Mutations in the Carboxyl-terminal Domain of Phospholipase C-β1 Delineate the Dimer Interface and a Potential GαqInteraction Site
2002; Elsevier BV; Volume: 277; Issue: 6 Linguagem: Inglês
10.1074/jbc.m109612200
ISSN1083-351X
AutoresOlga Ilkaeva, Lisa N. Kinch, Ruth H. Paulssen, Elliott M. Ross,
Tópico(s)Cell death mechanisms and regulation
ResumoThe carboxyl-terminal domain of phospholipase C-β is required for its stimulation by Gαq and for its Gαq-specific GTPase-activating protein (GAP) activity. We subjected this domain to a combination of deletion and alanine/glycine scanning mutagenesis to detect mutations that would inhibit either responsiveness to Gq or Gq GAP activity. Most mutations that altered either response or GAP activity diminished both in parallel. Many of these mutations map at the interface at which the carboxyl-terminal domain was recently shown to form a dimer (Singer, A. U., et al. (2001) Nat. Struct. Biol., 9, 32–36). Most others clustered in an area that is a plausible Gαq binding site. In addition, one mutation that differentially inhibited GAP activity relative to responsiveness to Gαq mapped in this region at a location modeled to be in close contact with the switch II region of Gαq. This is the site at which RGS proteins are thought to exert their GAP activity. Last, a deletion mutation differentially inhibited the response of phospholipase C-β1 to Gαqwithout blocking GAP activity. Its location in the molecule suggests that moving the attachment point of the catalytic domain can disrupt its ability to be activated by Gαq. The carboxyl-terminal domain of phospholipase C-β is required for its stimulation by Gαq and for its Gαq-specific GTPase-activating protein (GAP) activity. We subjected this domain to a combination of deletion and alanine/glycine scanning mutagenesis to detect mutations that would inhibit either responsiveness to Gq or Gq GAP activity. Most mutations that altered either response or GAP activity diminished both in parallel. Many of these mutations map at the interface at which the carboxyl-terminal domain was recently shown to form a dimer (Singer, A. U., et al. (2001) Nat. Struct. Biol., 9, 32–36). Most others clustered in an area that is a plausible Gαq binding site. In addition, one mutation that differentially inhibited GAP activity relative to responsiveness to Gαq mapped in this region at a location modeled to be in close contact with the switch II region of Gαq. This is the site at which RGS proteins are thought to exert their GAP activity. Last, a deletion mutation differentially inhibited the response of phospholipase C-β1 to Gαqwithout blocking GAP activity. Its location in the molecule suggests that moving the attachment point of the catalytic domain can disrupt its ability to be activated by Gαq. phospholipase C guanosine 5′-O-(thiotriphosphate) GTPase-activating protein phosphatidylinositol 4,5-bisphosphate Phospholipase C (PLC)1-βs are metazoan PIP2-selective PLCs whose activity is stimulated by heterotrimeric G proteins. They are structurally distinguished from the G protein-unresponsive PLCs γ, δ, and ε by a distinctive sequence amino-terminal to the catalytic domain and by a ∼40-kDa carboxyl-terminal extension (1Rhee S.G. Choi K.D. Putney Jr., J.W. Advances in Second Messenger and Phosphoprotein Research. Raven Press Ltd., New York1992: 35-61Google Scholar, 2Singer W.D. Brown H.A. Sternweis P.C. Annu. Rev. Biochem. 1997; 66: 475-509Crossref PubMed Scopus (353) Google Scholar, 3Rebecchi M. Pentyala S. Physiol. Rev. 2000; 80: 1291-1335Crossref PubMed Scopus (840) Google Scholar). PLC-β enzymes are stimulated both by GTP-bound Gα subunits of the Gq family (Gαq, Gα11, Gα14/15, and Gα16) and by Gβγ subunits, although relative sensitivity to Gαq and Gβγ varies among the four mammalian members of the PLC-β family (2Singer W.D. Brown H.A. Sternweis P.C. Annu. Rev. Biochem. 1997; 66: 475-509Crossref PubMed Scopus (353) Google Scholar, 3Rebecchi M. Pentyala S. Physiol. Rev. 2000; 80: 1291-1335Crossref PubMed Scopus (840) Google Scholar). Stimulation of phospholipase activity by G proteins can exceed 100-fold in vitro, although both activity and relative stimulation are strongly influenced by the lipid interface at which the PIP2 substrate is found (2Singer W.D. Brown H.A. Sternweis P.C. Annu. Rev. Biochem. 1997; 66: 475-509Crossref PubMed Scopus (353) Google Scholar). In addition to their role as Gq-regulated effector enzymes, PLC-βs are also GTPase-activating proteins (GAPs) specific for members of the Gαq family (4Berstein G. Blank J.L. Jhon D.-Y. Exton J.H. Rhee S.G. Ross E.M. Cell. 1992; 70: 411-418Abstract Full Text PDF PubMed Scopus (364) Google Scholar). PLC-β1 can increase the rate of hydrolysis of GTP bound to Gq over 1000-fold (5Mukhopadhyay S. Ross E.M. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 9539-9544Crossref PubMed Scopus (156) Google Scholar). The GAP activity of PLC-βs is thought to provide enhanced temporal responsiveness to signaling processes mediated by Gq and PLC-β and to minimize spontaneous activation (noise) in these pathways (6Ross E.M. Recent Prog. Horm. Res. 1995; 50: 207-221PubMed Google Scholar, 7Ross E.M. Wilkie T.M. Annu. Rev. Biochem. 2000; 69: 795-827Crossref PubMed Scopus (950) Google Scholar). The carboxyl-terminal domain of PLC-β, excluding the last ∼20 amino acid residues, is required for the enzyme to respond to activated Gαq (8Wu D. Jiang H. Katz A. Simon M.I. J. Biol. Chem. 1993; 268: 3704-3709Abstract Full Text PDF PubMed Google Scholar, 9Park D. Jhon D.-Y. Lee C.-W. Ryu S.H. Rhee S.G. J. Biol. Chem. 1993; 268: 3710-3714Abstract Full Text PDF PubMed Google Scholar). Fragments of this domain, some as small as 15 kDa, display Gq GAP activity independent of the PLC-β catalytic and amino-terminal domains (10Paulssen R.H. Woodson J. Liu Z. Ross E.M. J. Biol. Chem. 1996; 271: 26622-26629Abstract Full Text Full Text PDF PubMed Scopus (58) Google Scholar), and several mutations in this region inhibit response to Gαq (11Kim C.G. Park D. Rhee S.G. J. Biol. Chem. 1996; 271: 21187-21192Abstract Full Text Full Text PDF PubMed Scopus (194) Google Scholar). It is thus likely that the carboxyl-terminal region of PLC-β constitutes all or part of the Gαq binding site. This idea is supported by the observation that the carboxyl-terminal domain can inhibit stimulation of PLC-β by Gq either using purified proteins in solution (10Paulssen R.H. Woodson J. Liu Z. Ross E.M. J. Biol. Chem. 1996; 271: 26622-26629Abstract Full Text Full Text PDF PubMed Scopus (58) Google Scholar) or in cells (12Beekman A. Helfrich B. Bunn Jr., P.A. Heasley L.E. Cancer Res. 1998; 58: 910-913PubMed Google Scholar), presumably by binding and sequestering Gαq. In addition to displaying intrinsic GAP activity, the tails can potentiate and inhibit the GAP activity of PLC-β1 in a biphasic process and display biphasic stimulation/inhibition by themselves (10Paulssen R.H. Woodson J. Liu Z. Ross E.M. J. Biol. Chem. 1996; 271: 26622-26629Abstract Full Text Full Text PDF PubMed Scopus (58) Google Scholar). The complex and biphasic regulatory behavior of the tails, both alone and in combination with full-length PLC-β1, led us to suggest that the tails can oligomerize both with themselves and with full-length PLC-β1. The idea that PLC-β may function as a dimer also agrees with the concentration-dependent activity described for PLC-β2 (13Paterson A. Boyer J.L. Watts V.J. Morris A.J. Price E.M. Harden T.K. Cell. Signaling. 1995; 7: 709-720Crossref PubMed Scopus (23) Google Scholar) and with the observation of fluorescence resonance energy transfer between differently fluorophore-labeled PLC-β2 molecules. 2L. Runnels and S. Scarlata, personal communication. The structure of the carboxyl-terminal tail of turkey PLC-β2 was recently solved at the atomic level by Singer et al. (14Singer A.U. Waldo G.L. Harden T.K. Sondek J. Nature Struct. Biol. 2002; 9: 32-36Crossref PubMed Scopus (86) Google Scholar). The tail is a long, three-stranded coiled coil. Two tails bind to form an antiparallel dimer with its interface along the long axis of the coiled coils. The validity of the dimeric structure of the tails found in the protein crystal is supported by the functional data cited above. Prior to the solution of the tertiary structure of the PLC-β tail, we attempted to locate sites in the carboxyl-terminal region of PLC-β1 that, when mutated, would block its Gq GAP activity. We report here the location of one mutation that selectively inhibits GAP activity relative to responsiveness to stimulation by Gqand another that inhibits responsiveness with minimal loss of GAP activity. Most inactivating mutations, however, diminish both effects in parallel. Many of these mutations lie at the dimer interface. Others, along with the mutation that disproportionately inhibits GAP activity, point to a likely site of interaction with Gαq. Mutagenesis of the cDNA that encodes the carboxyl-terminal domain of rat PLC-β1 was performed on a 1130-bp AvaI cDNA fragment in pUC19. Most mutagenesis was performed using PCR, usually by the QuikChange protocol (Stratagene). After the sequences of the mutants were verified, the AvaI fragment was returned to the full-length cDNA in bacterial expression vector pQE60 (Qiagen), which was modified to encode a His6 tag and include aNotI site in the multiple cloning site. For expression in Sf9 cells, the cDNA was transferred as aNotI-BamHI fragment to pVL1392 (15Jenkinson S. Nahorski S.R. Challiss R.A.J. Mol. Pharmacol. 1994; 46: 1138-1148PubMed Google Scholar), and recombinant baculoviruses were prepared as described previously (16Parker E.M. Kameyama K. Higashijima T. Ross E.M. J. Biol. Chem. 1991; 266: 519-527Abstract Full Text PDF PubMed Google Scholar). The deletion mutant Δ1123–38 was prepared in the Tail32 construct (10Paulssen R.H. Woodson J. Liu Z. Ross E.M. J. Biol. Chem. 1996; 271: 26622-26629Abstract Full Text Full Text PDF PubMed Scopus (58) Google Scholar), and that fragment was inserted into the full-length cDNA in pVL1392. Some PLC-β1 constructs were expressed in Sf9 cells and purified as described previously (17Biddlecome G.H. Berstein G. Ross E.M. J. Biol. Chem. 1996; 271: 7999-8007Abstract Full Text Full Text PDF PubMed Scopus (213) Google Scholar). Gαq, Gβ1γ2, and m1 muscarinic acetylcholine receptor were expressed in Sf9 cells and purified as described previously (17Biddlecome G.H. Berstein G. Ross E.M. J. Biol. Chem. 1996; 271: 7999-8007Abstract Full Text Full Text PDF PubMed Scopus (213) Google Scholar). Phospholipid vesicles that contain Gq and muscarinic 1 acetylcholine receptor were reconstituted as described previously (17Biddlecome G.H. Berstein G. Ross E.M. J. Biol. Chem. 1996; 271: 7999-8007Abstract Full Text Full Text PDF PubMed Scopus (213) Google Scholar). PLC-β1 was routinely expressed in Escherichia coliSG13009/pREP4 as described previously (18Meij J.T.A. Ross E.M. Biochem. Biophys. Res. Commun. 1996; 225: 705-711Crossref PubMed Scopus (5) Google Scholar). For evaluation of the activities of mutant PLC-β1s, the protein was partially purified by one-step affinity chromatography on Ni2+-NTA-agarose (Qiagen). The high-salt membrane extract from 1 liter of cells (18Meij J.T.A. Ross E.M. Biochem. Biophys. Res. Commun. 1996; 225: 705-711Crossref PubMed Scopus (5) Google Scholar) was mixed with 0.5 ml of Ni2+-NTA-agarose (Qiagen) for 1 h at 4 °C. The slurry was poured into a column and washed with 5 ml of buffer Y (20 mm Tris-Cl (pH 7.5), 1 mmMgCl2, 0.2 mm phenylmethylsulfonyl fluoride, 20 μg/ml leupeptin, 10 μg/ml aprotinin, 10% glycerol, and 10 mm 2-mercaptoethanol), 5 ml of 0.3 m NaCl in buffer Y, and 10 ml of buffer Y plus 0.3 m NaCl and 10 mm imidazole. PLC-β1 was eluted in 1 ml of 0.1m NaCl plus 150 mm imidazole in buffer Y. PLC-β1 that has been partially purified from E. coli is only 2–4% pure. To normalize multiple preparations to a fixed concentration of active enzyme, we assayed PLC activity in the presence of 5 μm Ca2+, as described below. The specific activity of rat PLC-β1 in this assay is identical for highly purified preparations of enzyme from Sf9 cells or E. coli (18Meij J.T.A. Ross E.M. Biochem. Biophys. Res. Commun. 1996; 225: 705-711Crossref PubMed Scopus (5) Google Scholar), and total activity in an assay is proportional to the amount of enzyme added over the range of concentrations used here (data not shown). To determine the concentration of active PLC-β1, we divided its activity by the specific activity of the purified enzyme. This concentration was consistent with the total amount of immunoreactive PLC-β1 determined by Western blotting. PLC activity was determined by the method of Blank et al.(19Blank J.L. Ross A.H. Exton J.H. J. Biol. Chem. 1991; 266: 18206-18216Abstract Full Text PDF PubMed Google Scholar), which measures hydrolysis of [3H]PIP2added as mixed phospholipid vesicles. Reactions were initiated by the addition of enzyme, and the reaction was allowed to proceed at 30 °C for 10 min. Ca2+-stimulated activity was determined in the presence of 5 μm Ca2+, which was maintained with an EGTA buffer. Gαq-stimulated activity was measured in the presence of both 0.2 μm Ca2+ and Gαq that had been previously been activated by incubation with GTPγS (17Biddlecome G.H. Berstein G. Ross E.M. J. Biol. Chem. 1996; 271: 7999-8007Abstract Full Text Full Text PDF PubMed Scopus (213) Google Scholar). Routine Gαq-stimulated assays used for screening mutants contained 2 nm Gαq, which is approximately equal to the EC50 (median effective concentration) (20Berstein G. Blank J.L. Smrcka A.V. Higashijima T. Sternweis P.C. Exton J.H. Ross E.M. J. Biol. Chem. 1992; 267: 8081-8088Abstract Full Text PDF PubMed Google Scholar). Gq GAP activity was determined according to the increase in receptor-stimulated steady-state GTPase activity in phospholipid vesicles that contained trimeric Gαqβ1γ2 and muscarinic 1 acetylcholine receptor. Assays were performed as described previously (4Berstein G. Blank J.L. Jhon D.-Y. Exton J.H. Rhee S.G. Ross E.M. Cell. 1992; 70: 411-418Abstract Full Text PDF PubMed Scopus (364) Google Scholar, 21Wang J. Tu Y. Mukhopadhyay S. Chidiac P. Biddlecome G.H. Ross E.M. Manning D.R. G Proteins: Techniques of Analysis. CRC Press, Boca Raton, FL1999: 123-151Google Scholar). The structural model of the carboxyl-terminal domain of rat PLC-β1 was based on the atomic coordinates of the amino-terminal domain of turkey PLC-β2 (14Singer A.U. Waldo G.L. Harden T.K. Sondek J. Nature Struct. Biol. 2002; 9: 32-36Crossref PubMed Scopus (86) Google Scholar). This construct contains a designed deletion between residues 960 and 1001 (PLC-β1 residue numbers). The corresponding amino acid residues display 33% sequence identity and 52% similarity, with a 2-residue deletion and a 1-residue insertion (both in loops). The insertion, residues 947–950 in PLC-β1, is in a region that is poorly conserved among the PLC-βs. The deletion, 2 residues between PLC-β1 residues 1084 and 1085, is characteristic of PLC-β1; the sequence in this region is distinctive for each of the four PLC-β isoforms. Residues in the following groups were counted as similar: (A, S, T, C), (Y, F, I, L, V, M), (K, R), (E, D, Q, N), and (H, Y, F). To create the model, the turkey PLC-β2 side chains were replaced with the corresponding rat side chains using the Biopolymer module of the Insight II graphics package. Overlapping side chains were adjusted by manual rotamer, and energy was minimized using the Discover module of the same graphics package. Structural images of the resulting model were produced with the program Bobscript (22Esnouf R.M. J. Mol. Graph. Modl. 1997; 15 (132–134): 112-113Crossref Scopus (1796) Google Scholar). To delineate sites in PLC-β1 that are involved in regulatory interaction with Gαq, we initially mutated stretches of 6–15 amino acid residues in the carboxyl-terminal domain. Mutations were confined to residues 903–1143, which corresponds to the fragment previously referred to as Tail 32 (10). This region is required for stimulation by Gq (8Wu D. Jiang H. Katz A. Simon M.I. J. Biol. Chem. 1993; 268: 3704-3709Abstract Full Text PDF PubMed Google Scholar). Relatively small fragments of this domain, 15–33 kDa in size, also display intrinsic Gq GAP activity and inhibit stimulation of intact PLC-β1 by Gαq (10Paulssen R.H. Woodson J. Liu Z. Ross E.M. J. Biol. Chem. 1996; 271: 26622-26629Abstract Full Text Full Text PDF PubMed Scopus (58) Google Scholar). The region tested here excludes carboxyl-terminal residues 1145–1216, which are not required for interaction with Gαq (8Wu D. Jiang H. Katz A. Simon M.I. J. Biol. Chem. 1993; 268: 3704-3709Abstract Full Text PDF PubMed Google Scholar) (confirmed in this study). Bovine, rat, and human PLC-β1s differ at only 7 amino acid residues in this region, and all differences but one are highly conservative. Each mutant PLC-β1 protein was purified and assayed both for sensitivity of its phospholipase activity to stimulation by Gq and for Gq GAP activity. All mutant proteins displayed about the same specific activity of Ca2+-stimulated phospholipase activity according to the ratio of measured activity to the amount of PLC-β1 protein determined by Western blotting and were similar to wild type PLC-β1 according to total expression and relative purification. The initial multisite mutations produced three general phenotypes (Table I). Group I, mutations A1, A4, A5, A11, A12, A13, N17, and N18, displayed no obvious loss of either GAP activity or sensitivity to stimulation by GTPγS-activated Gαq. Group II, deletion mutations R14, R16, and R19, also displayed normal Gq GAP activity. Their sensitivity to stimulation by activated Gαq was also normal, but only when the concentration of the mutant PLC-β1 was >3 nm in the assay medium. At lower concentrations, their sensitivities to stimulation by Gαq were reduced up to 60%. Because the effect of these mutations is relatively minor and is detected only at low PLC-β1 concentrations, it is unlikely that they alter interaction with Gαq directly. We consider them to be without phenotype in this sense. Their subtle phenotype may reflect a decrease in stability of the PLC-β1 dimer (see below), but we have not characterized them further. Mutants A7 and A10 behaved much like others in Group II, but displayed a somewhat greater loss of responsiveness to Gαq (∼80%). We did not determine whether response to Gαq was greater at high concentrations of the mutant proteins.Table IDeletion and replacement mutations of PLC-β1MutationBeginEndGαq responseGqGAP activityGroupA110181028++IA210431052−−IIIA310571064−−IIIA410811082++IA510921097++IA611011109−−IIIA7904916m+IIA8917927−−IIIA9928943−−IIIA10944957m+IIA11965975++IA12976990++IA139981007++IR1410071012c+IIR1510621070−−IIIR16993998c+IIN1710761078++IN1810851091++IR19958964c+IIMutations are identified by the first and last residues mutated. A and N denote alanine/glycine substitutions, and R denotes deletions. Each mutant protein was purified and assayed for its response to GTPγS-activated Gαq and its Gq GAP activity. Regulatory behaviors are tabulated as +, wild type; c, response to activated Gαq decreased only at low concentrations of the mutant PLC-β1; m, minor decrease; −, significant decrease. Open table in a new tab Mutations are identified by the first and last residues mutated. A and N denote alanine/glycine substitutions, and R denotes deletions. Each mutant protein was purified and assayed for its response to GTPγS-activated Gαq and its Gq GAP activity. Regulatory behaviors are tabulated as +, wild type; c, response to activated Gαq decreased only at low concentrations of the mutant PLC-β1; m, minor decrease; −, significant decrease. Group III mutants displayed varied but significant loss of both responsiveness to Gαq and Gq GAP activity. The group III mutants were mutations A2, A3, A6, A8, A9, and R15 (TableI). To further delimit the amino acid residues that contribute to interaction with Gαq, we then mutated subsets of these regions. We concentrated on residues that are conserved among the four mammalian PLC-β isoforms because all four are sensitive to stimulation by Gαq, and all are Gq GAPs (3Rebecchi M. Pentyala S. Physiol. Rev. 2000; 80: 1291-1335Crossref PubMed Scopus (840) Google Scholar,23Smrcka A.V. Sternweis P.C. J. Biol. Chem. 1993; 268: 9667-9674Abstract Full Text PDF PubMed Google Scholar, 24Biddlecome G.H. Regulation of Phospholipase C-β1 by Gq and m1 Muscarinic Cholinergic Receptor.Doctoral dissertation. University of Texas Southwestern Graduate School of Biomedical Sciences, Dallas, TX1997Google Scholar, 25Lee C.-W. Lee K.-H. Lee S.B. Park D. Rhee S.G. J. Biol. Chem. 1994; 269: 25335-25338Abstract Full Text PDF PubMed Google Scholar). 3S. Mukhopadhyay, G. H. Biddlecome, and E. M. Ross, unpublished data. The second round of mutants helped localize regions important for regulatory interactions between PLC-β1 and Gαq, as shown in TableII. This initial screen also indicated that at least 61% (147 of 241) of the amino acid residues in the Tail 3 region of PLC-β1 may be either deleted or mutated to alanine or glycine without significant loss of regulatory interactions with Gαq.Table IICarboxyl-terminal uncoupling mutations in PLC-β1RegionAltered couplingNot testedA2ECQ1048T 1043VAE 1045A3/R15K1058I 1060K1063L1068K1069A6SY 1101I 1103VV 1106A8/A9K921SFV 918Q923L 922KHY 925MK 930LVKR 933KKTT 939Residues in regions of the group III mutants (Table I) were mutagenized in smaller groups and evaluated as described in Table I. Those residues whose mutations impaired coupling are listed in the second column. This column includes three lysine residues that were implicated in responsiveness to Gαq by the studies of Kim et al.(11Kim C.G. Park D. Rhee S.G. J. Biol. Chem. 1996; 271: 21187-21192Abstract Full Text Full Text PDF PubMed Scopus (194) Google Scholar). For grouped amino acid residues, the number given is that of the first residue. Open table in a new tab Residues in regions of the group III mutants (Table I) were mutagenized in smaller groups and evaluated as described in Table I. Those residues whose mutations impaired coupling are listed in the second column. This column includes three lysine residues that were implicated in responsiveness to Gαq by the studies of Kim et al.(11Kim C.G. Park D. Rhee S.G. J. Biol. Chem. 1996; 271: 21187-21192Abstract Full Text Full Text PDF PubMed Scopus (194) Google Scholar). For grouped amino acid residues, the number given is that of the first residue. The significantly uncoupled mutants displayed defects in both their responses to stimulation by activated Gαq and their Gq GAP activities, but specific patterns of behavior varied (Table III). As shown in Fig.1A, all of the class III mutants displayed decreased responsiveness to activated Gαq in terms of apparent affinity. The EC50for Gαq was increased by about 10-fold for each mutant. For several mutants (A3 and A1101 for example), the increased EC50 was the only significant cause of decreased response. Others also displayed a decreased maximal response to Gαq, and the amounts of protein used in the assays shown in Fig. 1A were varied over a 12.5-fold range to yield comparable maximal activities. The loss of response to Gqwas most pronounced for mutant A923 (residues Gln923, His926, Tyr927) in which the decrease in maximally stimulated activity approached 90%. Mutant A1048 also displayed a significant loss of maximal response. All mutants in class III that displayed a decreased maximally stimulated phospholipase activity displayed the same loss in apparent affinity for activated Gαq.Table IIIRegulatory properties of PLC-β1 mutantsResponse to Gαq-GTPγSGAP EC50GAPVmax(fold vs. Ca2+)(nm)(pmol/min)Wild type2.96 ± 0.451.3 ± 0.23.1 ± 0.3A9230.12 ± 0.22.4 ± 0.51.5 ± 0.2A30.68 ± 0.092.0 ± 0.40.49 ± 0.08A90.80 ± 0.142.4 ± 0.51.05 ± 0.16A11011.48 ± 0.321.4 ± 0.31.4 ± 0.25A10481.27 ± 0.192.3 ± 0.50.92 ± 0.13Mutant PLC-β1s were assayed for phospholipase activity in the presence either of 5 μm Ca2+ or 10 nmGTPγS-activated Gαq. Stimulation by Gαq is expressed as the ratio of that activity to activity measured in the presence of Ca2+. The Ca2+-stimulated activities varied from 0.11 pmol/min/ng to 0.14 pmol/min/ng. Gq GAP activity was measured in the steady-state, receptor-coupled assay over a range of PLC concentrations. Maximal activity and EC50 are both shown. Data shown are taken from at least three experiments with duplicate determinations in each assay. The error parameter is S.D. Open table in a new tab Mutant PLC-β1s were assayed for phospholipase activity in the presence either of 5 μm Ca2+ or 10 nmGTPγS-activated Gαq. Stimulation by Gαq is expressed as the ratio of that activity to activity measured in the presence of Ca2+. The Ca2+-stimulated activities varied from 0.11 pmol/min/ng to 0.14 pmol/min/ng. Gq GAP activity was measured in the steady-state, receptor-coupled assay over a range of PLC concentrations. Maximal activity and EC50 are both shown. Data shown are taken from at least three experiments with duplicate determinations in each assay. The error parameter is S.D. When the mutant PLC-β1 proteins were assayed for Gq GAP activity in the steady-state GTPase assay, all the class III alanine/glycine replacement mutants displayed decreased maximal Gq GAP activity. The decrease ranged from about 50%, for mutant A923 to 85% for mutant A3. However, concentration dependence for GAP activity was essentially unchanged for all these mutants (Fig.1B; Table III). This pattern contrasts with that observed for responses to Gαq, where the concentration of activated Gαq needed to increase phospholipase activity was increased, but often without significant loss of the maximum response. The behavior of the mutants in the two assays thus seems to differentiate between interactions needed to exert GAP activity and those needed to respond to activated Gαq. However, GAP activity in the steady-state GTPase assay depends on a relatively stable interaction among receptor, Gq, and GAP over the time course of multiple GTPase cycles (5Mukhopadhyay S. Ross E.M. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 9539-9544Crossref PubMed Scopus (156) Google Scholar, 7Ross E.M. Wilkie T.M. Annu. Rev. Biochem. 2000; 69: 795-827Crossref PubMed Scopus (950) Google Scholar, 17Biddlecome G.H. Berstein G. Ross E.M. J. Biol. Chem. 1996; 271: 7999-8007Abstract Full Text Full Text PDF PubMed Scopus (213) Google Scholar). Consequently, a decrease in maximal GAP activity in this assay may actually represent a decrease in affinity of the mutant PLC-β1 for Gαq. The simplest interpretation of these results is therefore that the alanine/glycine substitutions in the PLC-β1 carboxyl-terminal region decrease affinity for Gαq without altering the GAP activity of the PLC-β1 protein when it is bound to its Gαq target. Singer et al.(14Singer A.U. Waldo G.L. Harden T.K. Sondek J. Nature Struct. Biol. 2002; 9: 32-36Crossref PubMed Scopus (86) Google Scholar) recently determined the tertiary structure of a fragment of the carboxyl-terminal domain of turkey PLC-β2 that corresponds to residues 904–1174 of mammalian PLC-β1. To crystallize this fragment, they also deleted a region that is not well conserved among PLC-β isoforms, residues 961–1000 in the rat enzyme. Their tail construct contained 231 amino acid residues that are 33% identical and 60% similar to the corresponding portion of mammalian PLC-β1. The PLC-β tail is essentially a three-stranded coiled coil. Two tails bind to form a dimer with the dimer interface along the long axis of each monomer (14Singer A.U. Waldo G.L. Harden T.K. Sondek J. Nature Struct. Biol. 2002; 9: 32-36Crossref PubMed Scopus (86) Google Scholar). A dimeric structure for PLC-β mediated by the carboxyl-terminal domain is consistent with the regulatory behavior of tail constructs similar to the one used by Singer et al.(14Singer A.U. Waldo G.L. Harden T.K. Sondek J. Nature Struct. Biol. 2002; 9: 32-36Crossref PubMed Scopus (86) Google Scholar). Because of the strong sequence similarity between the rat PLC-β1 and turkey PLC-β2, we were able to model the carboxyl-terminal domain of PLC-β1 according to that of PLC-β2 (Fig.2). Those residues that could be mutated to alanine or glycine with minimal effect are cyan, and those whose mutation (individually or as part of a small group) caused significant uncoupling from Gαq are red; those generally unconserved residues that were not tested are light gray-green. Note that the number of residues grouped as being phenotypically sensitive to mutation is an overestimate because we did not mutate each residue in this group separately. Some residues shown in red may therefore be phenotypically neutral. The residues in the GAP domain of PLC-β1 whose mutations cause uncoupling are distributed in two distinct groups (Fig. 2). One group lies along the dimer interface, and the other lies on one surface of the dimer. The relevant mutations on the dimer interface include residues Lys1058, Lys1065, Lys1069, and Tyr1102 (Fig. 2B, vermilion). These conserved residues all contain large side chains that extend outward from the helices and make contact with residues in the opposite monomer. The side chains may contribute to the stability of the dimer, the correct alignment or orientation of the coils, or the regulatory interaction surface outlined by the second group of mutated residues. We have not attempted to assign the precise interfacial contacts in which these residues participate because the structure shown in Fig. 2remains a model and because several nonidentical residues are found in turkey PLC-β2. In addition, the residues in the monomer that lie across the dimer interface from these four side chains do not yield severe phenotypes when mutated individually or in groups (Fig.2C, cyan). Support of this idea is provided by the phenotypically neutral A1 mutation (residues 1018 through 1028). It includes several residues (Gln1021, Tyr1025, and Lys1028) that are apparently poised to interact with these functionally important side chains. Such data suggest that the stability or specificity of the dimer interface may not provide the only contribution to the loss of activity displayed by the Lys1058, Lys1065, Lys1069, and Tyr1102 mutations. The surface representation of these mutations diagrammed in Fig. 2C illustrates the accessibility of these side chains to the protein exterior and their position relative to the remaining group of essential residues. Lys1069 is particularly exposed. The second group of functionally important residues maps to one surface of the dimeric carboxyl-terminal domain and may thus outline a site of interaction with Gαq. These residues include Lys921, Lys925, Glu1048, and Lys1063 (Fig. 2A, fuchsia). They lie on two of the three helices of each monomer and are directed outward from a slightly concave surface near the center of the long axis of the dimer. The residues in each monomer are oriented essentially symmetrically with respect to the center of the long axis of the dimer. The region delimited by these mutations describes the surface onto which Singer and Sondek (14Singer A.U. Waldo G.L. Harden T.K. Sondek J. Nature Struct. Biol. 2002; 9: 32-36Crossref PubMed Scopus (86) Google Scholar) were able to dock two molecules of Gα. The phenotypes of these eight mutations are consistent with their direct interaction with Gαq and support the proposed binding of Gαq in this region. These residues are certainly involved in the coupling process that allows Gαq to activate the phospholipase activity of PLC-β. The other possible function for this region would be regulatory interaction of the carboxyl-terminal domain with the catalytic core of PLC-β. Distinguishing these possibilities or assigning specific residues to one function or the other awaits solution of the structure of the intact PLC-β molecule and its complex with activated Gαq. Several of the 13 residues whose mutation altered PLC-β-Gαq coupling could not readily be assigned either to the dimer interface or to potential interaction with Gαq. His926, Gln1050, and Gln1050 lie within the coiled coil of each monomer. They may participate in helix-helix interactions, but their importance has not been tested in single mutations. We speculate that Ser1101, Cys1049, and Tyr927 are actually phenotypically neutral because they were only mutated in combination with other residues that contribute to the dimer interface. In addition to the uncoupled mutations discussed above, we also characterized one mutation in which residues 1123–1138 of rat PLC-β1were deleted (Δ1123–38). This mutant displayed a Gq GAP phenotype qualitatively different from those of the mutants discussed above, and we therefore expressed it in Sf9 cells and purified it completely (17Biddlecome G.H. Berstein G. Ross E.M. J. Biol. Chem. 1996; 271: 7999-8007Abstract Full Text Full Text PDF PubMed Scopus (213) Google Scholar) to study its behavior more carefully. The Δ1123–38 PLC-β1 was stimulated only slightly by activated Gαq, to a maximal activity about 2% that of wild type PLC-β1 (Fig. 3A). In addition, stimulation occurred only at relatively high Gαq concentrations, with an EC50 at least 50-fold above that of wild type (Fig. 3A). In contrast, the maximal Gq GAP activity of Δ1123–38 PLC-β1 was unaltered (Fig. 3B). Its EC50 as a GAP was increased about 15-fold in the steady-state, receptor-coupled assay (20-fold in Fig. 3B). The increase in EC50 in both assays probably indicates decreased affinity for activated Gαq, but such a decrease cannot account significantly for the loss of responsiveness to Gαq. Although it was less sensitive than wild type to stimulation by Gαq, Δ1123–38 PLC-β1 displayed an ∼4-fold increased basal phospholipase activity in the absence of Gαq. No other mutant displayed this effect. Because maximal activation by Gαq is decreased and basal activity is increased, Gαq stimulates the activity of Δ1123–38 PLC-β1 only about 0.6% as much as that of wild type (see the Fig. 3legend). The elevated basal phospholipase activity of Δ1123–38 PLC-β1 does not result from increased sensitivity to Ca2+. Although both wild type and Δ1123–38 PLC-β1 have very low activity at low nanomolar concentrations of Ca2+, there was no obvious increase in the potency of Ca2+ at concentrations above 10 nm. We have not determined whether the elevated basal activity of Δ1123–38 PLC-β1 results from an increased Vmax, decreased Km, or both. The behavior of the Δ1123–38 mutant shows minimally that the deleted region is unnecessary for Gαq binding and GAP activity, although the deletion does decrease affinity for Gαqsomewhat. The loss of responsiveness to Gαq and the increase in basal activity suggest that this region may be involved in regulatory coupling between the carboxyl-terminal region of the PLC-β molecule and its central catalytic domain. The deletion of residues 1123–1138 falls in the middle of the third of the three helices in the PLC-β carboxyl-terminal domain (Fig. 3, A andB). It would therefore be predicted to either grossly change the packing of the basic coiled-coil structure of this domain or shift the registry of the helix-helix interaction. The latter seems more likely because GAP activity is retained. This deletion apparently leaves the basic structure of the carboxyl-terminal domain intact, either by terminating helix-helix interactions at the initiation of the deletion or by replacing these interactions with helical elements carboxyl-terminal to the deletion. Unbiased mutagenic scanning of the carboxyl-terminal region of PLC-β1 indicates that a significant majority of the amino acid residues can be replaced by alanine and/or glycine with little, if any, loss to its regulatory functions. This structural stability is consistent with the extended coiled-coil structure of this domain (14Singer A.U. Waldo G.L. Harden T.K. Sondek J. Nature Struct. Biol. 2002; 9: 32-36Crossref PubMed Scopus (86) Google Scholar). The naive mutagenesis approach used here, initial replacement of contiguous segments with alanine and a few glycine residues, may be generally useful when applied to protein domains of unknown structure, particularly if they are predicted to contain large runs of α-helix. This is true for the PLC-β carboxyl-terminal domain.
Referência(s)