Proc. Natl. Acad. Sci. USA Vol. 92, pp. 7819-7823, August 1995 Biochemistry

The WW domain of Yes-associated protein binds a -rich ligand that differs from the consensus established for Src homology 3-binding modules (protein-protein interactions/polyproline/modular domains) HENRY I. CHEN AND MARIUS SUDOL* Laboratory of Molecular Oncology, The Rockefeller University, 1230 York Avenue, New York, NY 10021 Communicated by Hidesaburo Hanafusa, The Rockefeller University, New York; NY, May 18, 1995 (received for review March 28, 1995)

ABSTRACT The WW domain has previously been de- Various studies have shown that SH3 domains may mediate scribed as a motif of 38 semiconserved residues found in protein localization to the plasma membrane or the cytoskel- seemingly unrelated proteins, such as , Yes- eton, as exemplified by the interaction between Grb2 and Sos associated protein (YAP), and two transcriptional regulators, that transduces signals from cell surface receptors to the Ras Rsp-5 and FE65. The molecular function of the WW domain pathway (13-15). Furthermore, similar to SH2 domains, SH3 has been unknown until this time. Using a functional screen domains may in fact mediate the targeting of proteins to their of a cDNA expression library, we have identified two putative substrates (16-21). ligands of the WW domain of YAP, which we named WBP-1 Recently we have characterized a Yes-associated protein and WBP-2. Peptide sequence comparison between the two (YAP) that binds to the SH3 domain of the Yes protoonco- partial clones revealed a homologous region consisting of a gene product via a proline-rich domain (22). Within the YAP proline-rich domain followed by a tyrosine residue (with the sequence, we have identified a protein motif of 38 amino acids shared sequence PPPPY), which we shall call the PY motif. first recognized to be shared by YAP, dystrophin, and a variety Binding assays and site-specific mutagenesis have shown that of other proteins (23, 24). This domain contains four well- the PY motif binds with relatively high affinity and specificity conserved aromatic amino acids among which are two tryp- to the WW domain of YAP, with the preliminary consensus tophan residues, hence the name WW domain. Computer- XPPXY being critical for binding. Herein, we have implicated aided structure analysis of the WW domain predicted a the WWV domain with a role in mediating protein-protein hydrophobic core consisting of these four aromatic residues interactions, as a variant of the paradigm set by Src homology flanked by 13 strands containing charged residues, reminiscent 3 domains and their proline-rich ligands. of the determined structure of the SH3 domain (25-27). The newly discovered domain has also been found in other proteins The progressive elucidation of the function of Src homology including , murine Nedd-4, yeast Rsp5, ORF-1 (a Ras (SH) 2 and SH3 domains, in addition to the identification of GTPase activator-related protein), and FE65 (a mammalian their ligands, has in the recent years spawned an era of protein implicated in the regulation of transcription) (23). remarkable developments in the fields of signal transduction These diverse proteins share the intriguing characteristic of and molecular oncology (1-3). As a result of this trend, two being involved in signaling or regulatory functions. As a result, new domains, the pleckstrin homology domain and the phos- it has been hypothesized that the WW domain may be a photyrosine interaction domain, have been identified and are putative protein-binding site important in the intracellular extensively studied in the context of signaling pathways (4-8). signaling network. The recognition of the existence of modular protein-binding Herein, we report the cloning and characterization of pro- domains has proven to be critical in understanding the intri- tein ligands for the WW domain of YAP-namely, WBP-1 and cacies of signal transduction where a variety of events occur. WBP-2 (WW domain bindingprotein). Furthermore, we have For example, proteins transiently interact with their upstream demonstrated that WBP-1 binds with high specificity to the and downstream partners, substrates are brought into prox- WW domain through a proline-rich domain with a sequence of imity of their catalytic centers, or proteins are localized to their PPPPY, which we have named the PY motif. Through site- subcellular compartments through the interaction of these directed mutagenesis, we have shown that the second and third domains with their ligands (9). proline residues, along with the tyrosine residue of the PY The SH2 domain was first described as a conserved sequence motif, are essential for binding to the domain of YAP (WW- of 100 amino acids in the noncatalytic domain of cytoplasmic YAP). In addition, the PY motif of WBP-1 does not bind with tyrosine kinases such as Src and Fps, but it rapidly became high affinity to the SH3 domains of arbitrarily chosen proteins. recognized in other molecules involved in signaling pathways Therefore, the PY motif seems to diverge from the classical and cellular transformation including Crk, Ras-GAP, and PXXP consensus of SH3-binding proline-rich domains and phospholipase C-,y. When the SH2 domain was later shown to may in fact represent another type of modular protein-binding bind with high affinity to phosphotyrosine residues, it was sequence, pending the discovery of additional WW domain quickly realized that the SH2 domain represented a means by ligands. which many molecules sharing a common signaling pathway may associate to regulate signal propagation from the extra- MATERIALS AND METHODS cellular environment into the cell and its nucleus (10). In Construction of Fusion Proteins. Primers corresponding to contrast, the SH3 domain was originally identified as a 50- the 5' and 3' of the cDNA for human amino-acid-long region of homology shared between the Crk ends, respectively, and Src oncogene products and phospholipase C-,y (11, 12). Abbreviations: GST, glutathione S-transferase; SH, Src homology; WBP, WW domain binding protein; YAP, Yes-associated protein. The publication costs of this article were defrayed in part by page charge *To whom reprint requests should be addressed at: Laboratory of payment. This article must therefore be hereby marked "advertisement" in Molecular Oncology, The Rockefeller University, Box 169, 1230 accordance with 18 U.S.C. §1734 solely to indicate this fact. York Avenue, New York, NY 10021.

Downloaded by guest on September 30, 2021 7819 7820 Biochemistry: Chen and Sudol Proc. Natl. Acad. Sci. USA 92 (1995) WW-YAP flanked by a 5' BamHI or a 3' EcoRI site (under- irradiation. The hybridization conditions were as described lined) are as follows: 5'-dCTATACGGATCCCAGTCT- (24, 32). The blots were washed for 30 min at room temper- TCTTTTGAGATACCT-3' and 5 '-dTACGACGAATIC- ature in 2x SSC/0.05% SDS and for 1 hr at 50°C in 0.1X GACTGGTGGGGGCTGTGACGTTCA-3'. The primers SSC/0.1% SDS. Removal of the WBP cDNA probes from the were subsequently used to amplify the WW-YAP gene frag- blot for subsequent hybridization with the human ,3-actin ment by PCR (Perkin-Elmer). The amplified fragments were probe was achieved by incubating the blot for 10 min in sterile subcloned in-frame into the pGEX-2TK vector between the H20 containing 0.5% SDS that was heated to 90°C. BamHI and EcoRI sites (Pharmacia). Expression of a gluta- thione S-transferase (GST) fusion protein migrating at the predicted molecular mass was verified by SDS/PAGE, and the RESULTS nucleic acid sequence of the construct was confirmed by direct Precipitation of the WW Domain Ligand from Cell Lysates. sequence analysis using the Sanger method (28). Fusion pro- To show the possible existence of specific protein ligands to teins of the cloned WBP-1 and fragments therein were con- WW-YAP, we conducted coprecipitation studies by incubating structed similarly. PCR-amplified fragments (using Pfu poly- GST-WW-YAP fusion proteins with lysates from various rat merase; Stratagene) of the full-length WBP-1 (5'-dAG- organs and established cell lines. Previous studies have shown TATCGGATCCAGCCACGGGGCTGGCCCTGTT-3' and that YAP is present at high levels in lung, ovary, cerebellum, 5 '-dGCATCCGAATTCGGTTCATGTCTCTT- and skeletal muscle, thus increasing the likelihood of detecting TAATGAG-3') and residues 1-74 of the N-terminal fragment the cognate ligand(s) in those organs as well (24). Western (BamHI primer, as above; EcoRI primer: 5'-dGCTATC- ligand blot analysis revealed a band of -38 kDa in size in lung, GAATTCAAACACCTTCTACATTTGTCC-3') were sub- ovary, and cerebellum, and an additional 34-kDa band present cloned into pGEX-2TK. An oligonucleotide and its comple- only in cerebellum when probed with 32P-labeled fusion pro- mentary strand coding for residues 34-43 (including the PY tein (Fig. 1B) but not with labeled GST (Fig. 1A). In addition, motif and flanking BamHI and EcoRI sites) of WBP-1 were precipitation with lysates of HeLa, A431, 3Y1, and v-src- annealed and then subcloned in-frame into the vector between transformed 3Y1 cell lines yielded a 38-kDa band as well (data the same two restriction sites (5'-dTACGTCGGATCCGG- not shown). CACACCGCCACCTCCTTACACTGTGGGCCGAAITCG- Cloning of Ligands for the WW Domain. We screened a TCTGC-3' and its complementary strand). Mutagenized PY in mouse embryo library with 32P-labeled GST-WW-YAP fusion GST fusion constructs was similarly constructed with a pair (sense protein and isolated three partial clones that were positive and antisense) of the above PY motif oligonucleotides, but each after four rounds of screening but that did not bind 32P-labeled of the respective codons for the residues PPPPY was alternately GST probe. Two of the clones proved to be identical. The replaced with the codon (GCA) for an alanine residue. predicted gene products of the two different clones, 700 bp and Purification and Labeling ofFusion Proteins. Recombinant 1.8 kb in length, were named WBP-1 and WBP-2, respectively SURE cells (Stratagene) were induced with 1.0 mM isopropyl (Fig. 2). Searches through GenBank and the European Mo- f3-D-thiogalactoside (Pharmacia) for 3-4 hr and then sonicated lecular Biology Laboratory data base with the predicted amino in phosphate-buffered saline (PBS; 137 mM NaCl/3 mM acid sequences revealed no significant homologies with known KCl/8 mM Na2HPO4 7H20/1.5 mM KH2PO4) with 1% Triton proteins. However, we then compared the sequences of WBP-1 X-100. Fusion proteins were then purified on a glutathione- and WBP-2 with each other and found a short proline-rich agarose column as described (29). For labeling purposes, -100 region of homology, which we named the PY motif. The ,jg of fusion protein was bound to 50 ,ul ofglutathione-agarose. invariant residues PPPPY make up the PY motif of both Labeling of the proteins was achieved as described (30). WBP-1 and WBP-2. WBP-1 possesses only one such PY motif, Precipitation of Cell Lysates. Cerebellum, lung, ovary, and but WBP-2 has two in the forward orientation and one in the skeletal muscle were dissected from rats, homogenized in reverse orientation. These data pointed to the PY motif as a RIPA buffer (10 mM Tris HCl, pH 7.4/5 mM EDTA/300 mM possible region of binding between the WBPs and WW-YAP. NaCl), and clarified by centrifugation. Agarose-glutathione beads (Pharmacia) bound with 100 ,ug of GST or GST-WW- A B YAP fusion protein were incubated with the above organ lysates [diluted 10-fold in Tris/Tween buffer (50 mM Tris HCl, 92- pH 7.5/100 mM NaCl/1 mM EDTA/0.1% Tween 20/1% ovalbumin/1 mM dithiothreitol; ref. 18)] with a final protein 68- concentration of 1 mg/ml for 24 hr at 4°C. The beads were then washed twice in 10 volumes of PBS and then separated by 43 - SDS/PAGE, which was subsequently transferred to nitrocel- lulose paper. The blots of GST or GST-WW-YAP precipi- -4- tated complexes were probed with 32P-labeled GST or GST- 29- WW-YAP fusion protein (Western ligand blotting), respec- tively. Molecular Cloning of Ligands. A 16-day mouse embryo 1 2 3 4 1 2 3 4 cDNA library (Novagen) plaque-lifted onto isopropyl P-D- FIG. 1. Coprecipitation of GST-WW-YAP and putative ligands thiogalactoside-saturated nitrocellulose filters was probed from various rat organ lysates. (A and B) Lanes: 1, lung; 2, ovary; 3, with the labeled GST-WW-YAP fusion protein. Four rounds cerebellum; 4, skeletal muscle. (A) GST was used for precipitation and of screening were performed, and the clones were finally subsequently as a labeled probe. (B) Putative ligands were precipitated purified by cre-mediated excision and incorporation into the and blots were probed with GST-WW-YAP. No specific bands were pEXlox vector as described (31). Both strands of the cDNA seen in skeletal muscle, even when the blots were exposed to radio- clones were analyzed by direct sequence analysis using the graphic film for >1 week. Notice the presence of diffuse bands representing the dimerization of GST and GST-WW-YAP fusion Sanger method (28). protein migrating at 26 kDa and 30 kDa, respectively (open arrows). Northern Blot Analysis. Poly(A)+ RNAs were isolated from Both GST and GST-WW-YAP precipitate from lung and ovary a 16 different human tissues from healthy donors of both sexes 28-kDa unknown protein, which apparently does not bind specifically (Clontech). The RNAs (2 jig per lane) were run on a dena- to WW-YAP. We were unable to discern any discrete bands above 70 turing formaldehyde/1.2% agarose gel, transferred to a kDa because of the high nonspecific binding in this region. Molecular charge-modified nylon membrane by blotting, and fixed by UV mass markers (in kDa) are indicated at left. Downloaded by guest on September 30, 2021 Biochemistry: Chen and Sudol Proc. Natl. Acad. Sci. USA 92 (1995) 7821

WBP-1 SHGAGPVPTGSLLDLRLLSAFKPPAYEDVVHIHPGTPPPPYT 41 molecular size of 2.0kb (top arrow). The probe also hybridized to a 1.3-kb transcript in testis, possibly an alternatively spliced VGPGYPWTTSSECTRCSSESSCSAHLEGTNVEGVSSQQSAL 82 message (bottom arrow). These experiments indicated that the putative ligands are encoded by distinct mRNAs, their protein PHQEGEPRAGLSPVHIPPSCRYRRLTGDSGIELCPCPDSSEGE 125 products are likely to have a low molecular mass, and they are PLKEARASASQPDLEDHSPCALPPDSVSQVPPMGLASSCGT 166 well conserved in mammals at the level of nucleic acid sequence. SHK* 169 Binding Assay of Cloned Ligands. We subsequently con- firmed the ability of the cloned proteins to bind to WW-YAP by constructing GST fusion proteins of the WBP-1 clone WBP-2 SSKNHSEGGGVIVNNTESILMSYDHVELTFNDMKNVPEAFK 41 (residues 1-169), the N-terminal region of WBP-1 (residues 1-74), and the 10-aa sequence (residues 34-43) GTPPP- GTKKGTVYLTPYRVIFLSKGKDAMQSFMMPFYLMKDCEIK 81 PYTVG, which includes the PY motif. WBP-1 was chosen for QPVFGANFIKGIVKAEAGGGWEGSASYKLTFTAGGAIEFGQ 122 these studies because of the existence of only a single PY motif in this particular clone, as opposed to the three PY motifs in RMLQVASQASRGEVPNGAYGYPYMPSGAYVFPPPVANGM 161 WBP-2, which may not all be functional. Western ligand blot analysis probed with 32P-labeled GST-WW-YAP showed YPCPPGYPYEPfPPEFYAGPPMMDGAMGYVQPEPfEYPGP 201 binding to the GST-ligand fusion proteins but not to GST MEPPVSGPSAPATPAAEAKAAEAAASAYYNPGNPHNVYM 241 alone (Fig. 4A). The WW domain bound to the GST- GTPPPPYTVG fusion protein with slightly lower affinity TSQPEEEYYPPEDKKTQ* 259 ('80% of the signal with N-terminal construct; Fig. 4A4, lanes FIG. 2. Deduced amino acid sequences from the open reading 2-5) despite the gel being slightly overloaded, suggesting that frames of partial cDNA clones for the WW-YAP ligands WBP-1 and optimal binding may rely on more distant residues flanking the WBP-2. The PY motifs, presumed to be the putative binding sites that core PY motif. When the ligands were probed with labeled are shared between the two independent clones, are shown in bold- GST, only low background binding was seen (Fig. 4B). The faced type and are underlined. The open reading frames were set by binding specificity of three GST-ligand fusion proteins to the frame of the T7 gene 10 product in the pEXlox cloning vector. labeled GST-WW-YAP probe was shown by competition with unlabeled peptides harboring the PY motif at concentrations Northern Blot Analysis of the Ligands. When probed with of 300 nM or 300 ,uM (Fig. 4 C and D, respectively). In random-primed fragments of the WBP-1 coding region, a addition, competition with a scrambled peptide at 300 ,uM survey of human organs revealed a transcript of -1.5 kb (Fig. 4E) showed no interference with binding. By incubating (arrow) that is present in most organs but at a significantly the blots of the GST-GTPPPPYTVG protein with various lower level in placenta, lung, liver, and kidney (Fig. 3A). The amounts of unlabeled PY peptides (GTPPPPYTVG) and a same blot probed with WBP-2 DNA (Fig. 3B) revealed tran- fixed concentration of labeled GST-WW-YAP, we estimated scripts that were more ubiquitously present but migrated at a the binding affinity of the WW domain to its ligand from the ± 15 16 observed IC50. At a competing peptide concentration of 750 A l 2 3 4 5 6 7 8 910 11 121314 250 nM, the binding of GST-WW-YAP to GST-GTPPP- *| | |I | - | || i i I i 7.5 PYTVG was reduced 50% from maximum binding (i.e., with- S S a s C | | X E -4.4 :gN N :FsI A 1 2 3 4 5 6 7 1 2 3 4 5 6 7 C M M - 2.4 68- |gS - 1.3 _J! 43- B _ -7.5 29- D -4.4 *l-|3__| - I -2.4 B __ - 1.3 68- -Em,, ^,,.,b,l ,,,.#,...... E 43- *:' S. C : -2.4 ^_^_ =§... -1.3 29- ....-'..... _! .::.: ::. :: FIG. 3. Northern blot of human tissues probed with cDNA of cloned ligands. (A-C) Lanes: 1, heart; 2, brain; 3, placenta; 4, lung; 5, FIG. 4. Binding assays with WBP-1 and putative binding domain. liver; 6, skeletal muscle; 7, kidney; 8, pancreas; 9, spleen; 10, thymus; Two independent clones of each GST fusion construct were chosen 11, prostate; 12, testis; 13, ovary; 14, small intestine; 15, colon; 16, and induced for protein expression to minimize the possibility of using peripheral blood leukocytes. (A) The blot was probed with labeled a mutated and nonfunctional form of the ligand. The amount of random-primed fragments of WBP-1 cDNA. There also seems to be protein in each lane (3-5 ,ug) did not vary significantly between each nonspecific hybridization of the probe to 18S and 28S rRNA (indicated blot as confirmed by Coomassie stain (data not shown). (A-E) Lanes: by dots). The difference in the migration patterns of the rRNAs was 1, GST; 2 and 3, GST-WBP-1-(34-43); 4 and 5, GST-WBP-1-(1-74); caused by a variation__f *_...... in the sizewof the gels. The apparent size of the 6 and 7, GST-WBP-1-(1-169). Nitrocellulose blot of GST-ligands ligand transcript, however, was aligned at 1.5 kb (arrow). (B) The blot probed with labeled GST-WW-YAP (A) with labeled GST, as a was probed with.labeled ,.,.s, ,,,random-primed;R.: WBP-2. (C) For normaliza- control for background binding (B), or with labeled GST-WW-YAP tion and to ensure the intactness of the RNA, the blot was hybridized competed with 300 nM (C) or 300 ,uM.....(D) of unlabeled: _zow decapeptide* with cDNA encoding human f3-actin. Two arrows indicate ,B-actin (biotin-GTPPPPYTVG; Research Genetics, Huntsville, AL) contain- mRNAs. Note that heart and skeletal muscle (and to a lesser degree, ing the PY motif. (E) Blot of GST-ligands*e probed with labeled prostate and small intestine) contain another isoform of ,3-actin GST-WW-YAP and competed with a scrambled:::..:decapeptide (biotin- mRNA that is of 1.6-1.8 kb (lower arrow). Molecular size markers (in GVYGPTPTPP). Molecular mass markers (in kDa) are indicated at kb) are shown at right. left. Downloaded by guest on September 30, 2021 7822 Biochemistry: Chen and Sudol Proc. Natl. Acad. Sci. USA 92 (1995)

out any competing peptide) as measured by densitometry A 1 2 3 4 5 6 7 (data not shown). Mutagenesis of PY Motif Sequence. Each of the residues in the motif PPPPY was systematically changed to alanine (A) by 43- aS constructing the appropriate coding oligonucleotides and then expressing them as GST fusion proteins, as described (33). Binding to WW-YAP was then assayed by probing blots of the mutant ligand proteins with 32P-labeled GST-WW-YAP (Fig. 5A). Binding was virtually abolished in the fusion proteins 43- GST-GTPAPPYTVG, GST-GTPPAPYTVG, and GST- GTPPPPATVG. In addition, binding was reduced -2-fold in 29-- GST-GTAPPPYTVG and GST-GTPPPAYTVG as com- pared with wild-type PY. Therefore, all residues of the PY motif appear to be important in binding the WW domain, but FIG. 6. Binding specificity of the PY motif. (A and B) Lanes: 1, P2 (amino acid numbered according to its position in the PY GST; 2, GST-WW-YAP; 3, GST-WW-dystrophin; 4, GST-SH3- motif), P3, and Y5 are crucial although not entirely sufficient GAP; 5, GST-SH3-Abl; 6, GST-SH3-Fyn; 7, GST-SH3-Yes. (A) Blots for optimal binding. In summary, the PY motif does not appear of the above purified fusion proteins were probed with 32P-labeled to conform to the PXXP consensus of SH3-binding proline- GST-GTPPPPYTVG protein. (B) A Coomassie stain of a replica gel confirmed equal amounts of protein in each lane (2 ,tg). Molecular rich domains. mass markers (in kDa) are indicated at left. Specificity of WW-PY Binding. The SH3 domains of Yes (22), Fyn [a gift from G. Cheng and D. Baltimore (Massachu- The structure of the WW domain appears to resemble that of setts Institute of Technology)], Abl, and Ras-GAP were SH3 domains in that a hydrophobic core of conserved aromatic expressed as GST fusion proteins utilizing the previously residues is surrounded by ,B loops containing charged amino described method [the latter two were kindly provided by B. acids. This observation, along with the fact that the WW ligand Knudsen (The Rockefeller University)]. Nitrocellulose blots of contains a polyproline sequence, suggests that the WW do- these fusion proteins were probed with 32P-labeled GST- main-cognate ligand interaction may be a variant of the GTPPPPYTVG protein. The labeled probe bound to GST and paradigm set by the SH3 domain and its proline-rich ligand. the SH3 domains at the level of background binding as The SH3 domain binds to proline-rich motifs with the prelim- compared to the PY-WW interaction (Fig. 6A). Moreover, no inary consensus of PXXP, which contributes to the formation significant interaction was evident between GST-WW- of a left-handed helical structure known as polyproline helix dystrophin [a gift of C. Bougeret (The Rockefeller University)] type II (26). It is possible that the PY motif may bind in a and the labeled probe. The binding region from WBP-1 thus similar manner to the hydrophobic pocket of the WW domain, seems to exhibit binding specificity for WW-YAP and not for although further work is necessary to confirm the actual the panel of SH3 domains and the WW domain of dystrophin structure of the WW domain and the PY motif. Furthermore, (WW-dystrophin). It is likely that the residues surrounding the it remains to be seen whether the aromatic residues are PY motif may impart specificity for a particular WW do- essential for binding as they are sometimes partially absent in main(s). a few SH3 domains [e.g., Crk SH3(2) and Grb2 SH3 domains]. It is not clear whether the PY motif typifies a variant of the DISCUSSION polyproline helix type II structure that binds to SH3 domains. However, since the PY motif does not conform to the PXXP We have isolated two cDNA clones that encode putative ligand motif of SH3-binding ligands in our mutational studies and proteins, named WBP-1 and WBP-2, which bind the human does not bind to the SH3 domains of GAP, Abl, Fyn, and Yes, WW-YAP in vitro. Within the sequence of the ligands, we it may in fact represent another type of modular protein- identified the pentapeptide PPPPY as being involved in pro- binding sequence. In any event, further analysis of the existing tein-protein interactions and mapped the specific amino acids PY motif, including the flanking amino acids, as well as the in this PY motif that are required for binding to WW-YAP. isolation of more WW domain ligands are crucial before we can establish a consensus sequence for binding to WW do- A 1 2 3 4 5 6 7 mains. Despite our results showing that WW-dystrophin fails

68- to bind to the binding domain of the WBP-1, we still are unable to exclude the possibility that the ligand of WW-dystrophin 43- may contain a PY motif as well. It is conceivable that residues 29- # - flanking the PY motif exert limitations on binding, while the B PY motif itself acts as the core component needed to interact 68 - with a WW domain. As a result, it is important to determine PY 43- the relative specificity and binding affinity of the motif for 29- _ the WW domains of different proteins. Furthermore, the WWWWWWU presence of the tyrosine residue in the PY motif invites speculation on the possible role of its in FIG. 5. Mutational analysis of PY motif. The residues making up regulating the binding to the WW domain. It is unknown the PY motif were each changed to alanine. (A and B) Lanes: 1, GST; whether the PY motifs of WBP-1 and WBP-2 are actually 2, GST-GTPPPPYTVG (wild type); 3, GST-GTAPPPYTVG; 4, phosphorylated in vivo. However, our studies were performed GST-GTPAPPYTVG; 5, GST-GTPPAPYTVG; 6, GST-GTPPPAY- on bacterially expressed proteins, suggesting that tyrosine TVG; 7, GST-GTPPPPATVG. (A) GST fusion proteins expressing phosphorylation of the PY motif is not required for binding. A each of these mutated PY motifs along with the five invariant flanking negatively charged phosphate group on the tyrosine residue residues were then assayed for binding activity to labeled GST-WW- in vivo as a means YAP (arrow). (B) The amount of protein loaded in each well (2 ,ug) could in fact disrupt the WW-PY interaction was equivalent as confirmed by Coomassie staining of a replica gel. of regulating the process of signal propagation. The lower band in each lane most likely represents a product of limited A data base search revealed that a PPPPY sequence is proteolysis of the fusion protein. Molecular mass markers (in kDa) are present in two viral proteins: Gag of avian retroviruses and indicated at left. LMP2 of the Epstein-Barr virus (EBV). Interestingly, both Downloaded by guest on September 30, 2021 Biochemistry: Chen and Sudol Proc. Natl. Acad. Sci. USA 92 (1995) 7823 proteins are believed to affect cellular transformation and 5. Haslam, R. J., Kolde, H. B. & Hemmings, B. A. (1993) Nature modulate protein-tyrosine kinases. The putative PY motif is (London) 363, 309-310. present in the Gag protein fused to Fps, Yes, and Crk 6. Kavanaugh, W. M. & Williams, L. T. (1994) Science 266, 1862- oncogene products. When overexpressed in retroviral con- 1865. structs without the in Crk transforms 7. Blaikie, P., Immanuel, D., Wu, J., Li, N., Yajnik, V. & Margolis, Gag portion, contrast, B. (1994) J. Bio. Chem. 269, 32031-32034. fibroblasts to a much lower extent (34, 35). Numerous studies 8. Bork, P. & Margolis, B. (1995) Cell 80, 693-694. have implicated the viral Gag sequences as modulators of 9. Cohen, G. B., Ren, R. & Baltimore, D. (1995) Cell 80, 237-248. transforming potential rather than simple vehicles for the 10. Koch, C. A., Anderson, D., Moran, M. F., Ellis, C. & Pawson, T. stable retroviral expression of a cellular gene (e.g., ref. 36). It (1991) Science 252, 668-674. would be interesting to investigate the transforming potential 11. Mayer, B., Hamaguchi, M. & Hanafusa, H. (1988) Nature (Lon- of various Gag-oncogene fusion proteins in which the PY don) 332, 272-275. motif is rendered inactive by site-directed mutagenesis. LMP2 12. Stahl, M. L., Ferenz, C. R., Kelleher, K. L., Kriz, R. W. & Knopf, is an integral membrane protein encoded by the EBV genome J. L. (1988) Nature (London) 332, 269-272. (37), and it interacts biochemically with LMP1, the transform- 13. Li, N., Batzer, A., Daly, R., Yajnik, V., Skolnik, E., Chardin, P., ing protein of EBV (38). The N-terminal domain of LMP2 was Bar-Sagi, D., Margolis, B. & Schlessinger, J. (1993) Nature shown to (London) 363, 85-88. associate with two protein-tyrosine linases of the Src 14. Rozakis-Adcock, M., Fernley, R., Wade, J., Pawson, T. & Bow- family, Lyn and' Fyn (39). Since there are two putative PY tell, D. (1993) Nature (London) 363, 83-85. motifs located in the cytoplasmic part of the N-terminal 15. Bar-Sagi, D., Rotin, D., Batzer, A., Mandiyan, V. & Schlessinger, domain of LMP2 (40), it is conceivable that a molecule, J. (1993) Cell 74, 83-91. perhaps similar to YAP, could serve as an adaptor to bring 16. Cicchetti, P., Mayer, B. J., Thiel, G. & Baltimore, D. (1992) LMP2 and the Lyn or Fyn kinases together to stimulate cell Science 257, 803-806. growth and transformation of B lymphocytes (39). Although it 17. Ren, R., Mayer, B. J., Cicchetti, P. & Baltimore, D. (1993) remains to be proven that the PPPPY sequences in Gag or in Science 259, 1157-1161. LMP2 constitute a functional protein-binding domain, one 18. Feller, S. M., Knudsen, B. & Hanafusa, H. (1994) EMBO J. 13, parallel to keep in mind is the Nef protein of the human 2341-2351. immunodeficiency virus, a 19. Flynn, D. C., Leu, T. H., Reynolds, A. B. & Parsons, J. T. (1993) which contains PXXP motif that Mol. Cell. Biol. 13, 7892-7900. binds to the SH3 domain of the tyrosine kinases Hck and Lyn 20. Fumagalli, S., Totty, N. F., Hsuan, J. J. & Courtneidge, S. A. to effect a higher replicative potential (41). It is not surprising (1994) Nature (London) 368, 871-874. that viruses through evolution have managed to take advan- 21. Taylor, S. J. & Shalloway, D. (1994) Nature (London) 368, tage of existing host signaling pathways for enhanced self- 867-871. preservation. The human immunodeficiency virus'Nef protein, 22. Sudol, M. (1994) Oncogene 9, 2145-2152. and perhaps Gag and LMP2, apparently could be a few 23. Bork, P. & Sudol, M. (1994) Trends Biol. Sci. 19, 531-533. examples of many' such strategies. 24. Sudol, M., Bork, P., Einbond, A., Kumar, K, Druck, T., Negrini With the identification of Crk, a group of proteins with M., Huebner, K. & Lehman, D. (1995) 1. Biol. Chem. 270, multiple modular protein-binding domains and no apparent 14733-14741. has considerable interest. 25. Wu, X., Knudsen, B., Feller, S. M., Zheng, J., Sali, A., Cowburn, catalytic domain gained These pro- D., Hanafusa, H. & Kuriyan, J. (1995) Structure 3, 215-226. teins, including Grb2, Nck, and Shc, act as adaptor molecules, 26. Musacchio, A., Wilmanns, M. & Saraste, M. (1994) Prog. Biophys. charged with the responsibility of bringing together various Mol. Biol. 61, 283-297. components of a pathway by virtue of protein-binding domains 27. Yu, H., Chen, J. K., Feng, S., Dalgarno, D. C., Brauer, A. W. & such as SH2 and SH3 to propagate the signal. YAP, although Schreiber, S. L. (1994) Cell 76, 933-945. still unknown in-function, possesses not only a WW domain but 28. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl. Acad. also a proline-rich domain that binds to the SH3 domain of Sci. USA 74, 5463-5467. Yes. YAP may in fact represent another example of an adaptor 29. Smith, D. B. & Johnson, K. S. (1988) Gene 67, 31-40. molecule. Nonetheless, at this time the biological significance 30. Kaelin, W. G., Jr., Krek, W., Sellers, W. R., DeCaprio, J. A., of the interaction between the WW domain and the PY motif Ajchenbaum, F., Fuchs, C. S., Chittenden, T., Yue, L., Famham, is not clear. The genetic manipulation of the WW module in P. J., Blanar, M. A., Livingston, D. M. & Flemington, E. K a (1992) Cell 70, 351-364. dystrophin, molecule that has been implicated in a specific 31. Palazzolo, M. J., Hamilton, B. A., Ding, D., Martin, C. A., disease phenotype (Duchenne and Becker muscular dystro- Raghavan, K. V., Mierendorf, R. C., Mead, D. A., Meyerowitz, phy), and other genetic approaches to analyze the WW domain E. M. & Lipschitz, H. D. (1990) Gene 88, 25-36. of the yeast protein Rsp-5 should provide us with useful 32. Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seid- biological correlates. man, J. G., Smith, J. A. & Struhl, K., eds. (1987) CurrentProtocols in Molecular Biology (Wiley, New York). We give special thanks to Dr. Hidesaburo Hanafusa, who was the 33. Knudsen, B. S., Zheng, J., Feller, S. M., Mayer, J. P., Burrell, first to notice subtle similarities between the linear sequences of the S. K, Cowburn, D. & Hanafusa, H. (1995) EMBO J. 14, 2191- WW and SH3 domains and who encouraged us to search for the 2198. potential protein ligand. We also acknowledge Drs. Peer Bork, Andrea 34. Reichman, C. T., Mayer, B. J., Keshav, S. & Hanafusa, H. (1992) Musacchio, Hartmut Oschkinat, and Matti Saraste for stimulating Cell Growth Differ. 3, 451-460. discussions. We thank Dr. Beatrice Knudsen for valuable comments on 35. Mayer, B. J. & Hanafusa, H. (1990) J. Virol. 64, 3581-3589. the manuscript. This work was supported by Grant CA01605 from the 36. Foster, D. A., Shibuya, M. & Hanafusa, H. (1985) Cell 42, National Cancer Institute, by the Council for Tobacco Research-USA, 105-115. Inc., Grant 3035, by a Human Frontier Science Program grant to M.S., 37. Kieff, E. & Liebowitz, D. (1990) in Virology, eds. Fields, B. N. & and by National Institutes of Health Grant 5T32GM07739-16 to H.I.C. Knipe, D. M. (Raven, New York), pp. 1889-1920. 38. Longnecker, R. & Kieff, E. (1990) J. Virol. 64, 2319-2326. 1. Pawson, T. & Schlessinger, J. (1993) Curr. Biol. 3, 434-442. 39. Burkhardt, A. L., Bolen, J. B., Kieff, E. & Longnecker, R. (1992) 2. Birge, R. B. & Hanafusa, H. (1993) Science 262, 1522-1524. J. Virol. 66, 5161-5167. 3. Sudol, M. (1993) in The Molecular Basis of Human Cancer, eds. 40. Sample, J., Liebowitz, D. & Kieff, E. (1989)J. Virol. 63, 933-937. Neel, B. & Kumar, R. (Futura, Mt. Kisco, NY), pp. 203-224. 41. Saksela, K., Cheng, G. & Baltimore, D. (1995) EMBO J. 14, 4. Mayer, B. & Baltimore, D. (1993) Trends Cell Biol. 3, 8-13. 484-491. Downloaded by guest on September 30, 2021