an sh3 domain and proline-rich sequence mediate an interaction

Communication Vol. 269, No. 19, Issue of May 13, pp. 13752-13755, 1994 THE JUURNAL OF BIUL~~ICAL CHEMISTRY

0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.

An SH3 Domain and Proline-rich Sequence Mediate an Interaction between Two Components of the Phagocyte NADPH Oxidase Complex* (Received for publication, February 24, 1994, and in revised form,

March 11, 1994)

Peter Finant§, Yazuaki Shimizum, Ivan Goutli, Justin Hsuanll, Oanh 'Ikuongjl, Christopher Butcher$, Paul Bennett$, Michael D. Waterfieldll**, and Stuart Kelliet From the SYamanouchi Research Institute, Littlemore Hospital, Oxford OX4 4xN, United Kingdom, the (Ludwig Institute for Cancer Research, 91 Riding House Street, London W l P 8Bl: United Kingdom, and the **Department of Biochemistry and Molecular Biology, University College, London WClE 6Bl: United Kingdom

Neutrophils possess a multicomponent NADPH oxidase system capable of producing large quantities of superoxide in a process known as the respiratory burst (1). Upon stimulation of a phagocytic cell, two cytosolic components of the oxidase, p67phm and p47phm, associate with a membrane-bound flavocytochrome b and a small GTP-binding protein to form a functional enzyme complex. Each of the Phox proteins contains two src homology 3 (SH3) domains, which are of unknown function but are potential mediators of protein-protein interactions between components of the activated oxidase. We have isolated a 47-kDa protein from lysates of differentiated HL60 cells that specifically bound to the carboxyl-terminal SH3 domain of p67Pho" and not to any other SH3 domain tested. This protein was identified as p47Ph"', and the putative SH3 domain binding site was located to a carboxyl-terminal proline-rich region. Pro- line-rich synthetic peptides based on this carboxyl-terminal region specifically inhibited the binding of ~ 4 7 ~ ~ - to the carboxyl-terminal SH3 domain of p67pho", and sequential truncation defined a unique minimal sequence, which, although similar, does not match the consensus sequence defined for other SH3-binding proteins.

The role of SH2 domains in the assembly of protein com- plexes containing phosphorylated tyrosine residues has been firmly established (2); however, the function of SH3 domains remains more elusive. Like SH2 domains, SH3 domains are thought to be involved in mediating protein-protein interactions between signaling components downstream of membrane-bound receptors (3), and recently, a number of potential

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solelv to indicate this fact. 6 To whom correspondence should be addressed. Tel.: 865-747100;

F&: 865-748974.

gaoka, Tsukuba-Shi, Ibaraki 305, Japan. ll Present address: Yamanouchi Pharmaceutical Co., Ltd., 21 Miyuki-

Printed in U.S.A.

SH3 domain-binding proteins have been reported. The first of these to be isolated, 3BP-1, binds to the SH3 domain of c-ab1 in vitro and contains a region with sequence homology to the GTPase-activating protein for the monomeric GTPase Rho (4). In addition, an association between the GTPase dynamin and selected SH3 domains has also been described (5). The SH3 domains of src-related tyrosine kinases have been shown to associate with phosphatidylinositol 3"kinase p85 subunit (6, 71, and the signaling pathway from cell-surface receptors to Ras utilizes an adaptor protein Grb2 that couples to activated receptors through an SH2 domain and interacts with a downstream signaling component, Sos, via at least one SH3 domain (8-12). The association of SH3 domains with cytoskeletal components has also been reported (13, 14).

The roles of SH2 and SH3 domains have been predomi- nantly investigated in proliferation-related signaling in adher- ent cells such as fibroblasts or nonadherent cells such as lym- phocytes; however, in some cells SH2 and SH3 domain- containing proteins may also be important regulatory molecules in responses unrelated to cell division. In particular, two components of the phagocyte NADPH oxidase, p47p"" and p67Phox, each possess two SH3 domains and are normally cytosolic proteins. Activation of the oxidase is preceded by translocation of p47phox and p67p"" to the plasma membrane to form a complex with cytochrome b,,, and a small GTP-binding protein (15, 16). It is therefore possible that these protein-protein interactions are mediated by SH3 domains. We report here the identification of p47P"" as a protein that binds to the carboxyl-terminal SH3 domain of p67ph0" via a novel proline- rich motif.

EXPERIMENTAL PROCEDURES Reagents-The bacterial expression vector pGEX-4T-1 was pur-

chased from Pharmacia LKB Biotechnology Inc. Glutathione-agarose beads were obtained from Sigma. Synthetic peptides were synthesized by Zinsser Analytic (Maidenhead, United Kingdom) and Alta Bioscience (Birmingham, United Kingdom). Rabbit polyclonal antisera raised to recombinant p47ph" expressed as a GST' fusion protein in Escherichia coli was a gift from A. W. Segal.

Cell Culture-HL6O cells were cultured in RPMI 1640 and 10% heat- inactivated fetal calf serum containing 2 mM L-glutamine at 37 "C in 5% CO,. Cells were differentiated down the neutrophilic pathway by the addition of Me,SO to a final concentration of 1.25% for 5 days.

GZutathione S-Zkansferase Fusion Proteins-DNA sequences encod- ing the NH,-terminal SH3 domains of ~ 4 7 ~ ~ " (amino acids 157-215) and ~ 6 7 ~ ~ " (amino acids 237-398) were amplified by polymerase chain reaction from a human neutrophil phage library (Clontech) and cloned into the pGEX-4T-1 expression vector. Polymerase chain reaction prim- ers corresponding to each end of the SH3 domains contained EcoRI and Sal1 sites (p47N-SH3) or EcoRI and XhoI sites (p67N-SH3) to facilitate cloning. The pGEX constructs were sequenced using the Sequenase system (U. S. Biochemical Corp.). The remaining pGEX-SH3 constructs were produced as described previously (5). Constructs were trans- formed into E. coli XL1 and expression of GST fusion proteins performed as described previously (17).

GSTSH3 Binding Assays-AEnity matrices were prepared by immobilizing 25 pg of fusion protein on 50 1.11 of glutathione-agarose beads (Sigma). Differentiated HL60 cells were pelleted and washed twice in phosphate-buffered saline. Cell pellets were solubilized in lysis buffer (50 mM Tris, pH 7.5,5 mM EGTA, 2% (v/v) Triton X-100,75 mM NaC1,0.5 mM phenylmethylsulfonyl fluoride) and clarified by centrifugation at 14,000 x g for 15 min at 4 "C. Cell lysate was mixed with the GST-SH3

The abbreviations used are: GST, glutathione S-transferase; PAGE, polyacrylamide gel electrophoresis.

13752

Function of NADPH Oxidase SH3 Domain 13753

68-

45-

ferentiated HL60 cells to immobilized GST-SH3 fusion proteins. FIG. 1. In v i t r o binding of detergent-soluble proteins from dif-

HL60 cell lysates were incubated with glutathione-agarose beads (Beads) and beads bound with glutathione S-transferase (GST) or GST fusion proteins containing the carboxyl-terminal SH3 domain of p67"h"x (p67C), amino-terminal SH3 domain of ~ 6 7 ~ ~ ' " (p67N) , carboxyl-terminal SH3 domain of p47p"'" (p47C), amino-terminal SH3 domain of p47ph"" (p47N) , phospholipase Cy SH3 domain (PLCy), c-fgr SH3 domain (fgr), SH3 domain of the p85a subunit of phosphatidylinositol 3-kinase (P85). or the SH3 domain of c-src (src). Bound proteins were eluted with SDS-PAGE sample buffer, resolved by SDS-PAGE (10% polyacrylamide), and visualized by silver staining.

1

56

111

166

221

2 7 6

331

386

MGDTFIRHIALUjFEKRFVPSOHYVYMFLVKWQDLSEKWYRR~EIYEFHKTLK HIALLGFEK(Pepl] FTEIYEFHK[PepZ]

EMFPIEAGAINPENRIIPHLPAPKWFDGQRAAENRPGTLTEYCSTLMSLPTKISR IIPHLPA(Pep31

CPHLLDFFKVRPDDLKLPTDNQTKKPETYLMPKDGKSTATDITGPIILQTY~IA

PSMYLQKSGQDVSQAQRQIKRGAPPRRSSIRNAHSIHQRSRKRLSQDAYRRNSVR

FUlQRRRQARPGPQSPGSPLEEERQTQRSKPOPAVPPRPSADLILNRCSESTKRK

LASAV

the 47-kDa protein that bound to p67Ph"C-SH3 domain with hu- FIG. 2. Alignment of tryptic peptide sequences derived from

man ~ 4 7 ~ ~ " sequence. Peptide sequences obtained from the tryptic peptides (PEPI, PEPZ, PEP3) are indicated in bold below the appro- priate sequence of p47ph". The proline-rich sequence that is a potential SH3-binding region is underlined, and a homologous peptide, P2, was synthesized and used in further studies.

affinity matrices for 3 h a t 4 "C. The beads were then washed exten- sively in wash buffer (50 mM Tris, pH 7.5, 0.1% (v/v) Triton X-100, 10% (v/v) glycerol). Binding proteins were eluted by boiling in SDS-PAGE sample buffer, resolved by SDS-PAGE, and visualized by silver staining. In addition, binding assays were performed in the presence of proline- rich synthetic peptides at concentrations of 750 and 400 PM, as indicated in the relevant figure legends.

Protein Sequencing-A 47-kDa polypeptide that bound to the COOH-terminal SH3 domain of p67ph" was purified from 5 x lo9 differentiated HL60 cells by GST-SH3 affinity chromatography as described under "GST-SH3 Binding Assays." The protein sample (300 pl) was concentrated by centrifugal filtration and applied to SDS-PAGE. The 47-kDa polypeptide was identified by Coomassie Blue staining, excised, and digested with trypsin in the gel slice. Peptides were ex- tracted from the gel using detergents and separated by tandem ion-ex- change and reverse phase high performance liquid chromatography using a Hewlett-Packard 1090M chromatograph and diode array de- tection. Purified peptides were sequenced using fast cycle, automated Edman chemistry on an Applied Biosystems 477A sequenator, modi- fied as described (18).

Western Blotting-Following protein sequencing, the identity of the 47-kDa SH3 domain-binding protein was confirmed by Western blotting. SH3 domain-binding proteins were purified as previously described (under "GST-SH3 Binding Assays") and were resolved by SDS- PAGE, transferred to nitrocellulose filters, and incubated with an antibody against p47Phox (a gift from A. W. Segal).

68-

45- -

FIG. 3. In vitro binding of p47phm to the COOH-terminal SH3 of p67ph". SH3 domain-binding proteins were purified as described under "Experimental Procedures" and in Fig. 1. Following SDS-PAGE, proteins were transferred to nitrocellulose and probed with antibodies raised to p47Phor. Gel lanes are labeled as for Fig. 1.

" "

p -200

0 * .. .

-116 -95

-68

-45

-29

FIG. 4. Inhibition of p47P"-p67P"C-SH3 domain binding by the p47Phar-derived proline-rich peptide, P2. Binding of ~ 4 7 ~ ~ ~ to the COOH-terminal SH3 domain of ~67'"'"' was determined in the presence of synthetic peptides derived from proline-rich sequences of dynamin, peptide P1 (lane 4 ) ; p47Ph"', peptide P2 (lane 5); ~ 6 7 ~ ~ ' " , peptide P3 (lane 6 ) ; cytochrome b,,,a, peptide P4 (lane 7 ) ; phosphoty- rosine protein phosphatase (PTPase) lB, peptide P5 (lane 8); vinculin peptide P6 (lane 9); or in the absence of any peptide (lane 3). Lane 1 shows p67Ph"C-SH3 domain without cell lysate or peptide, and lane 2 shows control GST affinity matrix plus cell lysate. The full peptide sequences and inhibitory activities are listed in Table I. HL60 cell lysate was incubated with the p67Ph"C-SH3 domain affinity resin in the presence of the proline-rich peptides a t a concentration of 750 PM. The beads were washed and SDS-PAGE samples prepared as described in Fig. 1. Proteins were resolved by SDS-PAGE and visualized by silver staining.

RESULTS AND DISCUSSION

To investigate the role of these SH3 domains in phagocytic cells, we investigated whether proteins from such cells could specifically bind to SH3 domains. Affinity matrices were prepared by immobilizing various GST-SH3 domain fusion proteins on glutathione-agarose beads. These matrices were used to isolate proteins from detergent extracts of HL60 cells that had been differentiated down the neutrophilic pathway. Sev- eral leucocyte proteins bound specifically to matrices consisting of immobilized GST-SH3 fusion protein, but not to immobilized GST or glutathione-agarose alone (Fig. 1). In particular, a 68- kDa polypeptide bound to the SH3 domains of c-src and phospholipase Cy, and the NH,-terminal SH3 domain of ~ 4 7 " ~ ' ~ bound several proteins that are currently under investigation. In addition, a polypeptide of about 47 kDa specifically bound to the COOH-terminal SH3 domain of p67"*" (p67"'W-SH3) but not to any of the other SH3 domains tested (Fig. 1, asterisk). This polypeptide was isolated from 5 x lo9 cells by binding to the p67phoxC-SH3 domain as described and identified by amino acid sequencing of tryptic peptides. Sequences derived from three different peptides from the 47-kDa protein showed 100%

13754 Function of NADPH Oxidase SH3 Domain TABLE I

Summary ofthe ability of different proline-rich peptides to inhibit binding ofp47 to p67C-SH3 domain Binding studies were carried out as described for Fig. 4.

Peptide no. Protein Sequence (residue no.) ~47:~67C-SH3

Inhibition of

binding

P1 Dvnamin PAVPPAFtPRGSGPAPGPPPAG (786-806) P2 P3 P4 P5 P6

p47~hox KPQPAVPPRPSADL (360-373) p67phm APLQPQAAEPPPRPKTPE (218-235) Cytochrome b,,a KQPPSNPPPRPPAEA (148-162) PTPase 1B DLEPPPEHIPPPPRPPKR (298-316) Vinculin APPKPPLPEGEVPPPRPPPPE (858-878)

A

RG. 5. Characterization of the p67PhmC-SHS binding site on p4'7Phm. A, the effect of truncations at either the NH, or COOH terminus of peptide P2 on the ability to inhibit ~ 4 7 ~ ~ " binding to ~ 6 7 ~ ~ " C - SH3. P2a and P2b are different batches of peptide P2. P2-N indicates NH,-terminal truncations, and P2-C indicates COOH-terminal truncations as listed in Table 11. B, effect of double truncations or amino acid substitutions on the ability of peptide P2 to inhibit the binding of p47ph" to p67Ph"C-SH3 domain. The full peptide sequences and inhibitory activities are listed in Table 11. Binding of p47Ph0r to immobilized ~67~"C-

protein was incubated with the p67phoxC-SH3 domain affinity resin in SH3 domain was assayed as described in Figs. 1 and 4. Solubilized

the presence of the proline-rich peptides a t a concentration of 400 p. Bound proteins were resolved by SDS-PAGE and visualized by silver staining.

identity with the published amino acid sequence of p47ph" (Fig. 2). In addition, the 47-kDa polypeptide was recognized by antibodies raised to p47ph" (a gift from A. W. Segal) and Western blotting confirmed that p47phoX did not associate with any other SH3 domain tested (Fig. 3). It should be noted that a 46-kDa polypeptide that bound to the p47PhO"N-SH3 domain (Fig. 1) did not cross-react with anti-p47Phox antibodies. These data demon- strate a specific interaction between p47Pbx and the COOH- terminal SH3 domain of p67PhoX.

Previous reports have suggested that proline-rich sequences are important for the recognition of SH3 domains (4-10, 19). Analysis of the p47Ph" sequence revealed the presence of a proline-rich motif near the COOH terminus of the protein. "0 investigate the role of this site in the binding to the p67ph"C- SH3 domain, a synthetic peptide corresponding to the proline-

TABLE I1 Summary ofthe ability of sequentially truncated peptides to inhibit

binding ofp47 to p67C-SH3 domain Binding studies were carried out as described for Fig. 5.

Peptide Code Inhibitory activity

KPQPAVPPRPSADL P2 + PQPAVPPRPSADL P2-N1 + QPAVPPRPSADL P2-N2 + PAVPPRPSADL P2-N3 AVPPRPSADL

+/- P2-N4 -

VPPRPSADL P2-N5 - PPRPSADL P2-N6 - PRPSADL P2-N7 - RPSADL P2-N8 -

KPQPAVPPRPSAD P2-c1 + KPQPAVPPRPSA P2-C2 + KPQPAVPPRPS P2-C3 +

QPAVPPRPS P2-N2C3 QPAVPPRP P2-N2C4

+ +

QPAVPPR P2-N2C5 - QPAVPPKPS P2-N2C3K -

rich region of p47ph" was used in competition experiments. This peptide, with the sequence KPQPAVPPRPSADL (peptide P2) inhibited the binding of p47phoX to the p67phoxC-SH3 domain at a concentration of 750 w (Fig. 4). Peptides corresponding to proline-rich sequences in p67Ph", the a subunit of cytochrome b,,,, PTPase lB, vinculin, and dynamin all failed to compete at the same concentration (Fig. 4, Table I), although of these proteins only dynamin has been shown to bind SH3 domains in vitro (5). Thus the interaction between the proline-rich sequence of p47ph" and p67PhmC-SH3 appears to be highly specific. From dose-response studies, the half-maximal concentration for inhibition of binding was calculated to be approxi- mately 100 w (data not shown).

Aseries of truncated variants of the synthetic peptide P2 was tested to further characterize the putative SH3 domain binding site (Fig. 5, A and B; Table 11). Truncations at either the NH, or COOH terminus led to the identification of a minimal active core sequence of QPAVPPRPS (Fig. 5A). This was refined using peptides truncated a t both the NH, and COOH terminus to a minimal inhibitory sequence of QPAVPPRP (Fig. 5B 1. The peptide sequences and activities are summarized in Table 11. In addition, a peptide QPAVPPKP, in which arginine was substi- tuted for lysine, had no inhibitory activity. This indicates that the arginine is supplying more than just a charge for activity and that requirements other than proline sequence or spacing are essential for SH3 domain binding. The SH3 binding motif of 3BP-1 has been previously identified as APTMPPPLPP, with the proline residues at positions 2, 7, and 10 appearing to be crucial for binding to the cabl SH3 domain. Similarly, the corresponding sequence of a second SH3 domain-binding protein, 3BP2, was identified as PPAYPPPPVP and the consensus sequence XPZMPPPZXP (where X is any amino acid and Z is a hydrophobic amino acid) was proposed (19). However, recently the p85a subunit of phosphatidylinositol 3"kinase has been

Function of NADPH Oxidase SH3 Domain 13755

shown to associate with SH3 domain of p59ry" and pp60v-src, and two inhibitory peptides NERQPAPLPPK and KPRPPRPLPVA have been described (11, 12) that do not fit this consensus. Furthermore, sequences that mediate binding of dynamin to several SH3 domains have been identified as containing PAVP- PARP (5). The binding site for p67PhoxC-SH3, although proline- rich, also does not match the consensus sequence, and it is likely that as more SH3 domain-binding proteins are charac- terized the consensus binding motif will be specific to indi- vidual SH3 domains.

Both p47ph" and p67ph" translocate from the cytosol to the plasma membrane after stimulation, and there is evidence to suggest that within the cytosol these components may exist as a 240-kDa complex (20). At present we do not know whether the association between p47phoX and the COOH-terminal SH3 domain of p67phoX occurs within this cytosolic complex or mediates protein-protein interactions following oxidase activation. It is possible that the interactions of other proteins within the NADPH oxidase enzyme complex may be mediated by SH3 domains; for example, both p67phoX and the cytochrome b,,, a subunit contain putative SH3-binding motifs, and another component of the oxidase, p40ph"", has recently been cloned and found to contain an SH3 domain (21). The regulation of SH3-mediated interactions is at present unknown. However, distinct phosphorylation events occur both prior to translocation of the Phox components and after association with the cytochrome (22,23). It is possible that phosphorylation of one or more of the oxidase components may alter protein structure to expose either SH3 domains or binding sites on the components of the oxidase.

The present study demonstrates that ~ 4 7 ~ ~ ' ~ and p67ph"x in- teract with each other via an SH3 domain and a unique proline- rich sequence. Whether this contact forms the sole interaction between the proteins remains to be determined. Elucidation of the interactions that occur in the NADPH oxidase will lead to a greater understanding of the regulation of phagocyte re-

sponses to inflammatory stimuli and should allow a more ra- tional approach to the development of inhibitors of tissue-dam- aging free radicals release which occurs in a variety of diseases such as ischemia and arthritis.

REFERENCES 1. Segal, A. W., and Abo, A. (1993) Dends Biochem. Sci. 18 ,4347 2. Mayer, B. J., and Baltimore, D. (1993) Dends Cell Biol. 3, 8-13 3. Pawson, T., and Schlessinger, J . (1993) Cum Biol. 3, 434442 4. Cicchetti, P., Mayer, B. J., Thiel, G., and Baltimore, D. (1992) Science 257,

803-807 5. Gout, I., Dhand, R., Hiles, I. D., Fry, M. J., Panayatou, G., Das, P., Truong, O.,

Totty, N. F., Hsuan, J., Booker, G. W., Campbell, I. D., and Waterfield, M. D. (1993) Cell 75, 25-36

13.5225-5232 6. Lui, x., Marengere, L. E. M., Koch, C. A., and Pawson, T. (1993) Mol. Cell. Biol.

7. Prasad, K. S., Janssen, O., Kapeller, R., Raab, M., Cantley, L. C., and Rudd, C.

8. Buday, L., and Downward, J. (1993) Cell 73,611-620 9. Egan, S. E., Giddings, B. W., Brooks, M. W., Buday, L., Sizeland, A. M., and

10. Li, N., Batzer, A,, Daly, R., Yajnik, V., Skolnik, E., Chardin, P., Bar-Sagi, D.,

11. Olivier, J. P., Raabe, T., Henkemeyer, M., Dickson, B., Mbamalu, G., Margolis,

12. Rozakis-Adcock, M., Fernley, R., Wade, J., Pawson, T., and Bowtell, D. (1993)

13. Weng, Z., Taylor, J . A., Turner, C. E., Brugge, J . S., and Seidel-Dugan, C. (1993)

14. Bar-Sagi, D., Rotin, D., Batzer, A., Mandiyan, V., and Schlessinger, J . (1993)

15. Clark, R. A,, Volpp, B. D., Leidal, K. G., and Nauseef, W. M. (1990) J. Clin.

16. Abo, A., Pick, E., Hall, A., Totty, N., Teahan, C. G., and Segal, A. W. (1991)

17. Smith, D. B., and Johnson, K. S. (1988) Gene (Amst.) 67, 31-40 18. Totty, N. F., Waterfield, M. D., and Hsuan, J. J. (1992) Protein Sci. 1, 1215-

19. Ren, R., Mayer, B. J., Cicchetti, P., and Baltimore, D. (1993) Science 269,

20. Park, J.-W., Ma, M., Ruedi, J. M., Smith, R. M., and Babior, B. M. (1992) J.

E. (1993) Proc. Natl. Acad. Sci. U. S. A. 90, 73667370

Weinberg, R. A. (1993) Nature 363, 45-51

Margolis, B., and Schlessinger, J. (1993) Nature 363,85-88

B., Schlessinger, J., Hafen, E., and Pawson, T. (1993) Cell 73, 179-191

Nature 363, 83-85

J. Biol. Chem. 268, 1495614963

Cell 74, 83-91

Invest. 85, 714-721

Nature 353, 668470

1224

1157-1161

21. Wientjes, F. B., Hsuan, J., Totty, N. F., and Segal, A. W. (1993) Biohem. J. 296, Biol. Chem. 267,17327-17332

557-561 22. Rotrosen, D., and Leto, T. L. (1990) J. Biol. Chem. 265, 19910-19915 23. Dusi, S., and Rossi, F. (1993) Biochem. J. 296, 367-371

an sh3 domain and proline-rich sequence mediate an interaction

Documents