Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

PI 3‐kinase: structural and functional analysis of intersubunit interactions.

PI 3‐kinase: structural and functional analysis of intersubunit interactions. EMBO Journal vol.13 no.3 pp.511 -521, 1994 The Pi 3-kinase: structural and functional analysis of intersubunit interactions 3, derivatives to produce PI 3-phosphate Ritu Dhand', Kenta Hara2, Ian Hiles1 phosphorylated [PI(3,4)P2] and PI 3,4,5-tris- [PI(3)P], PI 3,4-bisphosphate Ben Bax4, Ivan Gout', George Panayotou', [PI(3,4,5)P3]. Rapid increases in the levels of phosphate Michael J.Fryl 5, Kazuyoshi Yonezawa2, and PI(3,4,5)P3 have been observed when PI(3,4)P2 Masato Kasuga2 and Michael D.Waterfieldl"6,7 cells are stimulated with peptide growth factors quiescent for Cancer Research, 91 Riding House Street, ILudwig Institute such as platelet-derived growth factor (PDGF) and epidermal 8BT, 6Department of Biochemistry and Molecular London, WIP growth factor (EGF) (reviewed in Parker and Waterfield, University College, Gower Street, London, WC1E Biology, 6BT, or transformation by polyoma virus (Ling et al., 1992) 2The Second of Internal Medicine, Kobe University School Department 1992). The 3-phosphorylated lipids appear not to be of Medicine, 7-5-1 Kusunoki-cho, Chuo-ku, Kobe 650, Japan and 4ICRF Unit of Structural Molecular Biology, Department of substrates for known isoforms of phospholipase C and thus Birkbeck College, Malet Street, London, Crystallography, themselves play a direct second messenger role. may WC1E 7HX, UK Determination of the primary structure of the PI 3-kinase address: The Wellcome Research Laboratories, Beckenham 3Present fundamental towards understanding its role in signal has been BR3 3BS, UK Purified PI 3-kinase from brain or transducing processes. 5Present address: Institute of Cancer Research, Cell Biology and liver is a heterodimer of 85 and 110 kDa subunits (Carpenter Experimental Pathology, Haddow Laboratories, 15 Cotswold Road, et Morgan et al., 1990; Shibasaki et al., 1991; al., 1990; Sutton, Surrey SM2 5NG, UK Fry et al., 1992). Through cDNA cloning two forms of p85, 7Corresponding author termed p85a and p85,B, which lack PI 3-kinase activity have Communicated by M.D.Waterfield been described (Otsu et al., 1991). Analysis of the primary of p85 reveals a protein with a multidomain sequence Phosphatidylinositol (PI) 3-kinase has an 85 kDa subunit structure which contains a number of non-catalytic domains (p85a) which mediates its association with activated first described in studies of transforming proteins of the src a Src homology region 3 protein tyrosine kinase receptors through SH2 domains, family. At the N-terminus lies and an 110 kDa subunit (p110) which has intrinsic domain (SH3) (reviewed in Pawson and Schlessinger, 1993) sequence catalytic activity. Here p85a and a related protein p853 and next to this is a region with significant region gene are shown to form stable complexes with recombinant similarity to the product of the breakpoint cluster pllO in vivo and in vitro. Using a panel of glutathione BCR (Otsu et al., 1991). Since the BCR protein has a GAP interact with small fusion proteins of the inter-SH2 activity for rac, this region of p85 may S-transferase (GST) in The C- region of p85, 104 amino acids were found to bind GTP binding proteins (reviewed Fry, 1992). two SH2 the pllO protein, while deletion mutants within terminal half of the molecule is dominated by directly which flank this further defined the binding site to a sequence domains in Pawson and Gish, 1992) region (reviewed adopt a helical of 35 amino acids. Transient expression of the mutant a region that we have predicted may The of these in mouse L cells showed it was unable to conformation (Panayotou et al., 1992). presence p85ct protein domains that the p85 proteins may bind PI 3-kinase activity in vivo. Mapping of the distinct functional suggests and regulatory roles. site of interaction on the pllO protein have multiple interactive complementary of p85 has facilitated both structural defined 88 amino acids in the N-terminal region of pllO The modular structure studies on this and has led to rapid which mediate the binding of this subunit to either the and functional protein in our of its role in the PI 3-kinase or the The inter-SH2 region of p85 advances understanding p85a p853 proteins. little is known as is to be an folded module of a complex, however, by comparison, very predicted independently 110 a single form of pl 10 has coiled-coil of two long anti-parallel c-helices. The yet about the p protein. Only and studies have clearly structure of p85 suggests a basis for the been cloned to date expression predicted can alone encode a with a PI 3-kinase intersubunit interaction and the relevance of this shown that this protein to the of the PI et 1992). The cloned pl 10 is homologous interaction with respect regulation activity (Hiles al., to a PI 3-kinase which is involved in vacuolar 3-kinase complex is discussed. Vps34p, yeast Schu et structure/SH2 domains and Emr, 1990; al., 1993). Key words: PI 3-kinase/protein protein sorting (Herman reveals redundant kinase of the p1O sequence Analysis with and the more motifs and comparisons Vps34p, sequence another PI 3-kinase cloned Tor2, putative yeast recently Introduction for (Kunz et al., homologue required progression GI a domain to be assigned to the 3-kinase play a key role in 1993), has allowed catalytic Phosphatidylinositol (PI) may et of the al., 1992). linked directly or indirectly to C-terminal region protein (Hiles intracellular signal processes of 10 reflects in some cells be an of the N-terminal p1 and may important Sequence analysis region diverse receptor types to known and this lead to et no any proteins in the events which mitogenesis (Fantl al., significant homology mediator a role. 3-kinase region may play regulatory and PI phosphoryl- 1992; Valius Kazlauskas, 1993). a series of such of the inositol of PI and its PI 3-kinase may be regulated by events, ates the D3 ring position Press Oxford University R.Dhand et al. with proteins, little is known of the interaction between the p85 as translocation of PI 3-kinase to cell membranes In consequent access to substrate, and the subsequent and p1 10 subunits. this study we have used both the p85ct association with, and phosphorylation by, protein tyrosine and the p853 proteins in complementary studies to define in with kinases (PTKs) (reviewed Panayotou and Waterfield, the regions responsible for stable subunit association 1992). The binding of PI 3-kinase through its SH2 domains vivo the p110 protein both in and in vitro. We predict that is able to phosphotyrosine-containing sequences on PTKs the structure of the region of p85 responsible for the of of of to bring about independently some degree activation interaction with pl 10 is an independently folded module ct-helices. The of the Tyrosine phosphorylated peptides corresponding a coiled-coil of two possible significance enzyme. to potential PI 3-kinase sites on the insulin receptor substrate, the intersubunit interaction with respect to activation of the have been shown to IRS-1, or the intact IRS-1 protein, enzyme is discussed. et activate PI 3-kinase in vitro (Backer al., 1992; Giorgetti Results of a et al., 1993). Similarly, addition phosphotyrosine- to Y751 of the PDGF3 p110 can reconstitute with p85a or to form containing peptide corresponding p85,3 also causes activation active Pi 3-kinase complexes both in vivo and in vitro receptor, a known in vivo binding site, et it The p85a subunit is known to form a stable complex with of PI 3-kinase in vitro (Carpenter al., 1993). Further, this Y75 1 the p1O subunit when they are coexpressed in Sf9 cells has been shown that upon binding phosphopeptide, and its N-terminal SH2 (Hiles et al., 1992). To investigate whether p85(3 will also the p85 subunit of the PI 3-kinase conformational reconstitute with the p110 protein in an active complex, domain in particular undergo a change et This induced insect cells were coinfected with either p85ac and pl10 or (Panayotou et al., 1992; Shoelson al., 1993). of be transmitted p853 and p1 10 recombinant baculoviruses and PI 3-kinase change in conformation the p85 protein may subunit and thus contribute to its activity was measured. In double infections of baculoviruses to the associated p1O of the nature of the interaction expressing p85a and p 10, both proteins were detected in activation. An understanding subunit and its domain either anti-p85a or anti-p 1 10 immunoprecipitates, or when between the regulatory p85 catalytic on the mechanism of activation of the enzyme was bound to Y751 phosphopeptide affinity would shed more light 5 and All and immune there is a wealth of biochemical and beads (Figure lAi, lanes 4, 6). affinity PI 3-kinase. Although on the association of PI 3-kinase with other precipitates were found to possess PI 3-kinase activity structural data l 1iii ,. 1) /Ifi I, V III /S I II \Pi% __ h Is .2 s>4 %~ ).i . i, \l\\\ \I *kJI 1NM R1D£l k1I),i ... i- SO 1) (I 10 - tei "( (I ...i pi --I *1)11 1 *1) 1I1 q11p"~4P - 4- 5%4 p1S *- p's Bii I(ii 5 ) III 1i) l1 r.) IX5l /1 4 5 ii 2 x I'll, Fig. 1. Both p85ca and p85(3 are able to for-m stable active complexes when coinfected with pl10 in insect cells both in vivo and in vitro. (Ai) Sf9 cells were infected with either p85ca (lane 1), p85(3 (lane 2) or pllO (lane 3); viruses were coinfected with p85a and viruses (lanes 4-6), or pllO with p85,B and plIO viruses (lanes 7-9). Lysates of these cells were precipitated using polyclonal affinity purified antibodies raised against p85a (lanes 1 and 4), monoclonal antibodies raised against p85(3 (lanes 2 and 7), polyclonal affinity purified antibodies raised against 10 (lanes 3, 5 and pl 8) or bound to immobilized Y751 phosphopeptide affinity beads (lanes 6 and 9). Samples were analysed by Coomasssie blue staining of SDS-PAGE gels. (Aii) PI 3-kinase assays were performed on immunoprecipitates infected and treated as described in Ai. (Bi) and p85(3 were p85ai immunoprecipitated from insect cells and then incubated with lysates of Sf9 cells that had been infected with viruses (lanes 1 and 2). pllO Complexed proteins were resolved on SDS-PAGE and identified by staining with Coomassie blue. (Bii) 3-kinase assays were performed on PI immunoprecipitates infected and treated as described in (Bi) Lysates of insect cells that had been infected with viruses were (Biii) plIO immunoprecipitated and then either left untreated (lane 1) or incubated with p85a-containing Sf9 cells (lane 2). Samples were washed in lysis buffer, resolved on SDS-PAGE and identified by Coomassie blue staining. 512 Pi 3-kinase intersubunit interactions (Figure lAii, lanes 4, 5 and 6). Similar results were seen To study the interaction between the inter-SH2 region of with viruses expressing the fl-form of p85 and p 110. Infected p85a with pl10 in the context of the entire p85 protein cells immunoprecipitated with antibodies directed to the monoclonal antibodies which have epitopes within the inter- p850 if or the plO subunit, or bound to Y751 phosphopeptide SH2 region were used to block intersubunit associations. affinity beads, were found to contain both and p110 p85a from insect cells was immunoprecipitated, either using p850 proteins when analysed by SDS -PAGE (Figure lAi, lanes two monoclonal antibodies which bound located epitopes in 8 and 7, 9). Examination of associated PI 3-kinase activity the inter-SH2 region of p85a, or with two control antibodies in parallel showed the 10 complex to be that recognize the BCR domain (End et al., 1993). These p850-pl enzymatically active (Figure lAii, lanes 7, 8 and 9). The immunocomplexes were washed stringently and incubated levels of Coomassie blue staining protein precipitated in the with lysates of insect cells that had been infected with pl 10 p85,B-pl 10 complex are essentially the same as those seen virus. Bound proteins were then subjected to analysis by precipitated in the complex using p85oa-plO identical SDS -PAGE and PI 3-kinase assays. Similar quantities of amounts of both plO antibody or Y751 phosphopeptide p85ca were found to be immunoprecipitated by all four affinity beads. Since both p85a and p85fl are expressed to antibodies, as determined by staining with Coomassie blue approximately the same levels in insect cells (Gout et (Figure 3A, lanes 1-4). However, the amount of PI al., 1992), this would suggest that 10 complex 3-kinase activity which bound to p85ai immunoprecipitated p850-pl formation in Sf9 cells is as efficient as p85x-p1 10 complex with the two antibodies that recognize the inter-SH2 region formation. Indeed, as comparable amounts of PI 3-kinase was significantly less (Figure 3B, lanes 3 and 4), than the activity were precipitated with either the p85ac-p110 activity which bound to p85cx immunoprecipitated with complex or the 10 complex this suggests that p1 10 antibodies to the BCR domain (Figure 3B, lanes 1 and 2). p850l--pl is able to form a stable, active PI 3-kinase complex with either of the p85 isoforms in insect cells. Mapping the precise site in the inter-SH2 region To ascertain whether the two subunits of the PI 3-kinase required for the interaction with p1 10 complex were able to associate post-translationally in vitro, To define which component of the inter-SH2 region is one subunit of the PI 3-kinase complex was first immobilized involved in directly binding the pl 10 subunit, deletions were onto an appropriate matrix and then allowed to bind the other subunit in solution. First, the p85 proteins immunoprecipi- tated from infected insect cells using isoform-specific 110 78 274 108 329 41640 603623 706 724 monoclonal antibodies were collected on protein A- 1. GST-p85a SH3 , BCR Sepharose beads. These immunocomplexes were 1 | NSH21 InIer-SH2 then incu- | ,SH2I | bated with lysates of insect cells that had been infected with 2. GST-CSH2 pl 10 viruses, after which they were washed and subjected 1i[ to either analysis by SDS-PAGE or PI 3-kinase assays. 3. GST-InterSH2 Inter-SH2 Alternatively, p1 10 was immunoprecipitated from infected 329 433 insect cells, incubated with p85a-expressing lysates 4. GST-NSH2 and then treated as described above. Both p85a and p85,B were seen 113 279 S. GST-BCR to bind p1 10 by Coomassie blue - staining following SDS PAGE (Figure lBi, lanes 1 and 2) or upon assay for 84EI 6. GST-SH3 associated PI 3-kinase activity (Figure lBii, lanes 1 and 2). Similar results were obtained when p 110 was immobilized and the bound p85a protein was detected as a Coomassie blue-stained protein on SDS-PAGE following binding * P13K activity in vitro (Figure lBiii, lane 2). CPM bound (cpm) The inter-SH2 region of p85a is required for binding to p110 The specific region of p85a responsible for binding to p1 10 was defined initially by analysis of the domains of p85 which may be involved in the interaction. Glutathione S-transferase (GST) fusion proteins of the various subdomains of p85a were immobilized on glutathione-Sepharose beads (Figure 2A) and used for affinity analysis with cell lysates prepared from Sf9 cells infected with a recombinant baculovirus expressing pl 10, after which the complexes were alone and washed and subjected to PI 3-kinase assays. GST beads GST-p85a bound to glutathione-Sepharose °TI2 3 4o 5 6 7 were used as controls. incubated with wild-type lysate GST-Fusion protein associated Figure 2B shows that PI 3-kinase activity only with either the or full-length GST-p85a protein (lane 1) 2. The inter-SH2 of is for Fig. region p85a required binding p1l0. with the of the PI 3-kinase inter-SH2 region p85ca (lane 3). domains of and GST (A) Full-length GST-p85a (1) p85a (2-6). (B) did not bind to GST alone (lane and no activity 7) A and GST alone were Domains 2-6 of as described in (lane p85a 7) insect cell 3-kinase bound to full- endogenous PI activity and then incubated with bound to beads glutathione-Sepharose lysates when incubated with control infected with virus. After length GST-p85a protein of insect cells that had been plIO washing were to 3-kinase with lysis buffer, proteins subjected PI assays. not lysates (data shown). 513 R.Dhand al. at ,A ~ ~~ *X p8X5Lt i)>S t. .}j_ ..... "' 5--XFi, -- il -X 12,i 24 32) i2 4411 (03 f 23 0 '24 'k -.:. j__________ i 2 I .) 4~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ W C, [S lA K12 [ |.} | lilter-SH-1 12| 425l; l [ : 1113K activits (1 bound11( (I0000 6H5 8(X(O() it 6000-) l Pl.3Ka( uti%it I1111 ,4Nl 4000(1 I - 2 3 4 I 111r t'- SF12- 15 41'k'll ) _ _ .2 ))1 0( Iiter- S112-9 l ' 4 3,:1 in of Fig. 3. Antibodies with epitopes the inter-SH2 region p85a _2 4 ti inhibit the binding of the pllO protein. was (A) p85a 1. .. %-I 1 oltssbl 1 prolvil cells with monoclonal immunoprecipitated from infected insect in and antibodies which have epitopes either the BCR domain (lanes 1 or in (lanes 3 and and 2) with epitopes the inter-SH2 region 4) Fig. 4. Identification of the region of the inter-SH2 that with region directly incubated pl1O-containing lysates. Immunoprecipitated p85 binds plO. (A) Full-length GST-p85a (1) and varying sizes of the visualized Coomassie blue of SDS-PAGE proteins were by staining inter-SH2 region (2-6). (B) The different sized GST-inter-SH2 3-kinase were on gels. (B) PI assays performed inimunoprecipitates regions were bound to glutathione-Sepharose beads and incubated infected and treated as described in A. with plO-containing Sf9 cell lysates (lanes 1-6). Samples were washed and assayed for associated PI 3-kinase activity. introduced at both ends of the inter-SH2 of region p85a, reducing the size of this from 175 amino acids region (425 -600) down to 104 residues (451-555) (Figure However, mutants with deletions which advanced further 4A). into the as those in The truncated inter-SH2 regions expressed as GST fusion inter-SH2 region GST -p85,B(486-724) - which in proteins were bound to glutathione Sepharose beads and GST -p85((516-724) were unable to bind PI 3-kinase lanes 3 and These were then incubated with lysates of insect cells that had been activity (Figure SB, 4). results suggest infected with 10 and was that 71 amino acids in p1 associated PI 3-kinase (residues 445-516) were activity p85( measured. The results in 4B (lanes show that involved in the interaction between the Figure 2-6) p85,B protein and the deletion of amino acids from either end of the inter-SH2 0. In order to locate the site in p1 binding p85( more region gradually reduced the amount of bound PI 3-kinase precisely, mutants that contained deletions in this such region This be if the inter-SH2 as activity. may explained disrupting GST-p85(3A486-516, GST-p85,3A445-485 and structure is a domain and thus were destabilizing putative prevents GST-p85(3A445-469 constructed (Figure SA, mutants protein -protein interactions. In addition, these results 5, 6 and 7). The function of these mutants was then assayed identified a structural element of 104 amino acids, between as described above. Following binding, PI 3-kinase assays residues 451-555 of the inter-SH2 domain of p85a, which revealed that both and GST-p85OA486-516 is able to bind the directly p1O protein and associated PI GST-p85j3A445-469 were still able to bind -65% of the 3-kinase activity (Figure 4B, lane 6). PI 3-kinase activity (Figure SB, lanes 5 and 7), in A series of nested deletions of fusion proteins with comparison wild-type GST-p85(3. However, the GST-p850 from the 3' end of the N-terminal SH2 domain to the 5' end deletion mutant (analogous to residues GST-p85OA445-485 of the C-terminal SH2 domain were then constructed to 452 -492 in p85cx) was unable to bind any detectable PI refine the binding site within the inter-SH2 region 3-kinase activity (Figure SB, lane 6), thereby defining 40 (Figure SA). The interaction of these GST fusion proteins amino acid residues that are absolutely required for the with p1 10 was then measured as described above. Full-length binding of PI 3-kinase activity to p85(. GST-p853(1-724) and GST-p85((446-724), which To assess whether a similar region of p85a was involved contains most of the inter-SH2 region, were clearly able to in binding to the p1 10 protein, an analogous deletion to that bind PI 3-kinase activity (Figure 5B, lanes 1 and 2). made in was p85( constructed (Figure 5A, mutant 6) and 514 interactions Pi 3-kinase intersubunit 73 254 329 41644 463423 706724 110 129 (A) (B) I nca Ns1121: 14m 1.GSTpUp (1724) p85I~ lismi |H2 445 724 l wSH2, GSTpiP (445.724) 406 724 516 724 r GSTpLS( (516-724) 1 436-516 724 s G 5.GSTpt5L4516 |I I rs I S.z |1|1 II WR GSTpLVM4S485 ISM I I I -fq I .Mul 16. 0 0 I 724 I I vu KA W | 1 7. w CI HI G6Fp3SA4.6 I|| I I kH| w m 110 78 129 254 329 41644 43623 786724 G G12 G 12 NMG p85a IISH3 |a ' 1w lH2 LG8STp (1-724) 329 41644 463623 78 724 l-S2 .a||| 9. Fig. 6. Semi-confluent mouse L cells were transfected with either GSTpSSaNC H=40 SRa-Wp85cz or SRa-p85aA478-513 plasmid. Transiently expressed 329 416 478.13 623 786 724 with an Wp85a (W) or p85aA478-513 (M) were immunoprecipitated z 10. GSTp3SaNCA47S513 |NLSFj2 I anti-p8Ma monoclonal antibody (G12) or control mouse IgG (NMG). The immunoprecipitates were either subjected to Western blot analysis using a polyclonal anti peptide rabbit antiserum (A), or samples were * % P13K subjected to PI 3-kinase assays (B). (-) indicates untransfected cells. aqvk boud G12, which is unable to recognize murine p85a (Yonezawa et al., 1992) and then either visualized by Western blot C& the analysis using a polyclonal antibody which recognizes jI46 C-terminal of or assayed for associated PI 3-kinase p85cx Figure 6A show that equivalent activity. The results in wild-type and mutant p85CY were amounts of the from these cells (lanes 2 and 3) and no immunoprecipitated I.29 mouse p85a is immunoprecipitated from endogenous untransfected cells (lane 1). However, PI 3-kinase activity 1 2 3 4 5 6 7 3 9 10 is found to be associated with the wild-type p85a GST-Ih.I protel. only 3) and no PI 3-kinase activity is observed (Figure 6B, lane in of the mutant p85CY (Figure 6B, lane 5. Amino acid 478-513 of p85a and 445-485 of p85,B immunoprecipitates Fig. residues are for of (A) Full-length p85(3 (1), 5' necessary the binding pllO. or from untransfected cells 6B, lane 1). 4) (Figure nested deletions of p85(3 (2-4), deletion mutants of p85,B (5-7), full- p85aNC (9) and deletion mutant of p85aNC (10). length p85a (8), for binding on p110 Mapping the site p85 GST fusion proteins of and p85(3 were bound to (B) p85a To map the regions of p 110 which could interact with p85a beads and then incubated with plO-containing glutathione-Sepharose cell (lanes were then washed and and p853, GST-pl10 fusion proteins encompassing various Sf9 lysates 1-10). Samples to PI 3-kinase subjected assays. regions of p110 were made (Figure 7A) and their ability to bind p85 expressed in an insect cell expression system for its to bind PI 3-kinase activity as described assessed. Results in Figure 7 show that of the six fusion assayed ability 0. to bind to proteins piIO. I-pi 10.6, only p1 was able above. This deletion mutant, GST-p85cxNCA478-513 with No 471-508 in was p85t (lane 16) and p853 (lane 17) high affinity. (analogous to amino acid residues p85,B), in the control Sf9 cell and no lane binding was detected lysate unable to bind PI 3-kinase activity (Figure SB, 10) seen in made with which binding to either subunit was experiments compared with a similar construct, GST-p85aNC, This indicates that residues lane These results define any of the other constructs. lacks the deletion (Figure SB, 9). 10 site for the p85 proteins. in that is to mediate 1-128 of p1 contain a binding a 35 amino acid region p85a necessary for to subunit. To investigate further the sequences required binding binding of the p1O catalytic which seven GST fusion proteins (p1 10-1.1-pl 10-1.7) p85, the first 128 amino acids of were Association of PI 3-kinase activity with wild-type and encompassed p110 Of pI 10-1.3 and mutant in vivo constructed (see Figure 8A). these, p85a and lanes for 10 10-1.5 bound both 8, 20, of the necessary binding p1 p1 p85at p85( (Figure As the regions p85 proteins None of the other constructs been examined in 21 and 15 and 3-kinase have using 14, respectively). PI activity only to bind either p85 subunit. these and a mutant with showed ability Together vitro the full-length p85a p85a any analyses, acid residues 20-108 of are able to define amino were constructed results the site deleted (p85ctA478-513) binding which is for a minimum structural element in mouse L cells. The recombinant pl 10 as required for transient expression It should be noted that the the subunit. and mutant were binding p85 bovine wild-type p85a proteins specifically 0 to be fusion with an monoclonal GST-pl1 proteins appear very susceptible immunoprecipitated anti-p85a antibody 515 R.Dhand et al. to degradation which explains the presence of numerous The inter-SH2 region of p85 is predicted to be a proteins on SDS -PAGE gels even after affinity binding to of two anti-parallel a-helices. coiled-coil glutathione- Sepharose beads. that the inter-SH2 domain mediated inter- The demonstration led us to evaluate further the predicted subunit association of this region of the p85 protein. Previously we structure that this region would be largely a-helical in A had predicted et al., 1992) and suggested that it might character (Panayotou bundle. To extend this work, the sequence be a four-helix region of p85a was used to search the amino of the inter-SH2 in the Owl 19.0 database (Proterin acid sequences Leeds University, UK); 12 of the top 20 Engineering Club, from the proteins myosin or hits were with sequences have coiled-coil regions, the top five hits paramyosin which ., of p85. Although the percentage being with isoforms between the aligned amino acid sequences sequence identity low, heptad repeats (see below) were often was relatively The amino acid sequences of proteins found to be aligned. Brookhaven databank of protein structures (Bernstein in the B also searched with the sequence of the inter- et al., 1977) was Here two of the top three hits were with SH2 region. a two stranded oa-helical coiled-coil muscle tropomyosin, P2.{0-3 P11iO-4 P110-6 P110-5 P110-2 P110-l structure has been solved to 15 A resolution protein whose 0- .-- 1-- ---- - 40- I-.I-- . - -. 1979, 1986). Although the amino acid (Philips et al., with was quite low (20% over sequence identity tropomyosin the heptad repeats in the inter-SH2 region 175 amino acids), were in register for a considerable and that of tropomyosin Sequence analysis of tropomyosin part of the alignment. McLachlan and Stewart, 1975; Parry, (Hodges et al., 1972; an unbroken series of heptad repeats 1975) had shown where residues a and d are conserved as (abcdefg)n, residues, which suggested that the structure hydrophobic a coiled-coil of oa-helices. Our earlier sequence would be of the helical domain of p85a and p85,B (Panayotou analysis 7. Identification of a of which binds (A) region p1II0 p85. Fig. showed the presence of two long unbroken et al., 1992) view of the of as GST fusion regions pl110 expressed Diagrammatic of the form (abcdefg)n where residues a and heptad repeats Numbers to the amino acid residues at the proteins. correspond extremes of the of I1 0. (B) The different expressed regions p d are hydrophobic. fusion were bound to glutathione-Sepharose proteins GST-plI10 in the inter-SH2 region of p85 and the The heptad repeats beads and incubated with 519 cell p85j3 or lysate containing p85a, of its amino acid sequence with sequences of similarity control were washed and bound (CON) wild-type proteins. Samples such as tropomyosin, paramyosin and coiled-coil proteins resolved SDS-PAGE on a 10% gel. The arrow indicates proteins by that the domain might consist of a coiled- the of the myosin suggested position p85 proteins. 7 1 P110-1 4 P 1 10-1 3 P11- 1 .2 P 101- .6 fP1C0-1 ; 1- PI 1v ., '1r-1 Bsm 128 -,- .- C 12 X - , X c , i.X, A,,, ,, P112-1 l - 49 pi10-11 - 97 1 . RI pulO-i -66 1 1 pi10 3 .. _ _ a ;L _. ... 43 pulO-l -40 plO- 1.5 1 3 1 6 __ p110 .. .... 37 -_ 7 31 p110 L._ -t.- W ...... Fig. 8. Fine scale mapping of a region of pl10 capable of binding p85. (A) Diagrammatic view of the regions of p1 10-1 expressed as GST fusion proteins. The BsmI site used in the construction of p1 10-1.6 and pI10-1.7 is shown. Numbers correspond to the amino acid residues at the extremes of the expressed regions of pl10. (B) The different GST-pllO-l fusion proteins were bound to glutathione-Sepharose beads and incubated with Sf9 cell lysate containing p85a, p85,3 or control wild-type (CON) proteins. Samples were washed and bound proteins resolved by SDS-PAGE on a 10% gel. The arrow indicates the position of the p85 proteins. 516 interactions Pi 3-kinase intersubunit coil of a-helices. Analysis of the sequence of the inter-SH2 The coiled-coil has two faces, one having residues BEAaeb region showed that its sequence is consistent with the and the other residues we shall call these the AEB CGDdgc; structure being an anti-parallel coiled-coil of two 70-residue face and the DGC face (Figure 9C). In our model the first a-helices (residues 441-512 and 518-588 in p85a; see helix starts at residue 441, a G position, and it may be that Figures 9 and 10). The proposed register of the helices was preceding residues (431-440) pack against the DGC face that which gave the maximum number of possible ion pairs. at the bottom of the coiled-coil. The second helix is predicted A single proline residue (568 in p85a) which occurs on the to end at residue 588, an e position on the second helix, and outside of the second predicted a-helix should be tolerated it may be that residues C-terminal to the second helix (Blundell et al., 1983).T Sequence comparison with the (residues 589-618) pack against the AEB face at the bottom crystal structure of tropomyosin (Philips et al., 1979, 1986) of the coiled-coil. Possibly some of the seven acidic residues the coiled-coil region of p85 is expected between 604 and 615 could interact with the exposed basic suggests anti-parallel to have a length of -110 A and contains three-quarters of residues Arg577 (a), Arg574 (e) and Lys567 (e), some a turn of coiled-coil (see Figure IOA). A large number of 20-30 A from the bottom of the coiled-coil on the BAE charged residues are present at positions e and g on the two face. These results provide a possible structural basis for a-helices, as is commonly found in a-helical coiled-coil consideration of the intersubunit binding and enzyme proteins (Conway and Parry, 1990) and the amino acid regulation. residues found at the seven positions in the heptad repeat are fairly typical of those found in two-stranded coiled-coils Discussion (see Table I and Conway and Parry, 1990). Six ion pairs are predicted to help in stabilizing the coiled-coil region, PI-3 kinase has two distinct subunits: p85, which can serve giving a favourable 0.6 ionic interactions per heptad repeat. to link the enzyme to activated protein tyrosine kinase 431 446 453 460 467 474 481 488 495 502 509 lst SH2 domain GABCDEFGABCDEFGABCDEFGABCDEFGABCDEFGABCDEFGABCDEFGABCDEFGABCDEFGABCDEFGA Sa p8 ..NESLAQYNPKLDVKLLYPVSKYQQDQWKEDNIEAVGKKLHEYNTQFQEKSREYDRLYEDYTRTSQEIQMKRTAIEAFNETIKIFEEQCQTQERYSKEYIEKFKREGN ..HESLAQYNAKLDTRLLYPVSKYQQDQIVKEDSVEAVGAQLKVYHQQYQDKSREYDQLYEEYTRTSQELQMKRTAIEAFNETIKIFEEQGQTQEKCSKEYLERFRREGN p850 518 525 532 539 546 553 560 567 574 581 588 607 620 2nd SH2 efgabcdefgabcdefgabcdefgabcdefgabcdefgabcdefgabcdefgabcdefgabcdefgabcde 85a ETE IQR IMHNYEKLKSR IS E IVDSRRRLEEDLKKQAAEYREI DKRMNS I KPDL IQLRKTRDQYLMWLTQKGVRQKKLNEWLGN -ENTEDQYSLVEDDEDLPHHDEKTWN 5 ILLNSERLKSRIAE DKRMNSLKPDLMQLRK IRDQYLVWLTQKGARQKK INEWLG IKNETEDQYSLMEDEDDLPHHEERTWY p8 EKEMQR IHESRTKLEQELRAQASDNREI 446 453 474 495 502 509 460 467 481 488 F K T R E Q T E E R E B E H D T E K Q K Q Q E G N S Y S R N E E I ,. D V Y K L T K F F y G N K Q E D E A T Q Y K C A E E R R M A I T E F K T R E T E R E Q E 588 581 574 567 560 553 546 539 532 525 518 f m k p k a e d s h t b t d i n r k e q rr s e 9 1 r k d a e k m e r y a 1 r 1 m y 1 i i d k y 1 i i q 1 i 1 i w t d r e d s r n e c q r k q q s e k e f m k p k a e d s h t Ee G 1 N r S k Y d S a N v k E m I e g B e F. Aa I 1 L r F 1 Y m Y I I I r I i C y S i F Dd k V y Y 1 i L i T q 1 F i F 1 i Y IS E N w K t e A s T r n Y Gg Q d E r2 E d Q . I of bovine and are shown in the between the two SH2 domains. The C-terminal of the Fig. 9. (A) The sequences p85ai p85,B aligned region boundary 431 of is defined from the NMR structure of the domain et Two first SH2 domain (residue p85a) (Booker al., 1992). long heptad repeats to two helices residues 441-512 and are indicated above the GABCDEFG... and corresponding the predicted (helix-1, helix-2, 518-588) sequences; in Note that there are some 30 residues between the Positions A/a and are in character these efgabcdefg... D/d predominantly hydrophobic repeats. amino acid of the two helices have been drawn on a helical end of the second helix and the start of the C-terminal SH2-domain. The (B) sequences are in The solid line indicates the line of that for the has been drawn so that the two helices register. net. Note helix-2, sequence backwards, at and of the two helices are to interlock. A of anti- intersection of the two helices. Hydrophobic residues positions A/a D/d expected (C) pair two helices are residues at A/a and D/d are Note the from positions hydrophobic. parallel ca-helices interacting. Heptad repeats anti-parallel shown, register (height) of the residues in the two helices: A B C D E F GA d c bag f ed Residues on the two in that are to interact on how helices interact to form ca-helices 1953 for further details coiled-coils). (D) p85ca expected (see Crick, in while those from helix-2 are in lower case. Residues which could form ion between Residues from helix-i are pairs are shown case, superposed. upper the two helices are underlined. 517 R.Dhand et al. a form of p85 (Otsu et al., 1991; Hiles et al., 1992). The ce and the forms of are ( p85 highly conserved in all but the BCR and both region proteins have been used here in complementary studies to analyse the basis of the p85 -p1 10 i _ R~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ i~~~~~~~~~~~~~ interaction. We show that both p85a and can form p850 stable active complexes with plO. Indeed, PI 3-kinase activity specific to the p85( subunit has been identified in 478 492 441 512 several cell lines (Gout et al., 1992). In addition, T lymphocytes have been shown to contain both p85e and p853 proteins associated with a PI 3-kinase activity, which are differentially regulated during T cell activation (Reif et al., iwl 1993). To examine whether the same specific region of p85a and =40~~~~~4 interacts with we p1 10, expressed domains of p85 as p850 GST fusion proteins. The results clearly show that the inter- SH2 region of p85 is required for the association with p110 r,- and blocking of the inter-SH2 region on the intact p85 protein .1,3. with monoclonal antibodies that have epitopes in this region prevents p110 binding in vitro. These data preclude SH2 4s4 wi domains as binding PI 3-kinase activity in vivo, as shown by Cooper and Kashishian (1993). Indeed, monoclonal antibodies that have epitopes located in the N-terminal SH2 domain of do not inhibit the of 10 to p85a binding pl the p85 subunits (R.Dhand, unpublished observation). A more detailed analysis of the inter-SH2 region of p85a !-W sizes of this domain defined 104 amino using varying acids, residues 451-555, which are able to bind the directly pl 10 CIS protein and PI 3-kinase Deletion mutants of the activity. p85 of Numbers proteins identified a short stretch of - 35 amino acids in Fig. 10. (A) Schematic diagram of the coiled-coil domain both p85. residues are residues according to p85a numbering. The first helix (helix-1; p85 proteins, which includes residues 478 -513 in p85ae and residues 441-512) is depicted with light shading, the second, (helix-2, 445 -485 in p85( for 10 in vitro. necessary binding pl The that the deletion mutants in 518-588) is shaded more darkly. Note p8513 remains that some of the deletion mutants of possibility p85 to 452-476 -469 452 -492 correspond residues (445 p853 numbering), may be bound to an inactive form of 10 and thus are not and 493-523 p1 (445-485 p85,B numbering) (486-516 numbering). p850 The deletion 452-492 abolishes and (B) binding may completely disrupt detected in these low level of assays, however, expression the structure. The deletion 493 -523 will remove the lower of (C) quarter the p85 deletion mutants does not detection of permit the the coiled-coil The deletion 452-476 will remove the half of the (D). top associated p 10 protein by other means. of the Expression coiled-coil. full-length p85cA478-513 in mouse L cells showed that it was also unable to associate with PI 3-kinase activity, in et Otsu et receptors and other proteins (Gout contrast to wild-type p85ao, thus that the deleted al., 1992; al., verifying 1991), and p1 10, which has an intrinsic region of p85ae defined in this study is the binding site for catalytic activity of and (Hiles et al., 1992). The regulation the p 10 protein in vivo. In the of the enzyme activity addition, ability full which be involved the complex protein interactions in length p85a and this deletion mutant p85aA478-513 to bind may to the of the function of PI 3-kinase seem PI 3-kinase activity and the insulin receptor substrate IRS-I require presence both subunits. The most obvious of intersubunit example in vivo has been examined using stable cell lines regulation is the of ability phosphotyrosine-containing overexpressing these constructs and the insulin receptor proteins or to cause activation of the 3-kinase peptides PI (Hara et al., in preparation). Insulin stimulation of cells et et activity (Backer al., 1992; Carpenter al., 1993; expressing wild-type p85a showed that it associated with Giorgetti et al., 1993). The basis for activation could lie in IRS-I and PI 3-kinase activity, moreover an accumulation phosphopeptide-induced conformational changes which have in the levels of PI(3,4)P2 and PI(3,4,5)P3 was observed. been seen in the N-terminal SH2 domain of and the However, insulin of p85a stimulation cells expressing p85aA478- p85a protein itself (Panayotou et al., Shoelson et 513 showed that although p85a was 1992; al., still able to associate the structural 10 1993), although implications of this on p1 with IRS-I it was unable to bind PI 3-kinase activity and remain unknown. Here we seek to determine for the basis significant reduction in the levels of PI(3,4)P2 and the intersubunit interactions so that we can understand the PI(3,4,5)P3 was seen. These data would suggest that these function of this enzyme both in vitro and in vivo. constructs may be used in further studies as dominant- PI 3-kinase may exist in a number of distinct isoforms in negative mutants in the analysis of mitogenic signalling mammalian cells. It is clear that the p85 proteins are encoded pathways. The of mapping the p85 binding site on pl 10 has by a gene family of at least three members, a, and 'y (Otsu shown that amino acids 20-108 are capable of binding to et al., 1991; Volinia et al., 1992), and in situ hybridization the p85 proteins and no other expressed fragments and PCR analysis confirm the presence of five closely related (plO.2-plO.6) show any detectable binding to either sequences of human pl 10 cDNA (S.Violinia, unpublished subunit. Thus the inter-SH2 region of p85 and the N-terminal observation). The only fully characterized PI 3-kinase has region of p1 1O, as independently folded modules, are been from the bovine purified brain, sequenced at the protein responsible for mediating the interaction of the two subunits cloned at DNA level, the level and is found to contain the of PI 3-kinase. However, it cannot be ignored that other 518 interactions PI 3-kinase intersubunit Table I. Percentage occurrence of each of the 20 amino acids at each position in the heptad repeats a(21) b(20) (%) c(20) (%) d(20) (%) e(21) (%) f(20) (%) g(21) (%) Ala(5) 0 0 10 0 4.8 5 4.8 Ile(13) 28.6 5 5 20 4.8 0 0 Leu(9) 19.0 0 0 20 4.8 0 0 Met(4) 4.8 0 5 0 4.8 5 0 Val(2) 0 0 0 5 4.8 0 0 9.5 0 0 10 0 0 0 Phe(4) 19.0 0 0 15 4.8 0 4.8 Tyr(9) 9.5 10 20 0 9.5 10 9.5 Arg(14) 0 15 10 15 9.5 15 9.5 Lys(15) 0 10 0 0 4.8 5 14.3 Asp(7) 0 15 25 0 19 25 19 Glu(21) 0 5 0 0 9.5 0 9.5 Asn(5) Cys(l) 4.8 0 0 0 0 0 0 Gln(12) 0 20 15 10 0 5 9.5 0 0 0 0 9.5 0 0 Gly(2) 0 5 0 0 0 5 0 His(2) 0 0 0 0 0 5 0 Pro(l) 4.8 5 5 0 9.5 5 4.8 Ser(7) 0 10 5 5 0 15 9.5 Thr(9) 0 0 0 0 0 4.8 Trp(l) 0 TOTALS(143) 100 100 100 100 100 100 100 5 20 70 28.8 10 9.5 Apolar(46) 80.9 9.5 25 30 15 19 25 19 Basic(29) 0 25 25 0 23.8 30 33.3 Acidic(28) of Ile, Leu, Met, Val, Phe and Tyr), basic and acidic residues occurring at each of the seven positions in Also shown are the percentages apolar (Ala, shows the two helices in the inter-SH2 domain of p85 have many features in common with two-stranded a-fibrous the heptad repeat. The Table that predicted Note in that at positions a and d of the heptad repeat 75.6% of residues are apolar. Forty-nine proteins (see Tables in Conway and Parry, 1990). particular of the helices are whereas only 12% of residues at positions a and d are charged. percent of the amino acids at the external positions (bcefg) charged, at a and d compared with other coiled-coil proteins (Conway and Parry, 1990) might be The relatively high proportions of Ile, Phe, and Tyr positions are predicted to be anti-parallel, whereas in other coiled-coil proteins helices are parallel. related to the prediction that in p85 the helices and the regions of both the p85 and the p110 proteins may also play a loss of function observed complete (Figure lOB) to some role in the interaction between the subunits but have 10 is seen not to bind p85. Sequence analysis p protein that residues 478-492 of been overlooked in this study due to the inability of a of both the reveals p85 proteins 470-497 domain to fold correctly. are contained in four particular p85a heptad repeats (residues the two at the To search for a structural basis for the region of p85 which in which are identical between proteins p85CY) conservative amino acid binds 10 the structural studies of Panayotou et al. amino acid level for a pl predicted except single to residues were extended. Sequence analysis suggests the inter- A (1992) change (Figure 9A). peptide corresponding and SH2 to be an coiled-coil of two a-helices 470-497 of was region anti-parallel p85az synthesized chemically coupled it was unable to bind to 10 which we have termed helix-i and helix-2 (Figure IOA). to Sepharose beads, however, p1 Evidence the concept that this region may be a and 3-kinase supporting PI data). activity (R.Dhand, unpublished characteristics of the domain is seen various size GST fusion of the discrete using Analysis p1O binding p85 that while helix-i of the inter-SH2 It is demonstrated that as deletion mutants proteins of the inter-SH2 region. suggests for the interaction between boundaries of the inter-SH2 are decreased the is the region region primarily responsible of 3-kinase helix-2 amount of PI 3-kinase seen to bind is also decreased. the two subunits the activity PI complex, probably element to the of helix- This could be accounted for if disrupting the inter-SH2 donates a structural 1. Thus, folding in to a stable structure is the and thus order destabilizing putative domain, structure, regions corresponding produce helix-i of and the on interactions. When deletions in to 470-497 on preventing protein -protein region p85a adjacent interaction between the are for the the and those in the are compared helix-2 p85a protein p85(3 protein required probably of the and the subunit to take and considered with the loss activity, corresponding place. p85 p110 and the 10 can between the third of helix-i of the residues 478-492 The interaction quarter coiled-coil, p85 pl proteins in that the two between the two both in vivo and in is identified as the common region take p85a, vitro, place suggesting be cotranslated in order to associate. mutants that are unable do not need to that is absent in deletion p85 proteins proteins the subunit and the 10 subunit Deletions which The between to bind PI 3-kinase IOA). activity (Figure binding p85 p1 and no suitable conditions have been half of residues 452 -476 on is of remove either the top helix-i, very high affinity, of an active which result in dissociation PI or the lower of residues established (Figure lOD) quarter helix-i, p85a Ruiz-Larrea et et al., destabilize the coiled 3-kinase 492-523 on p85a (Figure lOC) al., 1992; complex (Fry known that the interaction between In it is and thus the association with the structure disrupt plO addition, 1993). not on the of subunits is but do not inhibit its when the two protein completely binding. Only presence dependent and which are contained in residues 478 -492 are observation) the removed, (R.Dhand, unpublished phosphotyrosine motifs in this of 10 residues 452 -492 in is there are no mutation which deletes p1 the p850, region proline-rich present 519 et al. R.Dhand which may provide a binding consensus for the SH3 domain Materials and methods of the p85 protein. Secondary structure prediction using the Materials Leeds Prediction Package (Eliopoulos, 1989), predicts that Restriction enzymes and DNA modification enzymes were obtained from some 60% of the N-terminal 120 residues, which have been standard commercial sources and used according to the manufacturers recommendations. Oligonucleotides were synthesized on an Applied shown to direcfly bind the p85 proteins, will adopt an a- Biosystems 380B DNA synthesizer and used directly in all subsequent helical conformation and some 20% a ,3-sheet conformation. procedures. Thus, the N-terminal region of pl 10 will consist of a cu/l Cells and viruses region, where the c-helices might form interactions with the Maintenance of insect cell (Sf9) culture was carried out as described by inter-SH2 region of p85. Since the major binding site on Summers and Smith (1987). Both p85a and p85,B proteins were expressed p85a is considered to be between residues 478 -492 on in insect cells using baculovirus vectors as described previously by Otsu helix-1, it is of interest that there is a small hydrophobic (1991) and Gout et al. (1992). Recombinant baculovirus expressing et al. 110 was described in Hiles et al. (1992). Mouse-L cells were maintained in this region formed by Met479 (C position on pocket in DMEM medium supplemented with 10% fetal calf serum. Figure 9C), Ala483 (G position on Figure 9C) and Ala486 (C position on Figure 9C). As the N-terminus of p1O is Recombinant DNA methods of GST-p85a fusion proteins. Sequences corresponding to - 48% hydrophobic in character, this pocket may be Construction the SH3 domain (amino acids 1-86), the BCR homology region (amino in forming a high affinity interaction with the p 10 important acids 125 -322) and the helical inter-SH2 domain (amino acids 425 -600, protein. 425-555, 451-616, 451-600, 451-555) of the bovine p85a subunit of is intriguing that a lipid binding site has also been It the PI 3-kinase (Otsu et al., 1991), were amplified by PCR and cloned into vector (Pharmacia). BamHI and EcoRI sites were identified in the inter-SH2 region of the p85 protein. Two the pGEX-2T expression the C-termini of oligonucleotides respectively to included at the N- and antibodies that specifically recognize this region monoclonal facilitate cloning. Stop codons were introduced at the end of each of the been shown to inhibit the binding of phospholipids, in have cloned sequences, 5' to the EcoRI site. PCR fragments were verified using to the p85 subunit (End et al., 1993). particular PI(4,5)P2 system (US Biochemicals). The full-length p85a (amino acids the Sequenase sequence analysis reveals the presence of a short Amino acid 1-724) was subcloned into pGEX-2T by digesting the pGEX-2T-SH3 with and EcoRI and replacing this fragment with a similarly cut cartridge the inter-SH2 region of the p85cx protein, Bgll basic motif in from p85a in the Bluescript vector (Otsu et al., 1991) which contained the found in profillin and gelsolin and similar to that remaining coding region of the p85ai protein. The C- and N- terminal SH2 to confer PI(4,5)P2 binding properties to these demonstrated of p85a and the p85aN-C construct have been described (Yonezawa domains et al., 1992; Yu et al., 1992). Moreover, proteins (Janmey 1992). p85aN-CA478-514 was constructed by the amplification of et al., residues are well-conserved in p85,B, which PCR products (P1 and P2) from the p85a cDNA. The first (P1) analogous two nucleotides 988-1434 and had a BamHI site introduced at encompassed that it might be important as a regulatory site for suggests the 5' end and a XbaI site at the 3' end. The second (P2) encompassed The function of this lipid binding site is these proteins. nucleotides 1540-2175 and had a XbaI site introduced at the 5' end and unknown, but there are several possible presently both EcoRI and KpnI sites at the 3' end. Both fragments, that is the As the p85 protein has been shown to bind pl 10 implications. BamHI-XbaI fragment of P1 and the XbaI-EcoRI fragment of P2, were ligated into the BamHI-EcoRI site of the pGEX-2T plasmid and sequenced. via the inter-SH2 region, the lipid binding site may form The expression of this mutant results in a mutant containing amino acids of the substrate binding pocket for PI 3-kinase, specific a part with the deletion of 35 amino acids from residues 330-724 of p85a Indeed, at the membrane where PI is the for PI(4,5)P2. 478-513 with the insertion of two amino acids, serine and arginine in the substrate, this region may be important in most abundant deleted region. pGEX-2T constructs were transferred into E.coli XLl-blue of PI(4,5)P2, thus increasing the affinity for and expression of GST fusion proteins was carried out as the selection (Stratagene) described previously (Smith and Johnson, 1988). Evidence from a number of cell lines this phospholipid. that PI(3,4,5)P3 is the key signalling lipid suggests of GST-p85i3 fusion proteins. Full-length p850, Construction by the PI 3-kinase (Stephens et al., 1991; Hawkins generated p85,B(486-724) and were described in p85,B(445-724), p850(516-724) Yonezawa et al. (1992). For p8513A486-516 GST-p8513(1-724) was 1992), thus implicating PI(4,5)P2 as the critical et al., with BamHI and BgEl to remove the fragment residues 486-5 16 digested substrate required for activation of the PI 3-kinase. It could and then self-ligated. p85I3A445-485 and -469 were constructed p85OA445 be that the conformational change induced in p85 speculated by amplifying two PCR products from p8513 cDNA (P3 for p8513A445 -485 upon peptide binding may cause a sterically masked and P4 for p8513A445-469). P3 encompassed nucleotides 1457-1966 and PI(4,5)P2 lipid binding site to become more accessible to P4 encompassed nucleotides 1408-1966, and both had an Accl introduced at the 5' end. Both P3 and P4 were digested with AccI and KpnI and then the substrate binding pocket of PI 3-kinase, thus resulting ligated into AccI and KpnI digested GST-p85,1-724 respectively. activation. The substrate specificity of purified in its 10 and comparison with purified p85a -p1 10 monomeric pl GST-pJJO fusion proteins. Regions of the pl O cDNA Construction of al., 1992) corresponding to amino acids 1-128 (pl 110-), 123-458 and mutagenesis of critical residues in the lipid (Hiles et complex (p110-3), 601-960 (p110-4), 760-960 (p110-5), (p110-2), 577-1068 binding site of p85 will be required to resolve this issue. (pl11-6), 1-49 (110-1.1), 1-81 (p110-1.2), 1-108 (p110-1.3), 760-1069 This study has localized the regions of p85a and p110 20-81 (pl 10-1.4) and 20-108 (p1 10-1.5), were amplified by PCR. BamHI involved in interacting with each other. This will permit the and EcoRI sites were added to 5' and 3' oligonucleotides respectively to of dominant negative mutants that may define a more use facilitate cloning. Following amplification, PCR products were cut with BamHI and EcoRI and ligated into pGEX-2T (Phamacia), GST fusion role for PI 3-kinase in mitogenic signalling pathways. precise constructs expressing amino acids 1-35 (p1 10-1.6) and 37-128 (p1 10-1.7) Blocking the association between the p85 and the p110 were constructed from (p 110-1) as follows. For plasmid p1 10-1.6, p 110-1 subunits may be a potential means for inhibiting signal was digested with BsmI and EcoRI, ends were made blunt by treatment transduction. The identification of the binding sites reported with T4 DNA polymerase and the plasmid was circularized with T4 DNA in this will aid the development of assays for inhibitors paper ligase. For plasmid p1 10-1.7, p 110-1 was digested with BsmI and the ends with T4 DNA polymerase. BgM linkers (12mers; New England made blunt of this interaction. Finally, a more detailed understanding Biolabs) were added to the ends and the plasmid further restricted with of the association between the two subunits of PI 3-kinase BamHI. The plasmid containing band was gel-purified and the plasmid will require solution and crystallographic studies of the two were transformed with T4 DNA ligase. Expression constructs recircularized fragments described and subsequent mutagenesis of critical into XL1-blue (Stratagene). Expression of GST fusion proteins was E.coli carried out as described previously by Smith and Johnson (1988). residues. 520 3-kinase intersubunit interactions Pi Downing,A.K., Panayotou,G., (Uout,l., Booker,G.W., Breeze,A.L., Antibodies and immunoprecipitation and Campbell,I.D. (1992) Nature, 358, 684-687. Waterfield,M.D. The pl10 C-terminal antiserum was raised as described by Hiles et al. Auger,K.R., Cohen,B., Carpenter,C.L., Duckworth,B.C., (1992). Affinity purified polyclonal antiserum raised against SDS-PAGE Schaffhausen,B.S. and Cantley,L.C. (1990) J. Biol. Chem., 265, purified p85a was described by Gout et al. (1992). Monoclonal antibodies 19704-19711. directed to the p85a and to the proteins are described in End et al. p85,0 Auger,K.R., Chanudhuri,M., Yoakim,M., (1993). Immunoprecipitations were carried out as described previously by Carpenter,C.L., Schaffhausen,B., Shoelson,S. and Cantley,L.C. (1993) J. Biol. Chem., Otsu et al. (1991). 268, 9478-9483. Pi 3-kinase assay Conway,J.F. and Parry,D.A.D. (1990) Int. J. Biol. Macromol., 12, assay for PI 3-kinase activity was carried out as described by End et al. The 328-334. (1993). and Kashishian,A. (1993) Mol. Cell. Biol., 13, 1737-1745. Cooper,J.A. (1953) Acta Crystallogr., 6, 689-697. Crick,F.H.C. Association assays Eliopoulos,E.G. (1989) Documentation for Leeds Prediction Programs. ofp110 with p85a and p85(3 in vivo. Sf9 cells were infected Association of Biophysics, University of Leeds. Department baculoviruses as described by Gout et al. (1992) and Hiles with appropriate Gout,I., Fry,M.J., Panayotou,G., Dhand,R., Yonezawa,K., End,P., Two days post-infection, cells were harvested and lysed in et al. (1992). and Waterfield,M.D. (1993) J. Biol. Chem., 268, Kasuga,M. mM pH 7.4, 150 mM NaCl, 50 mM NaF, 5 mM EDTA, 50 Tris-HCI 10066-10075. X-100, 500 sodium orthovanadate, 2 mM PMSF and 100 1% Triton uM Escobedo,J.A., Martin,G.A., Turck,C.W., Delrosario,M., Fantl,W.J., inhibitor units of aprotinin. Lysates were then either kallikrein and Williams,L.T. (1992) Cell, 69, 413-423. McCormick,F. using appropriate antibodies or were applied to a immunoprecipitated Fry,M.J. (1992) Curr. Biol., 2, 78-80. affinity column, DY751VPML(G), derived from sequence phosphopeptide Panayotou,G., Dhand,R., Ruiz-Larrea,F., Gout,I., Nguyen,O., Fry,M.J., Tyr751 of the human PDGF (3-receptor to which PI 3-kinase surrounding and Waterfield,M.D. (1992) Biochem. J., 288, Courtneidge,S.A. to bind (Otsu et al., 1991; Fry et al., 1992). Affinity has been shown 383-393. columns were washed as described. Enzyme complexes were analysed by Kowalskichauvel,A., Tartare,S. and Van Giorgetti,S., Ballotti,R., 7.5% SDS-PAGE gels or subjected to PI 3-kinase assays as described (1993) J. Biol. Chem., 268, 7358-7364. Obberghen,E. above. Panayotou,G., Fry,M.J., Hiles,I., Otsu,M. and Gout,I., Dhand,R., Waterfield,M.D. (1992) Biochem. J., 288, 395-405. Association ofPI 3-kinase activity with wild-type and mutant p85ac in vivo. Hawkins,P.T., Jackson,T.R. and Stephens,L.R. (1992) Nature, 358, SRa plasmids (Takebe et al., 1988) for the expression of wild-type bovine 157-159. (Wp85a) or a mutant bovine fiull-length 85a (p85aA478-513), which p85a Herman,P.K. and Emr,S.D. (1990) Mol. Cell. Biol., 10, 6742-6754. lacks a binding site for p 10 were constructed as described (Hara et al., Hiles,I.D. et al. (1992) Cell, 70, 419-429. in preparation) and designated as SRa-Wp85a and SRa-p85aA478-513, Hodges,R.S., Sodek,J., Smillie,L.B. and Jurasek,L. (1972) Cold Spring respectively. Semi-confluent mouse L cells were transfected with 3 of Ag Harbor Symp. Quant. Biol., 37, 299-310. either SRa-Wp85a or SRci-p85aA478-513 plasmids using the DEAE Jamney,P.A., Lamb,J., Allen,P.G. and Matsudaira,P.T. (1992) J. Biol. dextran method. After 72 h cells were lysed in a buffer containing 20 mM Chem., 267, 11818-11823. Tris pH 7.6, 1% Nonidet P-40, 10% glycerol, 137 mM NaCl, 1 mM Kunz,J., Henriquez,R., Schneider,U., Deuterreinhard,M., Movva,N.R. and MgCl2, 1 mM CaCl2, 1 mM DTT, 1 mM PMSF and 1 mM sodium Hall,M.N. (1993) Cell, 73, 585-596. and transiently expressed bovine p85s immunoprecipitated orthovanadate, Ling,L.E., Druker,B.J., Cantley,L.C. and Roberts,T.M. (1992) J. Virol., with 2 of a anti-p85a monoclonal antibody (G12) (Yonezawa et al., 1992) Ag 66, 1702-1708. to G -agarose. These samples were then subjected to Western bound protein Mclachlan,A.D. and Stewart,M. (1975) J. Mol. Biol., 98, 293-304. blotting analysis or PI 3-kinase assay. Morgan,S.J., Smith,A.D. and Parker,P.J. (1990) Eur. J. Biochem., 191, 761 -767. association In vitro complex formation of the p85a or the In vitro assay. Otsu,M. et al. (1991) Cell, 65, 91-104. with 110 was examined by immunoprecipitating p85 from p85(3 subunit p (1975) J. Mol. Biol., 98, 519-535. Parry,D.A.D. antibodies. The immunoprecipitate was collected insect cells using appropriate and Waterfield,M.D. (1992) Trends Cell Biol., 2,358-360. Panayotou,G. A beads and washed three times in lysis buffer. This on protein -Sepharose Bax,B., Gout,I., Federwisch,M., Wroblowski,B., Dhand,R., Panayotou,G., was then incubated for 1 h at 40C with cell lysates immune complex preMared Fry,M.J., Blundell,T.L., Wollmer,A. and Waterfield,M.D. (1992) EMBO infected with viruses expressing p1 10. The resultant enzyme from Sf9 cells 4261-4272. J., 11, were washed extensively and either analysed on 7.5 % complexes Parker,P. and Waterfield,M.D. (1992) Cell Growth Differ., 3, 747-752. or for associated PI 3-kinase activity. SDS-PAGE gels assayed directly and Gish,G.D. (1992) Cell, 71, 359-362. Pawson,T. of 10 to domains was performed in essentially the In vitro binding pl p85a Pawson,T. and Schlessinger.J. (1993) Curr. Biol., 3, 434-442. GST fusion proteins were immobilized on same manner. Lattman,E.E., Cummins,P., Lee,K.Y. and Cohen,C. Philips,G.N.,Jr, beads as described previously (Smith glutathione-Sepharose (Pharmacia) 278, 413-417. (1979) Nature, These columns were then incubated for 1 h and Johnson, 1988). affmity Fillers,J.P. and Cohen,C. (1986) J. Mol. Biol., 192, Philips,G.N.,Jr, from Sf9 cells infected with a baculovirus at with cell lysates prepared 4°C 111-131. 10. After extensive washing with lysis buffer, bound proteins expressing p1 Gout,I., Waterfield,M.D. and Cantrell,D.A. (1993) J. Biol. Chem., Reif,K., kinase For association of the GST-pl 10 constructs were subjected to assays. 10780-10788. 268, cell of Sf9 cells infected with baculoviruses with p85a and p85(3, lysates et al. (1993) Biochem. J., 290, 609-616. Ruiz-Larrea,F. or a control were added to either p85a, p85(3 wild-type expressing Fry,M.J., Stack,J.H., Waterfield,M.D. and Schu,P.V., Takegawa,K., 10 fusion and was allowed to proceed immobilized GST-pl proteins, binding (1993) Science, 260, 88-91. Emr,S.D. The beads were then washed four times with lysis buffer for 2 h at 4°C. Homma,Y. and Takenawa,T. (1991) J. Biol. Chem., 266, Shibasaki,F., in 100 ml 2 x Laenmli buffer. Samples were boiled and sample resuspended 8108-8114. 5 resolved SDS-PAGE on 10% gels. for min and proteins by Shoelson,S.E., Sivaraja,M., Williams,K.P., Hu,P., Schlessinger,J. and EMBO 12, 795-802. Weiss,M.A. (1993) J., Acknowledgements Smith,D.B. and Johnson,K.S. (1988) Gene, 67, 31-40. and (1991) Nature, 351, 33-39. Stephens,L.R., Hughes,K.T. Irvine,R.F. thank Oahn for peptide synthesis and Alistair We would like to Truong and Tex. Exp. Stat. Bull., 1555, 1-56. Summers,M.D. Smith,G.E. (1987) for Sterling oligonucleotide synthesis. Hoy,P., Yokota,K., Arai,K., Yoshida,M. Takebe,Y., Seiki,M., Fujisawa,J., Mol. Cell. Biol., 8, 466-472. and Arai,N. (1988) and Kazlauskas,A. (1993) Cell, 73, 321-334. Valius,M. References Gout,I., Calzolari,E., Patracchini,P., Otsu,M., Hiles,I., Volinia,S., and (1992) Oncogene, 7, EMBO 3469-3479. Rooke,L. Waterfield,M.D. et al. J., 11, Bernardi,F., Backer,J.M. (1992) Brice,M.D., 789-793. Williams,G.J.B., Meyer,E.F.Jr, Bernstein,F.C., Koetzle,T.F., et al. J. Biol. 25958-25965. Shimanouchi,T. and Tasumi,M. (1977) J. Chem., 267, Kennard,O., Rodgers,J.R., Yonezawa,K. (1992) and J. Biol. Chem. Mol. 535 -542. Yin,H.L. (1992) Biol., 112, Yu,F.X., Sun,H.-Q., Janrney,P.A. 14616-14621. and Thorton,J. (1983) Nature, 306, Borkakoti,N. 267, Blundell,T.L., Barlow,D., revised on October 1993 on 7, Received 25, 1993; 281-283. August http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The EMBO Journal Springer Journals

PI 3‐kinase: structural and functional analysis of intersubunit interactions.

Download PDF
11 pages

Loading next page...
 
/lp/springer-journals/pi-3-kinase-structural-and-functional-analysis-of-intersubunit-lC7jV08tII

References (2)

Publisher
Springer Journals
Copyright
Copyright © European Molecular Biology Organization 1994
ISSN
0261-4189
eISSN
1460-2075
DOI
10.1002/j.1460-2075.1994.tb06289.x
Publisher site
See Article on Publisher Site

Abstract

EMBO Journal vol.13 no.3 pp.511 -521, 1994 The Pi 3-kinase: structural and functional analysis of intersubunit interactions 3, derivatives to produce PI 3-phosphate Ritu Dhand', Kenta Hara2, Ian Hiles1 phosphorylated [PI(3,4)P2] and PI 3,4,5-tris- [PI(3)P], PI 3,4-bisphosphate Ben Bax4, Ivan Gout', George Panayotou', [PI(3,4,5)P3]. Rapid increases in the levels of phosphate Michael J.Fryl 5, Kazuyoshi Yonezawa2, and PI(3,4,5)P3 have been observed when PI(3,4)P2 Masato Kasuga2 and Michael D.Waterfieldl"6,7 cells are stimulated with peptide growth factors quiescent for Cancer Research, 91 Riding House Street, ILudwig Institute such as platelet-derived growth factor (PDGF) and epidermal 8BT, 6Department of Biochemistry and Molecular London, WIP growth factor (EGF) (reviewed in Parker and Waterfield, University College, Gower Street, London, WC1E Biology, 6BT, or transformation by polyoma virus (Ling et al., 1992) 2The Second of Internal Medicine, Kobe University School Department 1992). The 3-phosphorylated lipids appear not to be of Medicine, 7-5-1 Kusunoki-cho, Chuo-ku, Kobe 650, Japan and 4ICRF Unit of Structural Molecular Biology, Department of substrates for known isoforms of phospholipase C and thus Birkbeck College, Malet Street, London, Crystallography, themselves play a direct second messenger role. may WC1E 7HX, UK Determination of the primary structure of the PI 3-kinase address: The Wellcome Research Laboratories, Beckenham 3Present fundamental towards understanding its role in signal has been BR3 3BS, UK Purified PI 3-kinase from brain or transducing processes. 5Present address: Institute of Cancer Research, Cell Biology and liver is a heterodimer of 85 and 110 kDa subunits (Carpenter Experimental Pathology, Haddow Laboratories, 15 Cotswold Road, et Morgan et al., 1990; Shibasaki et al., 1991; al., 1990; Sutton, Surrey SM2 5NG, UK Fry et al., 1992). Through cDNA cloning two forms of p85, 7Corresponding author termed p85a and p85,B, which lack PI 3-kinase activity have Communicated by M.D.Waterfield been described (Otsu et al., 1991). Analysis of the primary of p85 reveals a protein with a multidomain sequence Phosphatidylinositol (PI) 3-kinase has an 85 kDa subunit structure which contains a number of non-catalytic domains (p85a) which mediates its association with activated first described in studies of transforming proteins of the src a Src homology region 3 protein tyrosine kinase receptors through SH2 domains, family. At the N-terminus lies and an 110 kDa subunit (p110) which has intrinsic domain (SH3) (reviewed in Pawson and Schlessinger, 1993) sequence catalytic activity. Here p85a and a related protein p853 and next to this is a region with significant region gene are shown to form stable complexes with recombinant similarity to the product of the breakpoint cluster pllO in vivo and in vitro. Using a panel of glutathione BCR (Otsu et al., 1991). Since the BCR protein has a GAP interact with small fusion proteins of the inter-SH2 activity for rac, this region of p85 may S-transferase (GST) in The C- region of p85, 104 amino acids were found to bind GTP binding proteins (reviewed Fry, 1992). two SH2 the pllO protein, while deletion mutants within terminal half of the molecule is dominated by directly which flank this further defined the binding site to a sequence domains in Pawson and Gish, 1992) region (reviewed adopt a helical of 35 amino acids. Transient expression of the mutant a region that we have predicted may The of these in mouse L cells showed it was unable to conformation (Panayotou et al., 1992). presence p85ct protein domains that the p85 proteins may bind PI 3-kinase activity in vivo. Mapping of the distinct functional suggests and regulatory roles. site of interaction on the pllO protein have multiple interactive complementary of p85 has facilitated both structural defined 88 amino acids in the N-terminal region of pllO The modular structure studies on this and has led to rapid which mediate the binding of this subunit to either the and functional protein in our of its role in the PI 3-kinase or the The inter-SH2 region of p85 advances understanding p85a p853 proteins. little is known as is to be an folded module of a complex, however, by comparison, very predicted independently 110 a single form of pl 10 has coiled-coil of two long anti-parallel c-helices. The yet about the p protein. Only and studies have clearly structure of p85 suggests a basis for the been cloned to date expression predicted can alone encode a with a PI 3-kinase intersubunit interaction and the relevance of this shown that this protein to the of the PI et 1992). The cloned pl 10 is homologous interaction with respect regulation activity (Hiles al., to a PI 3-kinase which is involved in vacuolar 3-kinase complex is discussed. Vps34p, yeast Schu et structure/SH2 domains and Emr, 1990; al., 1993). Key words: PI 3-kinase/protein protein sorting (Herman reveals redundant kinase of the p1O sequence Analysis with and the more motifs and comparisons Vps34p, sequence another PI 3-kinase cloned Tor2, putative yeast recently Introduction for (Kunz et al., homologue required progression GI a domain to be assigned to the 3-kinase play a key role in 1993), has allowed catalytic Phosphatidylinositol (PI) may et of the al., 1992). linked directly or indirectly to C-terminal region protein (Hiles intracellular signal processes of 10 reflects in some cells be an of the N-terminal p1 and may important Sequence analysis region diverse receptor types to known and this lead to et no any proteins in the events which mitogenesis (Fantl al., significant homology mediator a role. 3-kinase region may play regulatory and PI phosphoryl- 1992; Valius Kazlauskas, 1993). a series of such of the inositol of PI and its PI 3-kinase may be regulated by events, ates the D3 ring position Press Oxford University R.Dhand et al. with proteins, little is known of the interaction between the p85 as translocation of PI 3-kinase to cell membranes In consequent access to substrate, and the subsequent and p1 10 subunits. this study we have used both the p85ct association with, and phosphorylation by, protein tyrosine and the p853 proteins in complementary studies to define in with kinases (PTKs) (reviewed Panayotou and Waterfield, the regions responsible for stable subunit association 1992). The binding of PI 3-kinase through its SH2 domains vivo the p110 protein both in and in vitro. We predict that is able to phosphotyrosine-containing sequences on PTKs the structure of the region of p85 responsible for the of of of to bring about independently some degree activation interaction with pl 10 is an independently folded module ct-helices. The of the Tyrosine phosphorylated peptides corresponding a coiled-coil of two possible significance enzyme. to potential PI 3-kinase sites on the insulin receptor substrate, the intersubunit interaction with respect to activation of the have been shown to IRS-1, or the intact IRS-1 protein, enzyme is discussed. et activate PI 3-kinase in vitro (Backer al., 1992; Giorgetti Results of a et al., 1993). Similarly, addition phosphotyrosine- to Y751 of the PDGF3 p110 can reconstitute with p85a or to form containing peptide corresponding p85,3 also causes activation active Pi 3-kinase complexes both in vivo and in vitro receptor, a known in vivo binding site, et it The p85a subunit is known to form a stable complex with of PI 3-kinase in vitro (Carpenter al., 1993). Further, this Y75 1 the p1O subunit when they are coexpressed in Sf9 cells has been shown that upon binding phosphopeptide, and its N-terminal SH2 (Hiles et al., 1992). To investigate whether p85(3 will also the p85 subunit of the PI 3-kinase conformational reconstitute with the p110 protein in an active complex, domain in particular undergo a change et This induced insect cells were coinfected with either p85ac and pl10 or (Panayotou et al., 1992; Shoelson al., 1993). of be transmitted p853 and p1 10 recombinant baculoviruses and PI 3-kinase change in conformation the p85 protein may subunit and thus contribute to its activity was measured. In double infections of baculoviruses to the associated p1O of the nature of the interaction expressing p85a and p 10, both proteins were detected in activation. An understanding subunit and its domain either anti-p85a or anti-p 1 10 immunoprecipitates, or when between the regulatory p85 catalytic on the mechanism of activation of the enzyme was bound to Y751 phosphopeptide affinity would shed more light 5 and All and immune there is a wealth of biochemical and beads (Figure lAi, lanes 4, 6). affinity PI 3-kinase. Although on the association of PI 3-kinase with other precipitates were found to possess PI 3-kinase activity structural data l 1iii ,. 1) /Ifi I, V III /S I II \Pi% __ h Is .2 s>4 %~ ).i . i, \l\\\ \I *kJI 1NM R1D£l k1I),i ... i- SO 1) (I 10 - tei "( (I ...i pi --I *1)11 1 *1) 1I1 q11p"~4P - 4- 5%4 p1S *- p's Bii I(ii 5 ) III 1i) l1 r.) IX5l /1 4 5 ii 2 x I'll, Fig. 1. Both p85ca and p85(3 are able to for-m stable active complexes when coinfected with pl10 in insect cells both in vivo and in vitro. (Ai) Sf9 cells were infected with either p85ca (lane 1), p85(3 (lane 2) or pllO (lane 3); viruses were coinfected with p85a and viruses (lanes 4-6), or pllO with p85,B and plIO viruses (lanes 7-9). Lysates of these cells were precipitated using polyclonal affinity purified antibodies raised against p85a (lanes 1 and 4), monoclonal antibodies raised against p85(3 (lanes 2 and 7), polyclonal affinity purified antibodies raised against 10 (lanes 3, 5 and pl 8) or bound to immobilized Y751 phosphopeptide affinity beads (lanes 6 and 9). Samples were analysed by Coomasssie blue staining of SDS-PAGE gels. (Aii) PI 3-kinase assays were performed on immunoprecipitates infected and treated as described in Ai. (Bi) and p85(3 were p85ai immunoprecipitated from insect cells and then incubated with lysates of Sf9 cells that had been infected with viruses (lanes 1 and 2). pllO Complexed proteins were resolved on SDS-PAGE and identified by staining with Coomassie blue. (Bii) 3-kinase assays were performed on PI immunoprecipitates infected and treated as described in (Bi) Lysates of insect cells that had been infected with viruses were (Biii) plIO immunoprecipitated and then either left untreated (lane 1) or incubated with p85a-containing Sf9 cells (lane 2). Samples were washed in lysis buffer, resolved on SDS-PAGE and identified by Coomassie blue staining. 512 Pi 3-kinase intersubunit interactions (Figure lAii, lanes 4, 5 and 6). Similar results were seen To study the interaction between the inter-SH2 region of with viruses expressing the fl-form of p85 and p 110. Infected p85a with pl10 in the context of the entire p85 protein cells immunoprecipitated with antibodies directed to the monoclonal antibodies which have epitopes within the inter- p850 if or the plO subunit, or bound to Y751 phosphopeptide SH2 region were used to block intersubunit associations. affinity beads, were found to contain both and p110 p85a from insect cells was immunoprecipitated, either using p850 proteins when analysed by SDS -PAGE (Figure lAi, lanes two monoclonal antibodies which bound located epitopes in 8 and 7, 9). Examination of associated PI 3-kinase activity the inter-SH2 region of p85a, or with two control antibodies in parallel showed the 10 complex to be that recognize the BCR domain (End et al., 1993). These p850-pl enzymatically active (Figure lAii, lanes 7, 8 and 9). The immunocomplexes were washed stringently and incubated levels of Coomassie blue staining protein precipitated in the with lysates of insect cells that had been infected with pl 10 p85,B-pl 10 complex are essentially the same as those seen virus. Bound proteins were then subjected to analysis by precipitated in the complex using p85oa-plO identical SDS -PAGE and PI 3-kinase assays. Similar quantities of amounts of both plO antibody or Y751 phosphopeptide p85ca were found to be immunoprecipitated by all four affinity beads. Since both p85a and p85fl are expressed to antibodies, as determined by staining with Coomassie blue approximately the same levels in insect cells (Gout et (Figure 3A, lanes 1-4). However, the amount of PI al., 1992), this would suggest that 10 complex 3-kinase activity which bound to p85ai immunoprecipitated p850-pl formation in Sf9 cells is as efficient as p85x-p1 10 complex with the two antibodies that recognize the inter-SH2 region formation. Indeed, as comparable amounts of PI 3-kinase was significantly less (Figure 3B, lanes 3 and 4), than the activity were precipitated with either the p85ac-p110 activity which bound to p85cx immunoprecipitated with complex or the 10 complex this suggests that p1 10 antibodies to the BCR domain (Figure 3B, lanes 1 and 2). p850l--pl is able to form a stable, active PI 3-kinase complex with either of the p85 isoforms in insect cells. Mapping the precise site in the inter-SH2 region To ascertain whether the two subunits of the PI 3-kinase required for the interaction with p1 10 complex were able to associate post-translationally in vitro, To define which component of the inter-SH2 region is one subunit of the PI 3-kinase complex was first immobilized involved in directly binding the pl 10 subunit, deletions were onto an appropriate matrix and then allowed to bind the other subunit in solution. First, the p85 proteins immunoprecipi- tated from infected insect cells using isoform-specific 110 78 274 108 329 41640 603623 706 724 monoclonal antibodies were collected on protein A- 1. GST-p85a SH3 , BCR Sepharose beads. These immunocomplexes were 1 | NSH21 InIer-SH2 then incu- | ,SH2I | bated with lysates of insect cells that had been infected with 2. GST-CSH2 pl 10 viruses, after which they were washed and subjected 1i[ to either analysis by SDS-PAGE or PI 3-kinase assays. 3. GST-InterSH2 Inter-SH2 Alternatively, p1 10 was immunoprecipitated from infected 329 433 insect cells, incubated with p85a-expressing lysates 4. GST-NSH2 and then treated as described above. Both p85a and p85,B were seen 113 279 S. GST-BCR to bind p1 10 by Coomassie blue - staining following SDS PAGE (Figure lBi, lanes 1 and 2) or upon assay for 84EI 6. GST-SH3 associated PI 3-kinase activity (Figure lBii, lanes 1 and 2). Similar results were obtained when p 110 was immobilized and the bound p85a protein was detected as a Coomassie blue-stained protein on SDS-PAGE following binding * P13K activity in vitro (Figure lBiii, lane 2). CPM bound (cpm) The inter-SH2 region of p85a is required for binding to p110 The specific region of p85a responsible for binding to p1 10 was defined initially by analysis of the domains of p85 which may be involved in the interaction. Glutathione S-transferase (GST) fusion proteins of the various subdomains of p85a were immobilized on glutathione-Sepharose beads (Figure 2A) and used for affinity analysis with cell lysates prepared from Sf9 cells infected with a recombinant baculovirus expressing pl 10, after which the complexes were alone and washed and subjected to PI 3-kinase assays. GST beads GST-p85a bound to glutathione-Sepharose °TI2 3 4o 5 6 7 were used as controls. incubated with wild-type lysate GST-Fusion protein associated Figure 2B shows that PI 3-kinase activity only with either the or full-length GST-p85a protein (lane 1) 2. The inter-SH2 of is for Fig. region p85a required binding p1l0. with the of the PI 3-kinase inter-SH2 region p85ca (lane 3). domains of and GST (A) Full-length GST-p85a (1) p85a (2-6). (B) did not bind to GST alone (lane and no activity 7) A and GST alone were Domains 2-6 of as described in (lane p85a 7) insect cell 3-kinase bound to full- endogenous PI activity and then incubated with bound to beads glutathione-Sepharose lysates when incubated with control infected with virus. After length GST-p85a protein of insect cells that had been plIO washing were to 3-kinase with lysis buffer, proteins subjected PI assays. not lysates (data shown). 513 R.Dhand al. at ,A ~ ~~ *X p8X5Lt i)>S t. .}j_ ..... "' 5--XFi, -- il -X 12,i 24 32) i2 4411 (03 f 23 0 '24 'k -.:. j__________ i 2 I .) 4~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ W C, [S lA K12 [ |.} | lilter-SH-1 12| 425l; l [ : 1113K activits (1 bound11( (I0000 6H5 8(X(O() it 6000-) l Pl.3Ka( uti%it I1111 ,4Nl 4000(1 I - 2 3 4 I 111r t'- SF12- 15 41'k'll ) _ _ .2 ))1 0( Iiter- S112-9 l ' 4 3,:1 in of Fig. 3. Antibodies with epitopes the inter-SH2 region p85a _2 4 ti inhibit the binding of the pllO protein. was (A) p85a 1. .. %-I 1 oltssbl 1 prolvil cells with monoclonal immunoprecipitated from infected insect in and antibodies which have epitopes either the BCR domain (lanes 1 or in (lanes 3 and and 2) with epitopes the inter-SH2 region 4) Fig. 4. Identification of the region of the inter-SH2 that with region directly incubated pl1O-containing lysates. Immunoprecipitated p85 binds plO. (A) Full-length GST-p85a (1) and varying sizes of the visualized Coomassie blue of SDS-PAGE proteins were by staining inter-SH2 region (2-6). (B) The different sized GST-inter-SH2 3-kinase were on gels. (B) PI assays performed inimunoprecipitates regions were bound to glutathione-Sepharose beads and incubated infected and treated as described in A. with plO-containing Sf9 cell lysates (lanes 1-6). Samples were washed and assayed for associated PI 3-kinase activity. introduced at both ends of the inter-SH2 of region p85a, reducing the size of this from 175 amino acids region (425 -600) down to 104 residues (451-555) (Figure However, mutants with deletions which advanced further 4A). into the as those in The truncated inter-SH2 regions expressed as GST fusion inter-SH2 region GST -p85,B(486-724) - which in proteins were bound to glutathione Sepharose beads and GST -p85((516-724) were unable to bind PI 3-kinase lanes 3 and These were then incubated with lysates of insect cells that had been activity (Figure SB, 4). results suggest infected with 10 and was that 71 amino acids in p1 associated PI 3-kinase (residues 445-516) were activity p85( measured. The results in 4B (lanes show that involved in the interaction between the Figure 2-6) p85,B protein and the deletion of amino acids from either end of the inter-SH2 0. In order to locate the site in p1 binding p85( more region gradually reduced the amount of bound PI 3-kinase precisely, mutants that contained deletions in this such region This be if the inter-SH2 as activity. may explained disrupting GST-p85(3A486-516, GST-p85,3A445-485 and structure is a domain and thus were destabilizing putative prevents GST-p85(3A445-469 constructed (Figure SA, mutants protein -protein interactions. In addition, these results 5, 6 and 7). The function of these mutants was then assayed identified a structural element of 104 amino acids, between as described above. Following binding, PI 3-kinase assays residues 451-555 of the inter-SH2 domain of p85a, which revealed that both and GST-p85OA486-516 is able to bind the directly p1O protein and associated PI GST-p85j3A445-469 were still able to bind -65% of the 3-kinase activity (Figure 4B, lane 6). PI 3-kinase activity (Figure SB, lanes 5 and 7), in A series of nested deletions of fusion proteins with comparison wild-type GST-p85(3. However, the GST-p850 from the 3' end of the N-terminal SH2 domain to the 5' end deletion mutant (analogous to residues GST-p85OA445-485 of the C-terminal SH2 domain were then constructed to 452 -492 in p85cx) was unable to bind any detectable PI refine the binding site within the inter-SH2 region 3-kinase activity (Figure SB, lane 6), thereby defining 40 (Figure SA). The interaction of these GST fusion proteins amino acid residues that are absolutely required for the with p1 10 was then measured as described above. Full-length binding of PI 3-kinase activity to p85(. GST-p853(1-724) and GST-p85((446-724), which To assess whether a similar region of p85a was involved contains most of the inter-SH2 region, were clearly able to in binding to the p1 10 protein, an analogous deletion to that bind PI 3-kinase activity (Figure 5B, lanes 1 and 2). made in was p85( constructed (Figure 5A, mutant 6) and 514 interactions Pi 3-kinase intersubunit 73 254 329 41644 463423 706724 110 129 (A) (B) I nca Ns1121: 14m 1.GSTpUp (1724) p85I~ lismi |H2 445 724 l wSH2, GSTpiP (445.724) 406 724 516 724 r GSTpLS( (516-724) 1 436-516 724 s G 5.GSTpt5L4516 |I I rs I S.z |1|1 II WR GSTpLVM4S485 ISM I I I -fq I .Mul 16. 0 0 I 724 I I vu KA W | 1 7. w CI HI G6Fp3SA4.6 I|| I I kH| w m 110 78 129 254 329 41644 43623 786724 G G12 G 12 NMG p85a IISH3 |a ' 1w lH2 LG8STp (1-724) 329 41644 463623 78 724 l-S2 .a||| 9. Fig. 6. Semi-confluent mouse L cells were transfected with either GSTpSSaNC H=40 SRa-Wp85cz or SRa-p85aA478-513 plasmid. Transiently expressed 329 416 478.13 623 786 724 with an Wp85a (W) or p85aA478-513 (M) were immunoprecipitated z 10. GSTp3SaNCA47S513 |NLSFj2 I anti-p8Ma monoclonal antibody (G12) or control mouse IgG (NMG). The immunoprecipitates were either subjected to Western blot analysis using a polyclonal anti peptide rabbit antiserum (A), or samples were * % P13K subjected to PI 3-kinase assays (B). (-) indicates untransfected cells. aqvk boud G12, which is unable to recognize murine p85a (Yonezawa et al., 1992) and then either visualized by Western blot C& the analysis using a polyclonal antibody which recognizes jI46 C-terminal of or assayed for associated PI 3-kinase p85cx Figure 6A show that equivalent activity. The results in wild-type and mutant p85CY were amounts of the from these cells (lanes 2 and 3) and no immunoprecipitated I.29 mouse p85a is immunoprecipitated from endogenous untransfected cells (lane 1). However, PI 3-kinase activity 1 2 3 4 5 6 7 3 9 10 is found to be associated with the wild-type p85a GST-Ih.I protel. only 3) and no PI 3-kinase activity is observed (Figure 6B, lane in of the mutant p85CY (Figure 6B, lane 5. Amino acid 478-513 of p85a and 445-485 of p85,B immunoprecipitates Fig. residues are for of (A) Full-length p85(3 (1), 5' necessary the binding pllO. or from untransfected cells 6B, lane 1). 4) (Figure nested deletions of p85(3 (2-4), deletion mutants of p85,B (5-7), full- p85aNC (9) and deletion mutant of p85aNC (10). length p85a (8), for binding on p110 Mapping the site p85 GST fusion proteins of and p85(3 were bound to (B) p85a To map the regions of p 110 which could interact with p85a beads and then incubated with plO-containing glutathione-Sepharose cell (lanes were then washed and and p853, GST-pl10 fusion proteins encompassing various Sf9 lysates 1-10). Samples to PI 3-kinase subjected assays. regions of p110 were made (Figure 7A) and their ability to bind p85 expressed in an insect cell expression system for its to bind PI 3-kinase activity as described assessed. Results in Figure 7 show that of the six fusion assayed ability 0. to bind to proteins piIO. I-pi 10.6, only p1 was able above. This deletion mutant, GST-p85cxNCA478-513 with No 471-508 in was p85t (lane 16) and p853 (lane 17) high affinity. (analogous to amino acid residues p85,B), in the control Sf9 cell and no lane binding was detected lysate unable to bind PI 3-kinase activity (Figure SB, 10) seen in made with which binding to either subunit was experiments compared with a similar construct, GST-p85aNC, This indicates that residues lane These results define any of the other constructs. lacks the deletion (Figure SB, 9). 10 site for the p85 proteins. in that is to mediate 1-128 of p1 contain a binding a 35 amino acid region p85a necessary for to subunit. To investigate further the sequences required binding binding of the p1O catalytic which seven GST fusion proteins (p1 10-1.1-pl 10-1.7) p85, the first 128 amino acids of were Association of PI 3-kinase activity with wild-type and encompassed p110 Of pI 10-1.3 and mutant in vivo constructed (see Figure 8A). these, p85a and lanes for 10 10-1.5 bound both 8, 20, of the necessary binding p1 p1 p85at p85( (Figure As the regions p85 proteins None of the other constructs been examined in 21 and 15 and 3-kinase have using 14, respectively). PI activity only to bind either p85 subunit. these and a mutant with showed ability Together vitro the full-length p85a p85a any analyses, acid residues 20-108 of are able to define amino were constructed results the site deleted (p85ctA478-513) binding which is for a minimum structural element in mouse L cells. The recombinant pl 10 as required for transient expression It should be noted that the the subunit. and mutant were binding p85 bovine wild-type p85a proteins specifically 0 to be fusion with an monoclonal GST-pl1 proteins appear very susceptible immunoprecipitated anti-p85a antibody 515 R.Dhand et al. to degradation which explains the presence of numerous The inter-SH2 region of p85 is predicted to be a proteins on SDS -PAGE gels even after affinity binding to of two anti-parallel a-helices. coiled-coil glutathione- Sepharose beads. that the inter-SH2 domain mediated inter- The demonstration led us to evaluate further the predicted subunit association of this region of the p85 protein. Previously we structure that this region would be largely a-helical in A had predicted et al., 1992) and suggested that it might character (Panayotou bundle. To extend this work, the sequence be a four-helix region of p85a was used to search the amino of the inter-SH2 in the Owl 19.0 database (Proterin acid sequences Leeds University, UK); 12 of the top 20 Engineering Club, from the proteins myosin or hits were with sequences have coiled-coil regions, the top five hits paramyosin which ., of p85. Although the percentage being with isoforms between the aligned amino acid sequences sequence identity low, heptad repeats (see below) were often was relatively The amino acid sequences of proteins found to be aligned. Brookhaven databank of protein structures (Bernstein in the B also searched with the sequence of the inter- et al., 1977) was Here two of the top three hits were with SH2 region. a two stranded oa-helical coiled-coil muscle tropomyosin, P2.{0-3 P11iO-4 P110-6 P110-5 P110-2 P110-l structure has been solved to 15 A resolution protein whose 0- .-- 1-- ---- - 40- I-.I-- . - -. 1979, 1986). Although the amino acid (Philips et al., with was quite low (20% over sequence identity tropomyosin the heptad repeats in the inter-SH2 region 175 amino acids), were in register for a considerable and that of tropomyosin Sequence analysis of tropomyosin part of the alignment. McLachlan and Stewart, 1975; Parry, (Hodges et al., 1972; an unbroken series of heptad repeats 1975) had shown where residues a and d are conserved as (abcdefg)n, residues, which suggested that the structure hydrophobic a coiled-coil of oa-helices. Our earlier sequence would be of the helical domain of p85a and p85,B (Panayotou analysis 7. Identification of a of which binds (A) region p1II0 p85. Fig. showed the presence of two long unbroken et al., 1992) view of the of as GST fusion regions pl110 expressed Diagrammatic of the form (abcdefg)n where residues a and heptad repeats Numbers to the amino acid residues at the proteins. correspond extremes of the of I1 0. (B) The different expressed regions p d are hydrophobic. fusion were bound to glutathione-Sepharose proteins GST-plI10 in the inter-SH2 region of p85 and the The heptad repeats beads and incubated with 519 cell p85j3 or lysate containing p85a, of its amino acid sequence with sequences of similarity control were washed and bound (CON) wild-type proteins. Samples such as tropomyosin, paramyosin and coiled-coil proteins resolved SDS-PAGE on a 10% gel. The arrow indicates proteins by that the domain might consist of a coiled- the of the myosin suggested position p85 proteins. 7 1 P110-1 4 P 1 10-1 3 P11- 1 .2 P 101- .6 fP1C0-1 ; 1- PI 1v ., '1r-1 Bsm 128 -,- .- C 12 X - , X c , i.X, A,,, ,, P112-1 l - 49 pi10-11 - 97 1 . RI pulO-i -66 1 1 pi10 3 .. _ _ a ;L _. ... 43 pulO-l -40 plO- 1.5 1 3 1 6 __ p110 .. .... 37 -_ 7 31 p110 L._ -t.- W ...... Fig. 8. Fine scale mapping of a region of pl10 capable of binding p85. (A) Diagrammatic view of the regions of p1 10-1 expressed as GST fusion proteins. The BsmI site used in the construction of p1 10-1.6 and pI10-1.7 is shown. Numbers correspond to the amino acid residues at the extremes of the expressed regions of pl10. (B) The different GST-pllO-l fusion proteins were bound to glutathione-Sepharose beads and incubated with Sf9 cell lysate containing p85a, p85,3 or control wild-type (CON) proteins. Samples were washed and bound proteins resolved by SDS-PAGE on a 10% gel. The arrow indicates the position of the p85 proteins. 516 interactions Pi 3-kinase intersubunit coil of a-helices. Analysis of the sequence of the inter-SH2 The coiled-coil has two faces, one having residues BEAaeb region showed that its sequence is consistent with the and the other residues we shall call these the AEB CGDdgc; structure being an anti-parallel coiled-coil of two 70-residue face and the DGC face (Figure 9C). In our model the first a-helices (residues 441-512 and 518-588 in p85a; see helix starts at residue 441, a G position, and it may be that Figures 9 and 10). The proposed register of the helices was preceding residues (431-440) pack against the DGC face that which gave the maximum number of possible ion pairs. at the bottom of the coiled-coil. The second helix is predicted A single proline residue (568 in p85a) which occurs on the to end at residue 588, an e position on the second helix, and outside of the second predicted a-helix should be tolerated it may be that residues C-terminal to the second helix (Blundell et al., 1983).T Sequence comparison with the (residues 589-618) pack against the AEB face at the bottom crystal structure of tropomyosin (Philips et al., 1979, 1986) of the coiled-coil. Possibly some of the seven acidic residues the coiled-coil region of p85 is expected between 604 and 615 could interact with the exposed basic suggests anti-parallel to have a length of -110 A and contains three-quarters of residues Arg577 (a), Arg574 (e) and Lys567 (e), some a turn of coiled-coil (see Figure IOA). A large number of 20-30 A from the bottom of the coiled-coil on the BAE charged residues are present at positions e and g on the two face. These results provide a possible structural basis for a-helices, as is commonly found in a-helical coiled-coil consideration of the intersubunit binding and enzyme proteins (Conway and Parry, 1990) and the amino acid regulation. residues found at the seven positions in the heptad repeat are fairly typical of those found in two-stranded coiled-coils Discussion (see Table I and Conway and Parry, 1990). Six ion pairs are predicted to help in stabilizing the coiled-coil region, PI-3 kinase has two distinct subunits: p85, which can serve giving a favourable 0.6 ionic interactions per heptad repeat. to link the enzyme to activated protein tyrosine kinase 431 446 453 460 467 474 481 488 495 502 509 lst SH2 domain GABCDEFGABCDEFGABCDEFGABCDEFGABCDEFGABCDEFGABCDEFGABCDEFGABCDEFGABCDEFGA Sa p8 ..NESLAQYNPKLDVKLLYPVSKYQQDQWKEDNIEAVGKKLHEYNTQFQEKSREYDRLYEDYTRTSQEIQMKRTAIEAFNETIKIFEEQCQTQERYSKEYIEKFKREGN ..HESLAQYNAKLDTRLLYPVSKYQQDQIVKEDSVEAVGAQLKVYHQQYQDKSREYDQLYEEYTRTSQELQMKRTAIEAFNETIKIFEEQGQTQEKCSKEYLERFRREGN p850 518 525 532 539 546 553 560 567 574 581 588 607 620 2nd SH2 efgabcdefgabcdefgabcdefgabcdefgabcdefgabcdefgabcdefgabcdefgabcdefgabcde 85a ETE IQR IMHNYEKLKSR IS E IVDSRRRLEEDLKKQAAEYREI DKRMNS I KPDL IQLRKTRDQYLMWLTQKGVRQKKLNEWLGN -ENTEDQYSLVEDDEDLPHHDEKTWN 5 ILLNSERLKSRIAE DKRMNSLKPDLMQLRK IRDQYLVWLTQKGARQKK INEWLG IKNETEDQYSLMEDEDDLPHHEERTWY p8 EKEMQR IHESRTKLEQELRAQASDNREI 446 453 474 495 502 509 460 467 481 488 F K T R E Q T E E R E B E H D T E K Q K Q Q E G N S Y S R N E E I ,. D V Y K L T K F F y G N K Q E D E A T Q Y K C A E E R R M A I T E F K T R E T E R E Q E 588 581 574 567 560 553 546 539 532 525 518 f m k p k a e d s h t b t d i n r k e q rr s e 9 1 r k d a e k m e r y a 1 r 1 m y 1 i i d k y 1 i i q 1 i 1 i w t d r e d s r n e c q r k q q s e k e f m k p k a e d s h t Ee G 1 N r S k Y d S a N v k E m I e g B e F. Aa I 1 L r F 1 Y m Y I I I r I i C y S i F Dd k V y Y 1 i L i T q 1 F i F 1 i Y IS E N w K t e A s T r n Y Gg Q d E r2 E d Q . I of bovine and are shown in the between the two SH2 domains. The C-terminal of the Fig. 9. (A) The sequences p85ai p85,B aligned region boundary 431 of is defined from the NMR structure of the domain et Two first SH2 domain (residue p85a) (Booker al., 1992). long heptad repeats to two helices residues 441-512 and are indicated above the GABCDEFG... and corresponding the predicted (helix-1, helix-2, 518-588) sequences; in Note that there are some 30 residues between the Positions A/a and are in character these efgabcdefg... D/d predominantly hydrophobic repeats. amino acid of the two helices have been drawn on a helical end of the second helix and the start of the C-terminal SH2-domain. The (B) sequences are in The solid line indicates the line of that for the has been drawn so that the two helices register. net. Note helix-2, sequence backwards, at and of the two helices are to interlock. A of anti- intersection of the two helices. Hydrophobic residues positions A/a D/d expected (C) pair two helices are residues at A/a and D/d are Note the from positions hydrophobic. parallel ca-helices interacting. Heptad repeats anti-parallel shown, register (height) of the residues in the two helices: A B C D E F GA d c bag f ed Residues on the two in that are to interact on how helices interact to form ca-helices 1953 for further details coiled-coils). (D) p85ca expected (see Crick, in while those from helix-2 are in lower case. Residues which could form ion between Residues from helix-i are pairs are shown case, superposed. upper the two helices are underlined. 517 R.Dhand et al. a form of p85 (Otsu et al., 1991; Hiles et al., 1992). The ce and the forms of are ( p85 highly conserved in all but the BCR and both region proteins have been used here in complementary studies to analyse the basis of the p85 -p1 10 i _ R~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ i~~~~~~~~~~~~~ interaction. We show that both p85a and can form p850 stable active complexes with plO. Indeed, PI 3-kinase activity specific to the p85( subunit has been identified in 478 492 441 512 several cell lines (Gout et al., 1992). In addition, T lymphocytes have been shown to contain both p85e and p853 proteins associated with a PI 3-kinase activity, which are differentially regulated during T cell activation (Reif et al., iwl 1993). To examine whether the same specific region of p85a and =40~~~~~4 interacts with we p1 10, expressed domains of p85 as p850 GST fusion proteins. The results clearly show that the inter- SH2 region of p85 is required for the association with p110 r,- and blocking of the inter-SH2 region on the intact p85 protein .1,3. with monoclonal antibodies that have epitopes in this region prevents p110 binding in vitro. These data preclude SH2 4s4 wi domains as binding PI 3-kinase activity in vivo, as shown by Cooper and Kashishian (1993). Indeed, monoclonal antibodies that have epitopes located in the N-terminal SH2 domain of do not inhibit the of 10 to p85a binding pl the p85 subunits (R.Dhand, unpublished observation). A more detailed analysis of the inter-SH2 region of p85a !-W sizes of this domain defined 104 amino using varying acids, residues 451-555, which are able to bind the directly pl 10 CIS protein and PI 3-kinase Deletion mutants of the activity. p85 of Numbers proteins identified a short stretch of - 35 amino acids in Fig. 10. (A) Schematic diagram of the coiled-coil domain both p85. residues are residues according to p85a numbering. The first helix (helix-1; p85 proteins, which includes residues 478 -513 in p85ae and residues 441-512) is depicted with light shading, the second, (helix-2, 445 -485 in p85( for 10 in vitro. necessary binding pl The that the deletion mutants in 518-588) is shaded more darkly. Note p8513 remains that some of the deletion mutants of possibility p85 to 452-476 -469 452 -492 correspond residues (445 p853 numbering), may be bound to an inactive form of 10 and thus are not and 493-523 p1 (445-485 p85,B numbering) (486-516 numbering). p850 The deletion 452-492 abolishes and (B) binding may completely disrupt detected in these low level of assays, however, expression the structure. The deletion 493 -523 will remove the lower of (C) quarter the p85 deletion mutants does not detection of permit the the coiled-coil The deletion 452-476 will remove the half of the (D). top associated p 10 protein by other means. of the Expression coiled-coil. full-length p85cA478-513 in mouse L cells showed that it was also unable to associate with PI 3-kinase activity, in et Otsu et receptors and other proteins (Gout contrast to wild-type p85ao, thus that the deleted al., 1992; al., verifying 1991), and p1 10, which has an intrinsic region of p85ae defined in this study is the binding site for catalytic activity of and (Hiles et al., 1992). The regulation the p 10 protein in vivo. In the of the enzyme activity addition, ability full which be involved the complex protein interactions in length p85a and this deletion mutant p85aA478-513 to bind may to the of the function of PI 3-kinase seem PI 3-kinase activity and the insulin receptor substrate IRS-I require presence both subunits. The most obvious of intersubunit example in vivo has been examined using stable cell lines regulation is the of ability phosphotyrosine-containing overexpressing these constructs and the insulin receptor proteins or to cause activation of the 3-kinase peptides PI (Hara et al., in preparation). Insulin stimulation of cells et et activity (Backer al., 1992; Carpenter al., 1993; expressing wild-type p85a showed that it associated with Giorgetti et al., 1993). The basis for activation could lie in IRS-I and PI 3-kinase activity, moreover an accumulation phosphopeptide-induced conformational changes which have in the levels of PI(3,4)P2 and PI(3,4,5)P3 was observed. been seen in the N-terminal SH2 domain of and the However, insulin of p85a stimulation cells expressing p85aA478- p85a protein itself (Panayotou et al., Shoelson et 513 showed that although p85a was 1992; al., still able to associate the structural 10 1993), although implications of this on p1 with IRS-I it was unable to bind PI 3-kinase activity and remain unknown. Here we seek to determine for the basis significant reduction in the levels of PI(3,4)P2 and the intersubunit interactions so that we can understand the PI(3,4,5)P3 was seen. These data would suggest that these function of this enzyme both in vitro and in vivo. constructs may be used in further studies as dominant- PI 3-kinase may exist in a number of distinct isoforms in negative mutants in the analysis of mitogenic signalling mammalian cells. It is clear that the p85 proteins are encoded pathways. The of mapping the p85 binding site on pl 10 has by a gene family of at least three members, a, and 'y (Otsu shown that amino acids 20-108 are capable of binding to et al., 1991; Volinia et al., 1992), and in situ hybridization the p85 proteins and no other expressed fragments and PCR analysis confirm the presence of five closely related (plO.2-plO.6) show any detectable binding to either sequences of human pl 10 cDNA (S.Violinia, unpublished subunit. Thus the inter-SH2 region of p85 and the N-terminal observation). The only fully characterized PI 3-kinase has region of p1 1O, as independently folded modules, are been from the bovine purified brain, sequenced at the protein responsible for mediating the interaction of the two subunits cloned at DNA level, the level and is found to contain the of PI 3-kinase. However, it cannot be ignored that other 518 interactions PI 3-kinase intersubunit Table I. Percentage occurrence of each of the 20 amino acids at each position in the heptad repeats a(21) b(20) (%) c(20) (%) d(20) (%) e(21) (%) f(20) (%) g(21) (%) Ala(5) 0 0 10 0 4.8 5 4.8 Ile(13) 28.6 5 5 20 4.8 0 0 Leu(9) 19.0 0 0 20 4.8 0 0 Met(4) 4.8 0 5 0 4.8 5 0 Val(2) 0 0 0 5 4.8 0 0 9.5 0 0 10 0 0 0 Phe(4) 19.0 0 0 15 4.8 0 4.8 Tyr(9) 9.5 10 20 0 9.5 10 9.5 Arg(14) 0 15 10 15 9.5 15 9.5 Lys(15) 0 10 0 0 4.8 5 14.3 Asp(7) 0 15 25 0 19 25 19 Glu(21) 0 5 0 0 9.5 0 9.5 Asn(5) Cys(l) 4.8 0 0 0 0 0 0 Gln(12) 0 20 15 10 0 5 9.5 0 0 0 0 9.5 0 0 Gly(2) 0 5 0 0 0 5 0 His(2) 0 0 0 0 0 5 0 Pro(l) 4.8 5 5 0 9.5 5 4.8 Ser(7) 0 10 5 5 0 15 9.5 Thr(9) 0 0 0 0 0 4.8 Trp(l) 0 TOTALS(143) 100 100 100 100 100 100 100 5 20 70 28.8 10 9.5 Apolar(46) 80.9 9.5 25 30 15 19 25 19 Basic(29) 0 25 25 0 23.8 30 33.3 Acidic(28) of Ile, Leu, Met, Val, Phe and Tyr), basic and acidic residues occurring at each of the seven positions in Also shown are the percentages apolar (Ala, shows the two helices in the inter-SH2 domain of p85 have many features in common with two-stranded a-fibrous the heptad repeat. The Table that predicted Note in that at positions a and d of the heptad repeat 75.6% of residues are apolar. Forty-nine proteins (see Tables in Conway and Parry, 1990). particular of the helices are whereas only 12% of residues at positions a and d are charged. percent of the amino acids at the external positions (bcefg) charged, at a and d compared with other coiled-coil proteins (Conway and Parry, 1990) might be The relatively high proportions of Ile, Phe, and Tyr positions are predicted to be anti-parallel, whereas in other coiled-coil proteins helices are parallel. related to the prediction that in p85 the helices and the regions of both the p85 and the p110 proteins may also play a loss of function observed complete (Figure lOB) to some role in the interaction between the subunits but have 10 is seen not to bind p85. Sequence analysis p protein that residues 478-492 of been overlooked in this study due to the inability of a of both the reveals p85 proteins 470-497 domain to fold correctly. are contained in four particular p85a heptad repeats (residues the two at the To search for a structural basis for the region of p85 which in which are identical between proteins p85CY) conservative amino acid binds 10 the structural studies of Panayotou et al. amino acid level for a pl predicted except single to residues were extended. Sequence analysis suggests the inter- A (1992) change (Figure 9A). peptide corresponding and SH2 to be an coiled-coil of two a-helices 470-497 of was region anti-parallel p85az synthesized chemically coupled it was unable to bind to 10 which we have termed helix-i and helix-2 (Figure IOA). to Sepharose beads, however, p1 Evidence the concept that this region may be a and 3-kinase supporting PI data). activity (R.Dhand, unpublished characteristics of the domain is seen various size GST fusion of the discrete using Analysis p1O binding p85 that while helix-i of the inter-SH2 It is demonstrated that as deletion mutants proteins of the inter-SH2 region. suggests for the interaction between boundaries of the inter-SH2 are decreased the is the region region primarily responsible of 3-kinase helix-2 amount of PI 3-kinase seen to bind is also decreased. the two subunits the activity PI complex, probably element to the of helix- This could be accounted for if disrupting the inter-SH2 donates a structural 1. Thus, folding in to a stable structure is the and thus order destabilizing putative domain, structure, regions corresponding produce helix-i of and the on interactions. When deletions in to 470-497 on preventing protein -protein region p85a adjacent interaction between the are for the the and those in the are compared helix-2 p85a protein p85(3 protein required probably of the and the subunit to take and considered with the loss activity, corresponding place. p85 p110 and the 10 can between the third of helix-i of the residues 478-492 The interaction quarter coiled-coil, p85 pl proteins in that the two between the two both in vivo and in is identified as the common region take p85a, vitro, place suggesting be cotranslated in order to associate. mutants that are unable do not need to that is absent in deletion p85 proteins proteins the subunit and the 10 subunit Deletions which The between to bind PI 3-kinase IOA). activity (Figure binding p85 p1 and no suitable conditions have been half of residues 452 -476 on is of remove either the top helix-i, very high affinity, of an active which result in dissociation PI or the lower of residues established (Figure lOD) quarter helix-i, p85a Ruiz-Larrea et et al., destabilize the coiled 3-kinase 492-523 on p85a (Figure lOC) al., 1992; complex (Fry known that the interaction between In it is and thus the association with the structure disrupt plO addition, 1993). not on the of subunits is but do not inhibit its when the two protein completely binding. Only presence dependent and which are contained in residues 478 -492 are observation) the removed, (R.Dhand, unpublished phosphotyrosine motifs in this of 10 residues 452 -492 in is there are no mutation which deletes p1 the p850, region proline-rich present 519 et al. R.Dhand which may provide a binding consensus for the SH3 domain Materials and methods of the p85 protein. Secondary structure prediction using the Materials Leeds Prediction Package (Eliopoulos, 1989), predicts that Restriction enzymes and DNA modification enzymes were obtained from some 60% of the N-terminal 120 residues, which have been standard commercial sources and used according to the manufacturers recommendations. Oligonucleotides were synthesized on an Applied shown to direcfly bind the p85 proteins, will adopt an a- Biosystems 380B DNA synthesizer and used directly in all subsequent helical conformation and some 20% a ,3-sheet conformation. procedures. Thus, the N-terminal region of pl 10 will consist of a cu/l Cells and viruses region, where the c-helices might form interactions with the Maintenance of insect cell (Sf9) culture was carried out as described by inter-SH2 region of p85. Since the major binding site on Summers and Smith (1987). Both p85a and p85,B proteins were expressed p85a is considered to be between residues 478 -492 on in insect cells using baculovirus vectors as described previously by Otsu helix-1, it is of interest that there is a small hydrophobic (1991) and Gout et al. (1992). Recombinant baculovirus expressing et al. 110 was described in Hiles et al. (1992). Mouse-L cells were maintained in this region formed by Met479 (C position on pocket in DMEM medium supplemented with 10% fetal calf serum. Figure 9C), Ala483 (G position on Figure 9C) and Ala486 (C position on Figure 9C). As the N-terminus of p1O is Recombinant DNA methods of GST-p85a fusion proteins. Sequences corresponding to - 48% hydrophobic in character, this pocket may be Construction the SH3 domain (amino acids 1-86), the BCR homology region (amino in forming a high affinity interaction with the p 10 important acids 125 -322) and the helical inter-SH2 domain (amino acids 425 -600, protein. 425-555, 451-616, 451-600, 451-555) of the bovine p85a subunit of is intriguing that a lipid binding site has also been It the PI 3-kinase (Otsu et al., 1991), were amplified by PCR and cloned into vector (Pharmacia). BamHI and EcoRI sites were identified in the inter-SH2 region of the p85 protein. Two the pGEX-2T expression the C-termini of oligonucleotides respectively to included at the N- and antibodies that specifically recognize this region monoclonal facilitate cloning. Stop codons were introduced at the end of each of the been shown to inhibit the binding of phospholipids, in have cloned sequences, 5' to the EcoRI site. PCR fragments were verified using to the p85 subunit (End et al., 1993). particular PI(4,5)P2 system (US Biochemicals). The full-length p85a (amino acids the Sequenase sequence analysis reveals the presence of a short Amino acid 1-724) was subcloned into pGEX-2T by digesting the pGEX-2T-SH3 with and EcoRI and replacing this fragment with a similarly cut cartridge the inter-SH2 region of the p85cx protein, Bgll basic motif in from p85a in the Bluescript vector (Otsu et al., 1991) which contained the found in profillin and gelsolin and similar to that remaining coding region of the p85ai protein. The C- and N- terminal SH2 to confer PI(4,5)P2 binding properties to these demonstrated of p85a and the p85aN-C construct have been described (Yonezawa domains et al., 1992; Yu et al., 1992). Moreover, proteins (Janmey 1992). p85aN-CA478-514 was constructed by the amplification of et al., residues are well-conserved in p85,B, which PCR products (P1 and P2) from the p85a cDNA. The first (P1) analogous two nucleotides 988-1434 and had a BamHI site introduced at encompassed that it might be important as a regulatory site for suggests the 5' end and a XbaI site at the 3' end. The second (P2) encompassed The function of this lipid binding site is these proteins. nucleotides 1540-2175 and had a XbaI site introduced at the 5' end and unknown, but there are several possible presently both EcoRI and KpnI sites at the 3' end. Both fragments, that is the As the p85 protein has been shown to bind pl 10 implications. BamHI-XbaI fragment of P1 and the XbaI-EcoRI fragment of P2, were ligated into the BamHI-EcoRI site of the pGEX-2T plasmid and sequenced. via the inter-SH2 region, the lipid binding site may form The expression of this mutant results in a mutant containing amino acids of the substrate binding pocket for PI 3-kinase, specific a part with the deletion of 35 amino acids from residues 330-724 of p85a Indeed, at the membrane where PI is the for PI(4,5)P2. 478-513 with the insertion of two amino acids, serine and arginine in the substrate, this region may be important in most abundant deleted region. pGEX-2T constructs were transferred into E.coli XLl-blue of PI(4,5)P2, thus increasing the affinity for and expression of GST fusion proteins was carried out as the selection (Stratagene) described previously (Smith and Johnson, 1988). Evidence from a number of cell lines this phospholipid. that PI(3,4,5)P3 is the key signalling lipid suggests of GST-p85i3 fusion proteins. Full-length p850, Construction by the PI 3-kinase (Stephens et al., 1991; Hawkins generated p85,B(486-724) and were described in p85,B(445-724), p850(516-724) Yonezawa et al. (1992). For p8513A486-516 GST-p8513(1-724) was 1992), thus implicating PI(4,5)P2 as the critical et al., with BamHI and BgEl to remove the fragment residues 486-5 16 digested substrate required for activation of the PI 3-kinase. It could and then self-ligated. p85I3A445-485 and -469 were constructed p85OA445 be that the conformational change induced in p85 speculated by amplifying two PCR products from p8513 cDNA (P3 for p8513A445 -485 upon peptide binding may cause a sterically masked and P4 for p8513A445-469). P3 encompassed nucleotides 1457-1966 and PI(4,5)P2 lipid binding site to become more accessible to P4 encompassed nucleotides 1408-1966, and both had an Accl introduced at the 5' end. Both P3 and P4 were digested with AccI and KpnI and then the substrate binding pocket of PI 3-kinase, thus resulting ligated into AccI and KpnI digested GST-p85,1-724 respectively. activation. The substrate specificity of purified in its 10 and comparison with purified p85a -p1 10 monomeric pl GST-pJJO fusion proteins. Regions of the pl O cDNA Construction of al., 1992) corresponding to amino acids 1-128 (pl 110-), 123-458 and mutagenesis of critical residues in the lipid (Hiles et complex (p110-3), 601-960 (p110-4), 760-960 (p110-5), (p110-2), 577-1068 binding site of p85 will be required to resolve this issue. (pl11-6), 1-49 (110-1.1), 1-81 (p110-1.2), 1-108 (p110-1.3), 760-1069 This study has localized the regions of p85a and p110 20-81 (pl 10-1.4) and 20-108 (p1 10-1.5), were amplified by PCR. BamHI involved in interacting with each other. This will permit the and EcoRI sites were added to 5' and 3' oligonucleotides respectively to of dominant negative mutants that may define a more use facilitate cloning. Following amplification, PCR products were cut with BamHI and EcoRI and ligated into pGEX-2T (Phamacia), GST fusion role for PI 3-kinase in mitogenic signalling pathways. precise constructs expressing amino acids 1-35 (p1 10-1.6) and 37-128 (p1 10-1.7) Blocking the association between the p85 and the p110 were constructed from (p 110-1) as follows. For plasmid p1 10-1.6, p 110-1 subunits may be a potential means for inhibiting signal was digested with BsmI and EcoRI, ends were made blunt by treatment transduction. The identification of the binding sites reported with T4 DNA polymerase and the plasmid was circularized with T4 DNA in this will aid the development of assays for inhibitors paper ligase. For plasmid p1 10-1.7, p 110-1 was digested with BsmI and the ends with T4 DNA polymerase. BgM linkers (12mers; New England made blunt of this interaction. Finally, a more detailed understanding Biolabs) were added to the ends and the plasmid further restricted with of the association between the two subunits of PI 3-kinase BamHI. The plasmid containing band was gel-purified and the plasmid will require solution and crystallographic studies of the two were transformed with T4 DNA ligase. Expression constructs recircularized fragments described and subsequent mutagenesis of critical into XL1-blue (Stratagene). Expression of GST fusion proteins was E.coli carried out as described previously by Smith and Johnson (1988). residues. 520 3-kinase intersubunit interactions Pi Downing,A.K., Panayotou,G., (Uout,l., Booker,G.W., Breeze,A.L., Antibodies and immunoprecipitation and Campbell,I.D. (1992) Nature, 358, 684-687. Waterfield,M.D. The pl10 C-terminal antiserum was raised as described by Hiles et al. Auger,K.R., Cohen,B., Carpenter,C.L., Duckworth,B.C., (1992). Affinity purified polyclonal antiserum raised against SDS-PAGE Schaffhausen,B.S. and Cantley,L.C. (1990) J. Biol. Chem., 265, purified p85a was described by Gout et al. (1992). Monoclonal antibodies 19704-19711. directed to the p85a and to the proteins are described in End et al. p85,0 Auger,K.R., Chanudhuri,M., Yoakim,M., (1993). Immunoprecipitations were carried out as described previously by Carpenter,C.L., Schaffhausen,B., Shoelson,S. and Cantley,L.C. (1993) J. Biol. Chem., Otsu et al. (1991). 268, 9478-9483. Pi 3-kinase assay Conway,J.F. and Parry,D.A.D. (1990) Int. J. Biol. Macromol., 12, assay for PI 3-kinase activity was carried out as described by End et al. The 328-334. (1993). and Kashishian,A. (1993) Mol. Cell. Biol., 13, 1737-1745. Cooper,J.A. (1953) Acta Crystallogr., 6, 689-697. Crick,F.H.C. Association assays Eliopoulos,E.G. (1989) Documentation for Leeds Prediction Programs. ofp110 with p85a and p85(3 in vivo. Sf9 cells were infected Association of Biophysics, University of Leeds. Department baculoviruses as described by Gout et al. (1992) and Hiles with appropriate Gout,I., Fry,M.J., Panayotou,G., Dhand,R., Yonezawa,K., End,P., Two days post-infection, cells were harvested and lysed in et al. (1992). and Waterfield,M.D. (1993) J. Biol. Chem., 268, Kasuga,M. mM pH 7.4, 150 mM NaCl, 50 mM NaF, 5 mM EDTA, 50 Tris-HCI 10066-10075. X-100, 500 sodium orthovanadate, 2 mM PMSF and 100 1% Triton uM Escobedo,J.A., Martin,G.A., Turck,C.W., Delrosario,M., Fantl,W.J., inhibitor units of aprotinin. Lysates were then either kallikrein and Williams,L.T. (1992) Cell, 69, 413-423. McCormick,F. using appropriate antibodies or were applied to a immunoprecipitated Fry,M.J. (1992) Curr. Biol., 2, 78-80. affinity column, DY751VPML(G), derived from sequence phosphopeptide Panayotou,G., Dhand,R., Ruiz-Larrea,F., Gout,I., Nguyen,O., Fry,M.J., Tyr751 of the human PDGF (3-receptor to which PI 3-kinase surrounding and Waterfield,M.D. (1992) Biochem. J., 288, Courtneidge,S.A. to bind (Otsu et al., 1991; Fry et al., 1992). Affinity has been shown 383-393. columns were washed as described. Enzyme complexes were analysed by Kowalskichauvel,A., Tartare,S. and Van Giorgetti,S., Ballotti,R., 7.5% SDS-PAGE gels or subjected to PI 3-kinase assays as described (1993) J. Biol. Chem., 268, 7358-7364. Obberghen,E. above. Panayotou,G., Fry,M.J., Hiles,I., Otsu,M. and Gout,I., Dhand,R., Waterfield,M.D. (1992) Biochem. J., 288, 395-405. Association ofPI 3-kinase activity with wild-type and mutant p85ac in vivo. Hawkins,P.T., Jackson,T.R. and Stephens,L.R. (1992) Nature, 358, SRa plasmids (Takebe et al., 1988) for the expression of wild-type bovine 157-159. (Wp85a) or a mutant bovine fiull-length 85a (p85aA478-513), which p85a Herman,P.K. and Emr,S.D. (1990) Mol. Cell. Biol., 10, 6742-6754. lacks a binding site for p 10 were constructed as described (Hara et al., Hiles,I.D. et al. (1992) Cell, 70, 419-429. in preparation) and designated as SRa-Wp85a and SRa-p85aA478-513, Hodges,R.S., Sodek,J., Smillie,L.B. and Jurasek,L. (1972) Cold Spring respectively. Semi-confluent mouse L cells were transfected with 3 of Ag Harbor Symp. Quant. Biol., 37, 299-310. either SRa-Wp85a or SRci-p85aA478-513 plasmids using the DEAE Jamney,P.A., Lamb,J., Allen,P.G. and Matsudaira,P.T. (1992) J. Biol. dextran method. After 72 h cells were lysed in a buffer containing 20 mM Chem., 267, 11818-11823. Tris pH 7.6, 1% Nonidet P-40, 10% glycerol, 137 mM NaCl, 1 mM Kunz,J., Henriquez,R., Schneider,U., Deuterreinhard,M., Movva,N.R. and MgCl2, 1 mM CaCl2, 1 mM DTT, 1 mM PMSF and 1 mM sodium Hall,M.N. (1993) Cell, 73, 585-596. and transiently expressed bovine p85s immunoprecipitated orthovanadate, Ling,L.E., Druker,B.J., Cantley,L.C. and Roberts,T.M. (1992) J. Virol., with 2 of a anti-p85a monoclonal antibody (G12) (Yonezawa et al., 1992) Ag 66, 1702-1708. to G -agarose. These samples were then subjected to Western bound protein Mclachlan,A.D. and Stewart,M. (1975) J. Mol. Biol., 98, 293-304. blotting analysis or PI 3-kinase assay. Morgan,S.J., Smith,A.D. and Parker,P.J. (1990) Eur. J. Biochem., 191, 761 -767. association In vitro complex formation of the p85a or the In vitro assay. Otsu,M. et al. (1991) Cell, 65, 91-104. with 110 was examined by immunoprecipitating p85 from p85(3 subunit p (1975) J. Mol. Biol., 98, 519-535. Parry,D.A.D. antibodies. The immunoprecipitate was collected insect cells using appropriate and Waterfield,M.D. (1992) Trends Cell Biol., 2,358-360. Panayotou,G. A beads and washed three times in lysis buffer. This on protein -Sepharose Bax,B., Gout,I., Federwisch,M., Wroblowski,B., Dhand,R., Panayotou,G., was then incubated for 1 h at 40C with cell lysates immune complex preMared Fry,M.J., Blundell,T.L., Wollmer,A. and Waterfield,M.D. (1992) EMBO infected with viruses expressing p1 10. The resultant enzyme from Sf9 cells 4261-4272. J., 11, were washed extensively and either analysed on 7.5 % complexes Parker,P. and Waterfield,M.D. (1992) Cell Growth Differ., 3, 747-752. or for associated PI 3-kinase activity. SDS-PAGE gels assayed directly and Gish,G.D. (1992) Cell, 71, 359-362. Pawson,T. of 10 to domains was performed in essentially the In vitro binding pl p85a Pawson,T. and Schlessinger.J. (1993) Curr. Biol., 3, 434-442. GST fusion proteins were immobilized on same manner. Lattman,E.E., Cummins,P., Lee,K.Y. and Cohen,C. Philips,G.N.,Jr, beads as described previously (Smith glutathione-Sepharose (Pharmacia) 278, 413-417. (1979) Nature, These columns were then incubated for 1 h and Johnson, 1988). affmity Fillers,J.P. and Cohen,C. (1986) J. Mol. Biol., 192, Philips,G.N.,Jr, from Sf9 cells infected with a baculovirus at with cell lysates prepared 4°C 111-131. 10. After extensive washing with lysis buffer, bound proteins expressing p1 Gout,I., Waterfield,M.D. and Cantrell,D.A. (1993) J. Biol. Chem., Reif,K., kinase For association of the GST-pl 10 constructs were subjected to assays. 10780-10788. 268, cell of Sf9 cells infected with baculoviruses with p85a and p85(3, lysates et al. (1993) Biochem. J., 290, 609-616. Ruiz-Larrea,F. or a control were added to either p85a, p85(3 wild-type expressing Fry,M.J., Stack,J.H., Waterfield,M.D. and Schu,P.V., Takegawa,K., 10 fusion and was allowed to proceed immobilized GST-pl proteins, binding (1993) Science, 260, 88-91. Emr,S.D. The beads were then washed four times with lysis buffer for 2 h at 4°C. Homma,Y. and Takenawa,T. (1991) J. Biol. Chem., 266, Shibasaki,F., in 100 ml 2 x Laenmli buffer. Samples were boiled and sample resuspended 8108-8114. 5 resolved SDS-PAGE on 10% gels. for min and proteins by Shoelson,S.E., Sivaraja,M., Williams,K.P., Hu,P., Schlessinger,J. and EMBO 12, 795-802. Weiss,M.A. (1993) J., Acknowledgements Smith,D.B. and Johnson,K.S. (1988) Gene, 67, 31-40. and (1991) Nature, 351, 33-39. Stephens,L.R., Hughes,K.T. Irvine,R.F. thank Oahn for peptide synthesis and Alistair We would like to Truong and Tex. Exp. Stat. Bull., 1555, 1-56. Summers,M.D. Smith,G.E. (1987) for Sterling oligonucleotide synthesis. Hoy,P., Yokota,K., Arai,K., Yoshida,M. Takebe,Y., Seiki,M., Fujisawa,J., Mol. Cell. Biol., 8, 466-472. and Arai,N. (1988) and Kazlauskas,A. (1993) Cell, 73, 321-334. Valius,M. References Gout,I., Calzolari,E., Patracchini,P., Otsu,M., Hiles,I., Volinia,S., and (1992) Oncogene, 7, EMBO 3469-3479. Rooke,L. Waterfield,M.D. et al. J., 11, Bernardi,F., Backer,J.M. (1992) Brice,M.D., 789-793. Williams,G.J.B., Meyer,E.F.Jr, Bernstein,F.C., Koetzle,T.F., et al. J. Biol. 25958-25965. Shimanouchi,T. and Tasumi,M. (1977) J. Chem., 267, Kennard,O., Rodgers,J.R., Yonezawa,K. (1992) and J. Biol. Chem. Mol. 535 -542. Yin,H.L. (1992) Biol., 112, Yu,F.X., Sun,H.-Q., Janrney,P.A. 14616-14621. and Thorton,J. (1983) Nature, 306, Borkakoti,N. 267, Blundell,T.L., Barlow,D., revised on October 1993 on 7, Received 25, 1993; 281-283. August

Journal

The EMBO Journal – Springer Journals

Published: Feb 1, 1994

There are no references for this article.