckϩ or LckϪ mice were lysed and cleared of the preexisting detergent insoluble pellet. The detergent soluble supernatants were incubated in the presence of 0.2 mM Mg2ϩ (Mg2; lanes 1 and 2) or 0.2 mM Mg2ϩ and 2.5 Downloaded from mM ATP (ATP; lanes 3 and 4, respectively). The cell lysates were then incubated at 37°C and centrifuged to separate pellet from the supernatant. The pellets were separated by PAGE and Western blotted with anti-␨ and anti-actin mAbs. ␨-Associated with the MgATP pellets from both Lckϩ and LckϪ mice. Identical results were obtained in analyses of lymph node and splenic T cells from LckϪ mice (data not shown).

inhibition of ␨-cytoskeleton association, suggesting that ␨-micro- Discussion

filament binding is dependent on interactions between phosphor- http://www.jimmunol.org/ We have previously shown that TCR␨-chain can be induced to ylated tyrosine residues in ␨-chain activation motifs and the SH2 domain of the Lck protein tyrosine kinase. associate with the actin cytoskeleton in a cell-free system (1). To determine if ␨-cytoskeleton association may be inhibited by Here, we expanded our characterization of this system, and, using the Fyn SH2 domain at higher inhibitory thresholds, due to lower this cell-free system, examined the role of Lck and tyrosine phos- ␨ binding specificity, T cell lysates were incubated in the absence or phorylation in -cytoskeleton binding. This reconstituted system is presence of increasing concentrations of the SH2 domain of Fyn specific for ATP and divalent cations since other nucleotides and (residues 144–255). As shown in Fig. 9B, ␨-cytoskeletal binding ATP analogs did not substitute for ATP in supporting cytoskeletal ␨ diminished with increasing concentrations of the SH2 peptide, rel- binding. The results of kinetic studies showed that -cytoskeleton by guest on September 23, 2021 ative to MgATP alone (Fig. 9B, lane 3) or to BSA (Fig. 9B, lane association has an apparent Km for ATP of 0.9 mM, providing 4), but at a 4- to 8-fold higher concentration requirement than with evidence that this is a saturable enzymatic reaction that is affected the Lck SH2 peptide. Since the phosphotyrosine-binding sites of by concentration and/or temperature and that ATP is utilized as a the Src family members are conserved (13), with the Lck and Fyn substrate as a possible phosphoryl donor in kinase reactions. In- SH2 domains sharing 60% sequence homology, the dose-depen- deed, tyrosine-phosphorylated ␨-chain was a predominant form of dent inhibition of ␨-cytoskeleton association by Fyn-SH2 domains ␨ that associated with the cytoskeleton. Dephosphorylation of suggests that Src SH2-containing polypeptides are involved, either ␨-chain by alkaline phosphatase inhibited or reversed the ␨-cy- directly or indirectly, in the interaction of ␨ with the microfilament toskeletal association, suggesting that early phosphorylation cytoskeleton. events regulate this interaction.

FIGURE 8. ␨-Cytoskeleton association is unaffected in intact T cells or T cell lysates from pp59Fyn knockout mice. A, Isolated lymph node T cells (LN; upper panel) and thymocytes (Thy; lower panel) from wild-type (Fynϩ; lanes 1–4) or Fyn-negative (FynϪ; lanes 5–8) mice were incubated in the absence (lanes 1 and 5) or presence (lanes 2 and 6) of an anti-CD3␧ (␣-CD3␧) mAb (145-2C11) and cross-linked with a secondary goat anti-mouse (GAM) Ab (lanes 4 and 8) or treated with GAM (lanes 3 and 7) alone. The cells were then lysed in nonionic detergent and centrifuged to separate the detergent- insoluble pellet fractions. The pellets were separated by PAGE and Western blotted with anti-␨ mAb (1). Both Fynϩ- and FynϪ-activated lymph node T cells and thymocytes (lanes 4 and 8, respectively) showed increases in ␨-cytoskeleton association relative to nonactivated controls. B, Isolated lymph node T cells (LN; upper panel) and thymocytes (Thy; lower panel) from Fynϩ or FynϪ mice were lysed and cleared of the preexisting detergent-insoluble pellet. The detergent-soluble supernatants were incubated in the absence (lanes 1 and 4) or presence of 0.2 mM Mg2ϩ (Mg2; lanes 2 and 5) or 0.2 mM Mg2ϩ and 2.5 mM ATP (ATP; lanes 3 and 6). The cell lysates were then incubated at 37°C and centrifuged to separate the pellet from the supernatant. The pellets were separated by PAGE and Western blotted with anti-␨ mAb. ␨ associated with the MgATP pellets from both Fynϩ and FynϪ lymph node T cells (LN; upper panel) and thymocytes (Thy; lower panel). 5498 ROLE OF pp56Lck IN ␨-CHAIN-MICROFILAMENT ASSOCIATION

lets and fibroblasts with their substrates. These data suggest that the properties of the reconstituted ␨-cytoskeleton association mimic those observed following TCR ligation on intact cells. This lysed cell system may therefore be useful as a model system for the biochemical study of the mechanisms and molecules involved in the interaction between TCR and the cytoskeleton. We observed that Lck and Fyn, but not Src, are components of the detergent insoluble pellet from intact cells and of the reconstituted cytoskeletal pellet from lysed cells. This compartmentalization of Lck and Fyn, like ␨-chain, is enhanced under conditions of TCR ligation or reconstitution following incubation with MgATP. Compartmentalization of these specific Src family kinases with ␨-chain and with the cytoskeletal fraction suggested an involvement of Lck or Fyn in the tyrosine phosphorylation-induced protein-protein interaction. We showed that reduction of Lck, but not Fyn or Src, inhibited the induction of ␨-cytoskeleton association. We also showed that TCR ligation of T cells from LckϪ, but not from FynϪ, mice failed to induce ␨-chain phosphor- Downloaded from ylation or ␨-cytoskeleton association. These results suggest that Lck is required for the tyrosine phosphorylation of ␨-chain leading to microfilament binding. That ␨-cytoskeleton association can be reconstituted in cell lysates from LckϪ mice suggests there is a FIGURE 9. The SH2 domain of Src family tyrosine kinases inhibits redundancy of kinase activity (22, 25, 26) in the reconstituted sys- ␨-cytoskeleton association. A, GST fusion protein containing the Lck SH2 tem, and that the machinery, components, and pathways leading to http://www.jimmunol.org/ domain preferentially inhibits ␨-cytoskeleton association. Thymocyte cell ␨-cytoskeletal interaction are intact in these mice, except for the lysates were incubated for2hintheabsence (lanes 1 and 2) or presence lack of Lck. Why then did depletion of Lck from wild-type (Lckϩ) of 10 ␮M GST (lane 4) or GST fusion proteins containing the SH2 domain cell lysates inhibit ␨-cytoskeleton association (Fig. 6)? We hypoth- of Fyn (lane 3), Lck (lane 5), or the SHIP phosphatase (lane 6). Subse- esize that an important adaptor/effector in the pathway leading to 2ϩ quently, these lysates were incubated in the presence of 0.2 mM Mg ␨-cytoskeleton association is depleted by coimmunoprecipitation (lane 1), or 0.2 mM Mg2ϩ and 1 mM ATP (lanes 2–6), and processed as with Lck. As discussed below, a polypeptide containing an SH3 described. These data are representative of two experiments. The mean amount of ␨ cosedimenting with the pellet relative to baseline (assigned a domain may function as just such a molecule. In summary, our relative density value of 1) was 8.9 for the MgATP pellet, 6.7 for GST data suggest that Lck is required both in vitro and in vivo as a

alone, 7.6 for GST-Fyn SH2, 3.2 for GST-Lck SH2, and 8.9 for GST-SHIP kinase and/or effector in the pathways leading to ␨-cytoskeletal by guest on September 23, 2021 SH2. The Lck SH2 fusion protein, therefore, inhibited ␨-cytoskeleton as- interaction. sociation by 50 to 70% relative to the other fusion proteins. B, Recombi- The data presented here also show that the Lck SH2 peptide nant Fyn-SH2 peptide inhibits ␨-cytoskeleton association. Thymocytes preferentially inhibited the ␨-cytoskeletal interaction and that the were lysed and incubated in the absence (lane 1) or presence of 0.5 mM SH2 peptide from Fyn was also inhibitory, in a dose-dependent Mg2ϩ (Mg, lane 2), 0.5 mM Mg2ϩ and 5 mM ATP (MgATP, lane 3), manner. The Fyn SH2 peptide was inhibitory at higher concentra- MgATP and 80 ␮M BSA (MgATP ϩ BSA, lane 4), or MgATP with increasing concentrations of recombinant Fyn-SH2 peptide ([SH2]) at 10 tions than Lck, suggesting that SH2-containing proteins are critical ␮M(lane 5), 20 ␮M(lane 6), 40 ␮M(lane 7), and 80 ␮M(lane 8). to the formation of this heteromeric protein complex, but retain ␨-Association with the cytoskeleton decreases in the presence of increasing their binding specificities. Since actin does not have SH2 domains, concentrations of Fyn-SH2 peptide. Data are representative of at least three intermediary signaling molecules and/or cytoskeletal-binding pro- experiments. teins containing these motifs must bridge the ␨-actin interaction. These intermediary proteins might, in turn, bind to actin via SH3 domains (12). That an SH2 domain of Lck, and to a lesser extent, In the lysed cell system presented here, as in intact cells, ␨ cosedi- of Fyn, inhibits the interaction, suggests that these kinases are ␨ mented with actin (Figs. 1–3) and the ␨-cytoskeleton association was involved as 1) adaptor/bridging proteins for the -actin association, disrupted by cytochalasin D (1) and EDTA. These data suggest that an 2) activated kinases that regulate other molecules in the binding intact microfilament array is necessary for compartmentalization of pathway, or 3) regulators of the availability of phosphorylated ty- ␨-chain with the cytoskeletal pellet and that microfilaments associate, rosines on the signaling ITAMs for other molecules, such as ZAP- either directly or indirectly, with the ␨-chain following T cell activa- 70. Recent studies have shown that the SH2 domain of Lck is tion. This association is, however, not driven by actin polymerization, essential for signal transduction events following TCR ligation since microfilaments can be isolated with minimal bound ␨-chain, (23) including the tyrosine phosphorylation of the ␨-chain and IL-2 under conditions that polymerize filaments in the absence of ATP production (27). (Figs. 1 and 2). In addition, the reconstituted ␨-cytoskeleton associa- These data are consistent with a model in which ␨ binds to actin, tion is dynamic over time (Fig. 3) and ␨ constructs with deletions in via an SH2-containing intermediary protein, subsequent to its ty- the terminal tyrosine do not cosediment with cytoskeleton (1), arguing rosine phosphorylation by Lck. This interaction may stabilize actin that this in vitro interaction is specific and not due to protein trapping polymerization and the formation of a signaling complex. We pos- in the insoluble pellet. Interestingly, the interaction of ␨-chain with tulate that these nascent ␨-cytoskeletal complexes, which we call actin is associated with increased actin polymerization both in the “cytoskeletal organizing centers,” serve to compartmentalize and cell-free system and in intact cells, suggesting that induced actin po- anchor activated enzymes critical for T cell signal transduction. lymerization may require the involvement of receptor association Another intriguing possibility comes from recent data demonstrat- (24), analogous to that seen following interaction of integrins on plate- ing that forces applied to cell surface receptors anchored to the The Journal of Immunology 5499

cytoskeleton quickly propagate to the cell interior (28). Thus, Ag- 10. Wange, R. L., N. Isakov, T. R. Burke, Jr., A. Otake, P. P. Roller, J. D. Watts, ␨ receptor ligation could translate into mechanical control of DNA R. Aebersold, and L. E. Samelson. 1995. F2(Pmp)2-Tam 3, a novel competitive inhibitor of the binding of ZAP-70 to the T cell antigen receptor, blocks early T and gene expression and regulation. cell signaling. J. Biol. Chem. 270:944. In conclusion, we have developed methods that reconstitute the 11. Isakov, N., R. L. Wange, W. H. Burgess, J. D. Watts, R. Aebersold, and ␨-cytoskeleton interaction in a cell-free system. Our findings indi- L. E. Samelson. 1995. ZAP-70 binding specificity to T cell receptor tyrosine- based activation motifs: the tandem SH2 domains of ZAP-70 bind distinct ty- cate that 1) intact cell membranes are not required for this inter- rosine based activation motifs with varying affinity. J. Exp. Med. 181:375. action, 2) the intracellular machinery required for this interaction 12. Koch, C. A., D. Anderson, M. F. Moran, C. Ellis, and T. Pawson. 1991. SH2 and shows redundancy but remains intact under cell-free conditions, SH3 domains: elements that control interactions of cytoplasmic signaling proteins. Science 252:668. and 3) the interaction is regulated by tyrosine phosphorylation, 13. Sonyang, Z., S. E. Shoelson, J. McGlade, P. Olivier, T. Pawson, X. R. Bustelo, through the involvement of the Src-family kinase, Lck, and its SH2 M. Barbacid, H. Sabe, H. Hanafusa, T. Yi, et al. 1993. Specific motifs recognized domain. The in vitro reconstituted system predicted a critical role by the SH2 domains of Csk, 3BP2, fps/fes, GRB-2, HCP, SHC, Syk, and Vav. Mol. Cell. Biol. 14:2777. for a particular Src-family kinase, a role that was confirmed in vivo 14. Caplan, S., S. Zeliger, L. Wang, and M. Baniyash. 1995. Cell-surface-expressed in our studies of mutant mice. In intact cells, Lck was required for T-cell antigen-receptor ␨-chain is associated with the cytoskeleton. Proc. Natl. TCR-induced tyrosine phosphorylation of ␨ and for ␨-cytoskeleton Acad. Sci. USA 92:4768. association. This cell-free system thus provides a powerful tool 15. Valitutti, S., M. Dessing, K. Aktories, H. Gallati, and A. Lanzavecchia. 1994. Sustained signaling leading to T cell activation results from prolonged T cell with which to further define and map the ␨-actin interaction. In receptor occupancy: role of T cell actin cytoskeleton. J. Exp. Med. 181:577. particular, this system will facilitate analysis of postulated inhib- 16. Jugloff, L. S., and J. Jongstra-Bilen. 1997. Cross-linking of the IgM receptor itors of the interaction, without the associated difficulties of intro- induces rapid translocation of IgM-associated Ig ␣, Lyn, and Syk tyrosine kinases to the membrane skeleton. J. Immunol. 159:1096. Downloaded from ducing peptides, Abs, or vectors into cells. In turn, in vivo use of 17. Caplan, S., and M. Baniyash. 1996. Normal T cells express two T cell antigen such inhibitors identified in our in vitro system will facilitate anal- receptor populations, one of which is linked to the cytoskeleton via ␨ chain and ysis of the role of ␨-cytoskeleton association in downstream events displays a unique activation-dependent phosphorylation pattern. J. Biol. Chem. 271:20705. of T cell signal transduction. 18. Pleiman, C. M., M. R. Clark, L. K. Timson Gauen, S. Winitz, K. M. Coggeshall, G. L. Johnson, A. S. Shaw, and J. C. Cambier. 1993. Mapping of sites on the Src Acknowledgments family protein tyrosine kinases p55blk, p59fyn, and p56lyn which interact with the effector molecules phospholipase C-␥2, microtubule-associated protein kinase, http://www.jimmunol.org/ We thank Dr. B. Jockusch for the gift of anti-actin mAb, Dr. T. Potter for GTPase-activating protein, and phosphatidylinositol 3-kinase. Mol. Cell. Biol. 13:5877. the gift of anti-Lck and anti-Fyn polyclonal antibodies, Kazuhiro Naka- 19. Molina, T. J., K. Kishihara, D. P. Siderovski, W. van Ewijk, A. Narendran, mura for the gift of GST-SHIP SH2, Sarah Johnson for the SH2 polypep- E. Timms, A. Wakeham, C. J. Paige, K.-U. Hartmann, A. Veillette, et al. 1992. tides, Dr. Tak Mak for Lck knockout mice, and V. Angkachatchai for Profound block in thymocyte development in mice lacking p56Lck. Nature assistance with time course experiments. 357:161. 20. Stein, P. L., H. M. Lee, S. Rich, and P. Soriano. 1992. pp59fyn mutant mice display differential signaling in thymocytes and peripheral T cells. Cell 70:741. References 21. Leo, O., M. Foo, D. H. Sachs, L. E. Samelson, and J. A. Bluestone. 1987. Iden- 1. Rozdzial, M. M., B. Malissen, and T. H. Finkel. 1995. Tyrosine phosphorylated tification of a monoclonal antibody specific for a murine T3 polypeptide. Proc. T cell receptor-␨ chain associates with the actin cytoskeleton upon activation of Natl. Acad. Sci. USA 84:1374. mature T lymphocytes. Immunity 3:623. 22. Van Oers, N. S., N. Killeen, and A. Weiss. 1996. LCK regulates the tyrosine by guest on September 23, 2021 2. Omann, G. M., R. A. Allen, G. M. Bokoch, R. G. Painter, A. E. Traynor, and phosphorylation of the T cell receptor subunits and ZAP 70 in murine thymo- L. A. Sklar. 1987. Signal transduction and cytoskeletal activation in the neutro- cytes. J. Exp. Med. 183:1053. phil. Physiol. Rev. 67:285. 23. Straus, D. B., A. C. Chan, B. Patai, and A. Weiss. 1996. SH2 domain function is 3. Ashwell, J. D., and R. D. Klausner. 1990. Genetic and mutational analysis of the essential for the role of the Lck tyrosine kinase in T cell receptor signal trans- T cell receptor. Annu. Rev. Immunol. 8:139. duction. J. Biol. Chem. 271:9976. 4. Wegener, A-M. K., F. Letourneur, A. Hoeveler, T. Brocker, F. Luton, and 24. Melamed, I., G. P. Downey, K. Aktories, and C. M. Roifman. 1991. Microfila- B. Malissen. 1992. The T cell receptor/CD3 complex is composed of at least two ment assembly is required for antigen-receptor-mediated activation of human B autonomous transduction molecules. Cell 68:83. lymphocytes. J. Immunol. 147:1139. 5. Cambier, J. C. 1995. New nomenclature for the Reth motif (or ARH1/TAM/ 25. Goves, T., P. Smiley, M. P. Cooke, K. Forbush, R. M. Perlmutter, and ITAM/YXXL). Immunol. Today 16:110. C. J. Guidos. 1996. Fyn can partially substitute for Lck in T lymphocyte devel- 6. Beaufils, P., D. Choquet, R. Z. Mamoun, and B. Malissen. 1993. The (YXXL/I) 2 opment. Immunity 5:417. signalling motif found in the cytoplasmic segments of the bovine leukemia virus envelope protein and Epstein-Barr virus latent membrane protein 2A can elicit 26. van Oers, N. S. C., B. Lowin-Kropf, D. Finlay, K. Connolly, and A. Weiss. 1996. ␣␤ early and late lymphocyte activation events. EMBO J. 12:5105. T cell development is abolished in mice lacking both Lck and Fyn protein 7. Nakayama, T., A. Singer, E. D. Hsi, and L. E. Samelson. 1989. Intrathymic tyrosine kinases. Immunity 5:429. signalling in immature CD4ϩCD8ϩ thymocytes results in tyrosine phosphoryla- 27. Lewis, L. A., C. D. Chung, J. Chen, J. R. Parnes, M. Moran, V. P. Patel, and tion of the T-cell receptor ␨-chain. Nature 341:651. M. C. Miceli. 1997. The Lck SH2 phosphotyrosine binding site is critical for 8. Rozdzial, M. M., R. T. Kubo, S. L. Turner, and T. H. Finkel. 1994. Develop- efficient TCR-induced processive tyrosine phosphorylation of the ␨-chain and mental regulation of the TCR ␨-chain. J. Immunol. 153:1563. IL-2 production. J. Immunol. 159:2292. 9. Hsi, E. D., J. N. Siegel, Y. Minami, E. T. Luong, R. D. Klausner, and 28. Maniotis, A., C. S. Chen, and D. E. Ingber. 1997. Demonstration of mechanical L. E. Samelson. 1989. T cell activation induces rapid tyrosine phosphorylation of connections between integrins, cytoskeletal filaments, and nucleoplasm that sta- a limited number of cellular substrates. J. Biol. Chem. 264:10836. bilize nuclear structure. Proc. Natl. Acad. Sci. USA 94:849.