Archives of Biochemistry and Biophysics 453 (2006) 151–160 www.elsevier.com/locate/yabbi
A recombinant transductor–eVector system: In vitro study of G inhibitory protein (G-alpha-i1) direct activators Lorenzo Di Cesare Mannelli a,¤,1, Alessandra Pacini b,1, Annarita Toscano b, Carla Ghelardini a, Dina Manetti c, Fulvio Gualtieri c, Tarun B. Patel d, Alessandro Bartolini a b
a Department of Preclinical and Clinical Pharmacology, University of Florence, Viale Pieraccini 6, 50134 Florence, Italy Department of Anatomy, Histology and Forensic Medicine, Anatomy Section, University of Florence, Viale Morgagni 85, 50134 Florence, Italy c Department of Pharmaceutical Sciences, University of Florence, Via U. SchiV 6, I-50019 Sesto Fiorentino (FI), Italy d Department of Pharmacology, Stritch School of Medicine, Loyola University Chicago, 2160 First South Avenue, Maywood, IL 60153, USA
Received 29 May 2006, and in revised form 10 July 2006 Available online 2 August 2006
Abstract Mutations and altered functionality of the inhibitory subfamily of G proteins (Gi) are involved in pathological states. Compounds able to activate Gi in a receptor-independent manner would be useful to treat these pathological conditions. Aimed to study Gi direct activation we have reconstituted a recombinant transductor–eVector complex cloning both the mammalian Gi1 subunit and adenylate cyclase (AC). The myristoylation of G, fundamental for interaction with AC, was obtained in the procaryotic expression host Escherichia coli transformed with a single plasmid containing both the coding sequences for human Gi1 and Saccharomyces cerevisiae myristoyl transferase. AC-V isoform was obtained by the expression of its cytosolic domains. A recent synthesized molecule, named BC5, was tested to evaluate its pharmacological proWle in a Gi/AC cell-free complex model. In this functional transductor–eVector system BC5 was able to activate Gi signalling, moreover providing a new tool to give a better insight into G-protein receptor-independent modulation. © 2006 Elsevier Inc. All rights reserved. Keywords: G protein; Myristoylation; Protein coexpression; Direct activator; Adenylate cyclase; Mastoparan
Heterotrimeric guanine-nucleotide binding proteins (G proteins) form a superfamily of signal transduction proteins that are peripherally associated with the plasma membrane and provide signal coupling to seven transmembrane receptors [1,2]. Upon activation G-GDP- complex dissociate in G-GTP subunit and G dimer, both able to initiate cellular response by altering the activity of speciWc eVector molecules that generate intracellular signals and alter cell functions. Given the immense diversity and cross-talk of signal transduction pathways controlled by heterotrimeric G-protein activation [3], and the numerous regulatory proteins that modulate signalling by directly competing with receptor, G-protein or eVector interactions
*
1
Corresponding author. Fax: +39 55 4271280. E-mail address:
[email protected] (L. Di Cesare Mannelli). These authors contributed equally to this work.
0003-9861/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.abb.2006.07.006
[4,5], this model of G protein signalling has grown increasingly complex. A growing body of evidence shows that mutations in genes encoding G proteins are an important cause of human disease [6]. In particular, mutations in the G inhibitory subunit (Gi)2 codifying genes have been associated with tumours of the adrenal cortex, endocrine cancers of the ovary [7], and pituitary ACTH-secreting adenomas [8]. A somatic mutation in the GNAI2 gene, that codes for the isoform of the G inhibitory subunit, has been correlated to an idiopathic ventricular tachycardia [9]. Recently, our group has demonstrated a hypofunctionality of Gi in lymphocytes of cephalalgic and Wbromialgic patients [10,11]. 2 Abbreviations used: Gi, G inhibitory subunit; AC, adenylate cyclase; LIC, ligation-independent cloning system; RT-PCR, reverse transcriptionpolymerase chain reaction; ORF, open reading frame; UTR, untranslated region.
152
L.Di Cesare Mannelli et al. / Archives of Biochemistry and Biophysics 453 (2006) 151–160
The increasingly complex model for G-protein signalling and their involvement in many pathological conditions, have driven the need for new compounds able to modulate G-protein functions. Classically, drug discovery eVorts have primarily focused on compounds that interact directly with G-protein-coupled receptors. Nevertheless, considering that the activation of G-proteins-coupled receptors can lead to the recruitment of several eVectors [12] or, on the contrary, that some pathological conditions, such as cardiac hypertrophy, are mediated by persistent stimulation of multiple receptors coupled with a common G protein [13], targeting intracellular G proteins directly would provide new approaches and selectivities for drug treatment [14– 16]. This would allow to modulate individual eVector pathways, to alter speciWc signals from particular G-protein classes or subclasses, and to modify the kinetics of G-protein signalling. Drugs, that are already known to modulate G proteins in a receptor-independent manner included: peptides, such as Mastoparan, extracted from wasp venom, and low molecular-weight compounds [17], such as alkyl-substituted amino acid amines, like N-dodecyl-lysinamide (ML250) [18; see Fig. 1], and lipoamines [19]. Unfortunately, although successful, most of these molecules are weak modulators of signalling, exhibiting their activities at relatively high doses (from M to mM concentrations). Recently, our group, using human lymphocytes have demonstrated that new synthesized compounds can interfere with the Gi-protein signalling pathway [20]. Moreover, in a Gi-protein reconstituted system, free from the eVects of the receptor, we demonstrated that these new compounds directly activated Gi subunits [20,21]. Among the active compounds that have been tested, a 4-aminopiperidinic derivative named BC5 (see Fig. 1) showed high activity with a good correlation between the in vivo and in vitro experiments. Here, we examine the eVect of this new activator (BC5) in an in vitro system where we reconstituted a transductor– eVector complex of the mammalian myristoylated Gi1 (isoform 1 of Gi) and the adenylate cyclase (AC). G myristoylation, fundamental for interaction with AC, was obtained in the procaryotic expression host Escherichia coli transformed with a plasmid containing both the coding
Fig. 1. Structure of ML250 and BC5.
sequences of the human Gi1 and Saccharomyces cerevisiae myristoyl transferase. AC was obtained by the expression of the cytosolic domains (C1, C2) of the V enzyme isoform in E. coli. C1 and C2 domains create a pseudosymmetrical heterodimer that forms the catalytic moiety of the enzyme and that is the target for most known intracellular regulators, such as Gi [22,23]. In the current study, a recombinant Gi1-ACV–C1–C2 system has been expressed and, after optimisation of the functional conditions, has been challenged with compounds such as Mastoparan, ML250, and BC5 that were previously shown to interact with isolated Gi proteins. Materials and methods Plasmid construction To ensure synthesis of myristoylated G protein as described by [24], mammalian Gi1 subunit and yeast Nmyristoyltransferase were coexpressed in a single plasmid using the vector ligation-independent cloning system (LIC; Novagen, San Diego, CA). The entire ORF (open reading frame) of both human G inhibitory subunit isoform 1 and yeast myristoyl transferase were obtained by reverse transcription-polymerase chain reaction (RT-PCR). Total RNA from human brain (for Gi1) and from S. cerevisiae (for MT) was puriWed using TriReagent (Sigma, Milan, Italy) according to the manufacturer’s instructions and resuspended in diethyl pyrocarbonate-treated water. QuantiWcation and purity of the RNA was assessed by A260/A280 absorption, and RNA samples with ratios greater than 1.6 were stored at ¡80 °C until needed. One micrograms of the total RNA was reverse-transcribed and ampliWed using the SuperScript One-Step RT-PCR kit (Invitrogen, Paisley, UK). The cDNA coding for the human i1 isoform was generated by two sequential ampliWcation reactions as described in [21]. For MT, RT-PCR was carried out in onestep with upstream primers annealing to the 5⬘-end starting from ATG and downstream primers complementary to the 3⬘UTR (untranslated region). RT-PCRs were performed in a total volume of 50 L, as follows: after cDNA synthesis for 30 min at 55 °C, the proWle of the replication cycles was as follows: denaturation at 94 °C for 15 s, annealing at 60 °C for 30 s, and polymerisation at 68 °C for 1 min for 35 cycles; the Wnal extension step was at 72 °C for 5 min. The primers used for PCRs were as follows: Gi1 (ORF 1) forward 5⬘-GACGACGACAAG ATGGGCTGCACGCTGAG-3⬘, and Gi1 reverse 5⬘-CGC GGGCGGCCGTTAAAAGAGACCACAATCTTTTAG ATT-3⬘ (GenBank Accession No. NM_002069); MT (ORF2) forward 5⬘-GCGGGCCCGGCCTTGATGTCA GAAGAGGATAAAGC-3⬘, and MT reverse 5⬘-GAGG AGAAGCCCGGTCTACAACATAACAACACCGA-3⬘ (GenBank Accession No. M23726). PCR products (shown in Fig. 2a) were puriWed by SpinPrep PCR Clean-Up Kit (Novagen) and, in order to generate compatible overhangs, ORF1 and ORF2 were treated with T4 DNA Polymerase
L.Di Cesare Mannelli et al. / Archives of Biochemistry and Biophysics 453 (2006) 151–160
153
expression was induced by adding isopropyl--D-thiogalactoside (IPTG) to a Wnal concentration of 1 mM. After 16 h of incubation, bacteria were collected by centrifugation and conserved at ¡80 °C. PuriWcation of His6-subunits fusion proteins
(Novagen). Then, the ORFs were annealed to the vector LIC Duet Mini Adaptor (Novagen), a vector designed to anneal to the 3⬘-end of the ORF-1 and the 5⬘-end of the ORF-2. The obtained ORF1–Adaptor–ORF2 complex was subsequently cloned in the expression vector pRSF2 Ek LIC (Novagen). Plasmid pRSF2 containing both Gi1 and MT sequences (pRSF2/Gi1/MT) was ampliWed in E. coli DH5- competent cells (Novagen) and the construct was veriWed by automatic sequencing. The canine ACV C1–C2 fragments plasmid was obtained as previously described [25].
Bacterial pellets were resuspended in 4 mL of Tris–HCl 50 mM, pH 8.0, cold buVer containing NaCl 200 mM, mercaptoethanol 5 mM and protease inhibitor cocktail (Roche, Milan, Italy). After an incubation for 30 min on ice in presence of 8 mg of lysozyme, cells were sonicated with six 10-s bursts at high intensity. DNA was sheared by passing the preparation through an 18-gauge syringe needle several times. The lysates were then centrifuged at 16,000g for 15 min to pellet the insoluble debris. Supernatants were aliquoted and stored at ¡80 °C or used for puriWcation of recombinant protein. The His-tag fusion proteins were puriWed by a Ni aYnity chromatography, using 1 mL of Ni-Nitriloacetic acid aYnity resin (Novagen) and according to the protocol provided by the manufacturer. Gi1 was eluted by an EDTA gradient as follows: column was washed two times with a buVer containing Tris–HCl 50 mM, pH 8.0, NaCl 200 mM, and EDTA 1 mM. Fusion protein was eluted with four 0.5 mL aliquots of a buVer containing Tris–HCl 50 mM, pH 8.0, NaCl 200 mM and EDTA 80 mM. AC was eluted by both EDTA and pH gradients, the last obtained with an elution buVer consisting of Tris–HCl 50 mM, pH 4.5, NaCl 500 mM. Protein eluates were dialyzed against Tris–HCl 50 mM, pH 8.0. The yield of the fusion proteins that we obtained was estimated to be about 8–10 mg/L of cultures. To verify the presence of the fusion proteins, an aliquot (500 L) of each culture was separated by electrophoresis on a 4–12% polyacrylamide gradient gels (Invitrogen). After transfer, the non-speciWc sites of nitrocellulose membrane were blocked with 5% non-fat dry milk in PBST (PBS containing 0.1% Tween 20) for 1 h at room temperature. The incubation with the anti-Gi1 subunit (1:1000 dilution) was carried out overnight at 4 °C in PBST with 5% non-fat dry milk. Detections were performed using goat anti-rabbit IgG HRP (horseradish peroxidase)-conjugated antibody (Santa Cruz, Milan, Italy) and the reactions were revealed using the Opti-4CN Substrate kit (Bio-Rad, Milan, Italy) according to the manufacturer’s instructions.
Expression of recombinant proteins in E. coli
Fusion proteins cleavage and puriWcation of subunits
For production of myristoylated Gi1 (Gi1-mir) and C1 and C2 AC domains, competent BL21(DE3) cells (Novagen) were transformed with the plasmid pRSF2/Gi1/MT or pTrcHisB/C1–C2 and selected for ampicillin (100 g/ mL) or kanamicine (30 g/mL) resistance, respectively. For large-scale production of recombinant proteins, 100 mL of Luria–Bertani broth were inoculated with 3 mL of a 0.5 OD culture. Cultures were grown up to 0.5–1.0 OD and then
The fusion proteins were cleaved by Enterokinase digestion according to the manufacturer’s protocol (Invitrogen). BrieXy, approximately 20 g of fusion protein was cleaved through incubation with EKMax™ for 16 h at 37 °C in a total volume of 30 L EKMax™ BuVer (500 mM Tris–HCl, pH 8.0, 10 mM CaCl2, 1% Tween 20). After digestion EKMax™ was removed by EK-Away resin (Invitrogen) following the instruction of the producer.
Fig. 2. PCR products, protein expression levels and binding of myristoylated human Gi1. (a) The cDNA encoding for the human i1 isoform of G protein was obtained by two sequential ampliWcation reactions: A and B fragments were produced by the Wrst ampliWcation reaction (lanes A and B), whereas the entire Open Reading Frame was assembled in a second PCR reaction (lane ORF). The cDNA coding for myristoyl transferase was obtained by RT-PCR starting from total RNA of S. cerevisiae (lane MT). (b) Lanes A and B: Western blot analysis of G inhibitory subunit isoform 1 (lane A) and C1–C2 ACV fragments (lane B). Lane C, Gi1 and AC binding assessed by immunoprecipitation. Lane M, marker of molecular weight.
154
L.Di Cesare Mannelli et al. / Archives of Biochemistry and Biophysics 453 (2006) 151–160
Immunoprecipitation In order to activate the Gi1-mir, 1 g of Gi1-mir was incubated 1 h at room temperature in a buVer containing Tris–HCl 50 mM, pH 8.0, GTPS 200 M and MgCl2 5 mM. The activated Gi1-mir was mixed with 20 g (320 ng of puriWed C1–C2) of the supernatant obtained from the lysis of bacteria expressing the ACV C1–C2 fragment, and incubated for 10 min at room temperature. The mix underwent to immunoprecipitation: after an incubation of 2 h, under shaking at 4 °C, with 5 g of the anti-Gi1-mir antibody; 50 L protein A sepharose beads (Sigma) were added and incubated for 2 h on a rocking platform. After precipitation, the protein A sepharose–Gi1-mir complex was washed with Tris–HCl 50 mM, pH 8.0, and analyzed by Western blotting (see the paragraph entitled “PuriWcation of His6-subunits fusion proteins” of this section for details). Complex formation was revealed by immunoblot using a speciWc anti-C1–C2 antibody (dilution 1:200), purchased from Santa Cruz, highlighted with the Opti-4CN Substrate kit. As negative controls of the immunoprecipitation experiments we omitted the ACV C1–C2 fragment from the reaction mix and we performed the same experiment with a non-myristoylated Gi1 subunit (data not shown). Liposome preparation Soy phosphatidylcholine was purchased from Avanti Polar Lipids Inc. (Alabama). Multilamellar vesicles (10 mg/ mL total lipid concentration) have been prepared by dispersing a dry lipid Wlm in the buVer (50 mM Tris–HCl buVer, pH 8.0). These lipids spontaneously form in water polydisperse vesicular suspensions by simple shaking. This dispersion was subjected to 5 min of forceful shaking, frozen under liquid nitrogen, and thawed (40 °C) six times. Afterward, it was sized down to unilamellar vesicles of approximately 90 nm radius by eleven repeated extrusions through two stacked polycarbonate Wlters with 200 nmsized pores, followed by eleven repeated extrusions through 100 nm-sized pores membranes [26]. Filtration was performed at room temperature with the Extruder by Lipex Biomembranes Inc., Vancouver (Canada) and Nuclepore polycarbonate membranes. The size distribution of unilamellar liposomes was veriWed by dynamic light scattering. Reconstitution of phospholipidic vesicles Gi protein solution was mixed with liposome to a Wnal concentration of 150 nM for subunit and 0.3 mg/mL for liposomes, in Tris–HCl 50 mM (pH 8.0). The system was incubated at room temperature for 20 min, at 4 °C overnight.
was incubated at 20 °C for 30 min in 100 L of a reaction mixture containing 50 mM Tris–HCl (pH 8.0), 1 mM EDTA, 1 mM DTT, 0.1% Lubrol, 1.1 mM MgSO4, and 0.5 M [-32P]GTP (0.1 Ci/tube; PerkinElmer). Reaction was stopped by adding 700 L of an ice-cold 5% (w/v¡1) charcoal suspension in 50 mM NaH2 PO4. The mixture was centrifuged at 13,000g for 18 min at 4 °C. A fraction of the supernatant (200L) was counted in 2 mL scintillation solution. The high-aYnity GTPase activity was calculated subtracting the 32Pi released in presence of 100 M cold GTP from total 32Pi hydrolyzed. GTPS binding Binding of GTPS (a non-hydrolysable GTP analogue) to recombinant Gi protein was measured according to a standard method [27]. Three picomoles of subunit in native conditions or reconstituted into lipidic vesicles were incubated at 30 °C for 30 min in 100 L of a reaction mixture containing 50 mM Tris–HCl (pH 8.0), 1 mM EDTA, 1 mM DTT, 1.1 mM MgSO4, and 0.1 M [35S]GTPS (0.1 Ci/tube; Perkin-Elmer). Reaction was stopped by the addition of 1 mL of an ice-cold stop buVer (consisting of 10 mM Tris–HCl, pH 8.0, 25 mM MgCl2, 100 mM NaCl). The diluted samples were Wltered through cellulose nitrate membranes (0.45 m pore size) under weak vacuum. The Wlters were washed with 12 mL of the same buVer and dried. The retained radioactivity was quantiWed by scintillation spectroscopy. SpeciWc binding was calculated by subtracting the GTPS bound in presence of 100 M cold GTPS from total binding. Adenylate cyclase activity Five micrograms of the supernatant obtained from the lysate of the ACV C1–C2 fragment-expressing bacteria, or 80 ng of puriWed C1–C2 were incubated for 10 min at 30 °C in 100 L of a buVer containing 50 mM Tris–HCl (pH 8.0), 1 mM DTT, 0.1 mg/mL bovine serum albumine, 0.6 mM EDTA, 0.05 M GTPS (except when diVerent speciWed), 5 mM MgS04, and 1 mM ATP. The process was conducted in the presence or in the absence of Gi1 and activators (Mastoparan, Alexis, San Diego, CA; ML250 and BC5, compounds synthesized by Prof. Fulvio Gualtieri and coworkers, Department of Pharmaceutical Science University of Florence according to the synthetic pathway, respectively, proposed in [18] or in [20], and with or without liposomes (10 g/sample). Reaction was stopped by adding 100 L of cold 100 mM EDTA. cAMP production was measured by an immunoenzymatic assay (cAMP Biotrak EIA assay kit; Amersham, Milan, Italy). As negative control, in the evaluation of activator eYcacy, the same experiments were carried out in the absence of Gi1.
GTP hydrolysis Statistical analysis GTPase activity of the recombinant Gi1 protein was measured according to a standard procedure [27]. Three picomoles of G, alone or incorporated into lipidic vesicles,
All experimental results were expressed as the means § SEM and statistical analyses were performed by ANOVA.
L.Di Cesare Mannelli et al. / Archives of Biochemistry and Biophysics 453 (2006) 151–160
p < 0.05 or p < 0.01 where indicated, were considered as signiWcant. Results The isoform 1 of inhibitory G-protein subunit was coexpressed in E. coli with the S. cerevisae N-myristoyltransferase protein in order to ensure synthesis of myristoylated G protein [24]. Human Gi1 coding sequence (CDS) was obtained with two sequential ampliWcation reactions as described in an our previous report [21], whereas myristoyl transferase sequence was obtained from S. cerevisiae total RNA with an one-step RT-PCR reaction. The results of the Gi1 and MT PCRs are shown in Fig. 2a. Coexpression of both proteins was achieved by using a single vector (pRSF2 Ek/LIC). A Western blot analysis of puriWed Gi1 and ACV–C1–C2 fragments was performed in order to verify protein expression levels and protein speciWcity and integrity. As shown in Fig. 2b, lanes A and B, Gi1 molecular weight (40 kDa) was comparable with the actually determined molecular weight and with standard G proteins (data not shown) whereas the molecular weight of ACV– C1–C2 (»70 kDa) was comparable with that obtained by Patel et al. [25]. The measurement of Gi1 myristoylation was performed evaluating the G-AC interaction by immunoprecipitation. Fig. 2b, lane C shows that activated Gi1 binds to ACV fragments constituting a complex that can be precipitated by G antibody and revealed by AC antibody. To test the Gi1-mir functionality the GTP hydrolysis and GTPS binding were evaluated in the presence and in the absence of phospholipidic vesicles. Fig. 3 shows basal Gi1-mir GTPase activity: according to published data [27] the level of hydrolyzed phosphate was 0.12 mol 32P/mol G/ min, both in the absence (Basal – white histogram) and in
Fig. 3. GTP hydrolysis and eVect of liposomes. Three picomoles of myristoylated Gi1 were incubated for 30 min at 30 °C in the absence (white histograms) and in the presence (black histograms) of phospholipidic vesicles and the level of hydrolysed phosphate was measured. The graph shows GTPase activity in the basal condition (Basal) and in the presence of the direct activator ML250 (1 £ 10¡4 M) (ML250). **p < 0,01 vs basal level; 䊊䊊 p < 0.01 vs ML250 stimulation obtained in the absence of liposomes.
155
the presence (Basal – black histogram) of liposomes. The presence of the direct activator ML250 (1 £ 10¡4M) stimulated G hydrolytic activity up to 0.17 mol 32P/mol G/min in the absence of liposomes (41% stimulation in respect to basal level; Fig. 3, ML250 – white histogram) and up to 0.20 mol in the presence of phospholipides (67% stimulation; Fig. 3, ML250 – black histogram). The GTPS binding experiments demonstrated that the direct activators, ML250 and BC5, are able to stimulate the recombinant Gi1-mir. Fig. 4 shows the dependence of the GTPS binding from the concentration of ML250 (䉲), with 5 £ 10¡5M as the concentration able to increase binding from 6 to 52 fmol (EC50 48.3 § 1.8 M), and of the compound BC5 (䉬) that was able, at the same concentration, to stimulate the GTPS binding up to 200 fmol (EC50 11,4 § 1,2 M). The BC5 stimulation of Gi1-mir was comparable to the results obtained in the presence of the unmodiWed i1 [28]. Moreover, the presence of phospholipidic vesicles signiWcantly increased both ML250 (䊉) and BC5 (䊏) stimulation up to 74 and 270 fmol, respectively, at 5 £ 10¡5M. Finally the eVects of ML250 (5) and BC5 (䉫) on G protein were conWrmed by the presence of 1 £ 10¡4M suramin, a G-protein inhibitor, that completely reverted the eVect of two activators on GTPS binding test. In order to set up the optimal conditions to evaluate the AC activity, we measured the fmol of cAMP released in various conditions. First we compared the basal AC activity measured on the lysate obtained from bacteria expressing the C1- C2 soluble form of ACV fragments, and on the puriWed fragments. As regards the puriWed fragments the demonstration that an imidazole-based elution denaturates AC protein [25], forced us to compare two other diVerent
Fig. 4. GTPS binding and eVect of direct activators. Three picomoles of myristoylated Gi1 was incubated for 30 min at 30 °C in the presence of increasing concentration (1 £ 10¡4 to 1 £ 10¡6 M range) of ML250 and BC5 in the absence (ML250 䉲; BC5 䉬) and in the presence (BC5 䊏; ML250 䊉) of liposomes. EVect of the G-protein inhibitor suramin (1 £ 10¡4 M), on the stimulatory eVect of ML250 (5) and BC5 (䉫). **p < 0.01 vs BC5 stimulation obtained in the absence of liposomes. 䊊䊊 p < 0.01 vs ML250 stimulation obtained in the absence of liposomes.
156
L.Di Cesare Mannelli et al. / Archives of Biochemistry and Biophysics 453 (2006) 151–160
elution buVers: EDTA- or pH-gradient type of elution. The results in Fig. 5A show that the AC activity measured on 5 g of bacterial lysate (lysate – white histogram) or on 80 ng of puriWed C1–C2 produced the same amount (300 fmol) of cAMP, and the elution buVers had no inXuence on this activity (EDTA and pH – white histograms). The same experiments were conducted in the presence of 50 nM Gi1-mir activated by 100 M GTPS (Fig. 5A – dark histogram), and the results show that the presence of the Gi1-mir was able to inhibit AC activity up to 30% on lysate (lysate – dark histogram) and on puriWed fragments eluted with EDTA (EDTA – dark histogram), whereas the fragments eluted with pH gradient showed only a 10% of AC inhibition (lane pH – dark histogram). The eVect of an acid pH is probably responsible of an alteration of the interaction between AC and G. Finally, we measured the AC activity on the lysate in the presence of liposomes (Fig. 5A, lysate + lipo). The results show that the presence of liposomes increased the basal cAMP production more than 80% (lysate + lipo – white histogram). In the presence of activated Gi protein, this high basal level did not allow more than 23% inhibition (lysate + lipo – dark histogram). Hence, we evaluated the entity of the AC inhibition on bacterial lysate at increasing concentrations of Gi1 protein. In Fig. 5B is shown that the inhibition is dependent by the Gi1 concentration up to 50 nM; at higher concentrations of Gi1 a plateau has been achieved. In order to eliminate any side eVects due to the lack or the excess of some known parameters that inXuence the Gi1 protein activity (such as the concentrations of Mg2+ and GTPS or the incubation time), we measured the cAMP production varying these parameters. Given the positive correlation that exist between Mg2+ concentration and Gi1 and AC activity we measured the C1–C2 fragment activity, in the presence of increasing concentration of Mg2+. Fig. 5C – white histograms demonstrates that the presence of Mg2+ gradually increased the cAMP production from 160 fmol (at 0.5 mM Mg2+) to 600 fmol (at 50 mM Mg2+). The same experiment, performed in the presence of Gi1 activated by GTPS (Fig. 5C – dark histograms and inner graph) showed that the inhibition was a direct function of Mg2+ concentration too, at least up to 5 mM of Mg2+, the concentration that elicits the highest inhibition (30%). In Fig. 5D a GTPS concentration-eVect relationship is presented: 100 M GTPS was able to induce a good G-protein activation (32%) while a 0.05 M GTPS induced a low stimulation, useful to highlighted G protein activator eVects. Finally, we evaluated if the incubation time for GTPS stimulation of Gi1-mir had an inXuence on AC activity and, as a consequence, on the eVects of G protein activators. Fig. 5E shows that optimal optimal AC activity inhibition (32%) was reached at 30 minutes of incubation time for 50 nM Gi1-mir with 100 M GTPS. Once we had established the best conditions for our in vitro system, we evaluated the eYcacy and the potency of reference compounds (Mastoparan and ML250) and of a new synthetic molecule (BC5), in the presence and in the
absence of liposomes. Results are shown in Fig. 6A–C. In the presence of 50 nM Gi1-mir, 5 mM Mg2+ and 0.05 M GTPS we have evaluated cAMP production by 5 g of bacterial lysate after 30 min incubation of Gi1-mir with Mastoparan and ML250. In the absence of liposomes, Mastoparan (from 5 £ 10¡8 to 1 £ 10¡5 M) signiWcantly stimulated Gi1-mir. 1 £ 10¡7 Mastoparan was able to induce maximum inhibition reducing cAMP level from 314 fmol (basal level) to 217 fmol (31% inhibition) (Fig. 6A, white circles). The presence of phospholipidic vesicles did not alter Mastoparan eYcacy (31% maximum inhibition), whereas the potency range was slightly shifted (1 £ 10¡8 to 5 £ 10¡6 M) (Fig. 6A, dark squares). In Fig. 6B ML250 eVect is shown. cAMP production was reduced in a signiWcative manner but ML250 showed a minor potency (1 £ 10¡6 to 5 £ 10¡6 M) with respect to MP, both in the absence and in the presence of phospholipides. In particular, in the absence of liposomes ML250 was able to reduce cAMP production up to 29%, whereas the presence of liposomes reduced ML250 activity (maximum of inhibition: 17%) and increased its potency. Finally, the eVect of the newly synthesized molecule BC5 was evaluated. Fig. 6C shows the concentration-eVect relationship in the absence (dark points) and in the presence (dark squares) of liposomes. In the Wrst experiment cAMP production from 5 g of bacterial lysate was signiWcatively reduced in the presence of 1 £ 10¡7 to 1 £ 10¡6 M BC5 (maximum inhibition 39% at 5 £ 10¡7 M). The incubation with phospholipidic vesicles reduced BC5 eYcacy (maximum inhibition 22% at 5 £ 10¡8 M) whereas, in these conditions BC5 signiWcatively stimulated Gi1-mir from 1 £ 10¡8 to 1 £ 10¡5 M showing a higher potency (Fig. 6C, dark squares). To evaluate if the activators exhibited some aspeciWc eVects on AC, parallel experiments were conducted in the absence of Gi1-mir protein. Our results showed that, at least in the concentration range from 1 £ 10¡9 to 5 £ 10¡5 M, Mastoparan, ML250, and BC5 did not directly interact with AC (data not shown). Discussion G-protein signalling necessitates the development of new tools for probing G-protein function. Upon stimulation of a heptahelical receptor by the appropriate agonists, the cognate G proteins interact sequentially with several reaction partners by exposing appropriate binding sites. For most of these domains, low molecular weight ligands have been identiWed that either activate or inhibit signal transduction. Because compounds that directly interact with G proteins can provide new forms of selectivity, G-protein subunits may, therefore, be considered as potential drug targets. Nevertheless, to evaluate the pharmacological proWle of such drugs it is necessary to create an isolated system, similar as much as possible to the intracellular environment, in order to nullify the inXuence of the receptor. For this reason we designed a system consisting of G
L.Di Cesare Mannelli et al. / Archives of Biochemistry and Biophysics 453 (2006) 151–160
157
Fig. 5. Adenilate cyclase activity is dependent on chromatography elution buVer, Mg2+ concentration and Gi1-mir. (A) Cyclic AMP basal production was measured in the absence (white histograms) or in the presence of 50 nM Gi1-mir activated by 100 M GTPS (black histograms) on bacterial lysate expressing the C1–C2 soluble form of ACV (lysate) and on puriWed fragments eluted by an EDTA-based (EDTA) or a pH-based buVer (pH). The histograms on the right side of the graph show the eVect of phospholipids in bacterial lysate AC activity. (B) EVect of Gi1-mir concentration on AC activity. cAMP production level from 5 g of lysate from bacteria expressing the ACV C1–C2 fragment was evaluated in the presence of increasing concentrations of Gi1-mir (0–500 nM). (C) EVect of Mg2+ concentration (0.5–50 mM range) on AC and Gi1-mir (50 nM) functionality. (D) EVect of increasing concentration of GTPS (from 0.01 M up to 1000 M) on Gi1-mir (50 nM) activation in the presence of 5 mM Mg++. (E) EVect of increasing GTPS incubation time on Gi1-mir activation. 50 nM Gi1-mir in the presence of 5 mM Mg++ was stimulated with 100 M GTPS and incubated from 0 to 30 min at room temperature before adding 5 g of AC bacterial lysate. In each panel the inner graphs show the percentage of Gi1-mir-dependent AC inhibition. All data reported in the Wgure are expressed as means § SEM of three independent experiments conducted in triplicate.
158
L.Di Cesare Mannelli et al. / Archives of Biochemistry and Biophysics 453 (2006) 151–160
Fig. 6. EVect of Mastoparan, ML250 and BC5 on Gi1-dependent AC activity. (A) Increasing concentrations of Mastoparan (1 £ 10¡9 to 5 £ 10¡5 M) were incubated, in the absence (䊊) or in the presence of liposomes (䊏) with 50 nM Gi1-mir (GTPS 0.05 M; Mg++ 5mM) and added to 5 g of lysate from bacteria expressing the ACV C1–C2 fragment. (B) EVects of increasing concentrations of ML250 (from 1 £ 10¡9 M up to 1 £ 10¡5 M) on cAMP production evaluated in the absence (䊐) or in the presence of liposomes (䉬). (C) Activity of BC5. Comparison of increasing concentrations of BC5 (from 1 £ 10¡9 M up to 1 £ 10¡5 M) on ACV C1– C2 cAMP production in the absence (䊉) and in the presence (䊏) of liposomes. All data reported in the Wgure are expressed as mean § SEM of three independent experiments conducted in triplicate. 䊊䊊p < 0.01 vs AC activity (basal value) in the presence of phospholipids. **p < 0.01 vs basal level in the absence of phospholipids.
inhibitory protein (Gi) and its eVector, the adenylate cyclase (AC). Considering that the binding with AC needs the myristoylation of G protein [28,29], in order to obtain protein myristoylation in a procaryotic system (such as E. coli), we expressed the Gi protein and the N-myristoyl transferase simultaneously. Classically, coexpression of multiple target proteins has been obtained transforming bacteria with two distinct plasmids [30]. The success of this strategy depends on the stable maintenance of both plasmids in the same cell [31] that compete with each other during replication and partitioning, such that, after few generations, cells can loose one of the plasmids [32]. Theoretically, any combination of plasmids in diVerent incompatibility groups could be used for coexpression, but their diVerent duplication speeds can give raise to unbalanced protein expression. Aimed to obtain comparable expression of both Gi1 and MT we have employed a multiple in vitro expression model, using a single high copy plasmid in which both the sequences have cloned. The eYciency of this system was evaluated by immunoprecipitation, considering as myristoylation index the binding of the two proteins. Gi1-mir/ ACV–C1–C2 complex molecular weight was about 50 kDa, a molecular weight diVerent from the sum of the single protein weights. Probably, the interaction itself can induce a diVerent protein folding that elicits a more compact quaternary structure. Moreover, considering that an environment analogous to plasmatic membrane could be a better condition for the G-protein folding and functionality, we created a reconstituted system in the presence of phospholipidic vesicles. The functional evaluation of the Gi1-mir chimeric protein in terms of GTP hydrolysis and GTPS binding highlighted that the presence of liposomes did not inXuence the Gi1-mir protein behaviour, whereas the stimulation by the well known activator ML250 was signiWcantly higher in the presence of phospholipids. Moreover, the same experiments conducted in presence of 4-aminopiperidinic derivative BC5 conWrmed our previous data, obtained with non-myristoylated i1 subunit which showed that ML250 is less eVective and potent that BC5 [20,21]. Finally, the inhibitory eVect of suramin on BC5 (and ML250) activity clearly conWrmed the speciWcity of the interaction between this molecule and G protein. Taken together these results demonstrated that Gi1mir reconstituted system is a good tool for evaluating new direct activators. Nevertheless, it was necessary to analyse the activity of this system in modulating the physiological Gi eVector, the adenylate cyclase (AC). For this purpose adenylate cyclase cytosolic fragments (C1–C2) were expressed and assessed for enzymatic activity. Basal cAMP production was comparable to literature kinetic data [25] and, interestingly, the presence of liposomes raised up to 80% the cAMP production, conWrming the role of phospholipides in improving protein functionality in a cellular-free system. The presence of AC fragments in this Gi1-mir system demonstrated that the Gi protein,
L.Di Cesare Mannelli et al. / Archives of Biochemistry and Biophysics 453 (2006) 151–160
activated with GTPS, was able to reduce cAMP production in a concentration dependent manner. Moreover, the system was sensitive to reference compounds, such us Mastoparan and ML250, both able to stimulate Gi1-mir with consequent reduction in cAMP levels. Mastoparan is a classical, powerful but poisonous, peptidic Gi activator, known to modulate GTP hydrolysis and GTPS binding of Gi subunits only in the presence of subunits and phospholipidic vesicles [21,27]. In our system, MP showed activity also in the absence of subunits and liposomes. Probably, the presence of AC induces a Gi1-mir conformation more functional in terms of MP interaction. The non-peptidic activator ML250 showed the same eYcacy of MP but a lower potency in the absence of liposomes. Hence, both MP and ML250 showed a higher potency in the eVector–transductor system in respect to Gi1-mir in the absence of AC. Also BC5-activated Gi1 was able to inhibit adenylate cyclase activity. In the absence of liposomes BC5 (5 £ 10¡7 M) induced a 39% inhibition of cAMP production, showing a higher eYcacy than Mastoparan and ML250. In the presence of phospholipids maximal inhibition was lower (22%), whereas potency was increased. Probably, the presence of phospholipidic vesicles induces a conformational change that increases the AC basal activity and reduces the eYcacy of the aminopiperidinic derivative. This liposomes eVect was observed also with the Gi protein activated by GTPS. On the contrary, MP was able to maintain the same eYcacy both in the absence and in the presence of phospholipids. One explanation could be that BC5 is a non-peptidic compound, thus less sensitive to structural rearrangement elicited by phospholipids. Finally, both reference compounds and BC5 exhibited a biphasic behaviour on Gi modulation: starting from a concentration of 1 £ 10¡9 up to a concentration of 5 £ 10¡6 M there was a proportional decrease of AC activity, whereas for concentration higher than 5 £ 10¡6 M the inhibitory eVect progressively decrease. This biphasic proWle could be due to many reasons, such as a not-speciWc eVect of activators on AC. Nevertheless, our experiments performed in the absence of Gi1-mir induced us to reject this hypothesis, at least in the examined concentration range. Moreover, at concentration higher than 5 £ 10¡5 M Mastoparan and ML250 stimulated AC, whereas BC5 inhibited the enzyme; thus, it can be excluded that the biphasic proWle, a common feature for all the activators, was due to a non-speciWc eVect on AC. In conclusion, our aim to develop strategies and tools that will allow the structural and physio-pathological investigations of Gi proteins was successful. The reconstitution of Gi1-mir with C1–C2 fragments in a transductor–eVector system is a valuable tool to study Gi direct activators and for potential high throughput assay to study biological signalling pathways. BC5 has demonstrated to be able to modulate intracellular signal by receptor-independent activation of Gi proteins, opening new perspectives on G proteins as drug target. Taken together this results suggest that
159
our cell-free system and direct activation provide new tools to clarify the molecular mechanisms in Gi protein signalling. Acknowledgments We are grateful to Dr. Martina Fortini, Dr. Debora Berti and Prof. Piero Baglioni for liposome preparation. These studies were supported by NIH Grants HL 59679 and GM 073181 to TBP. References [1] [2] [3] [4]
[5] [6] [7]
[8] [9] [10] [11] [12]
[13] [14] [15] [16] [17] [18] [19]
[20]
[21] [22] [23] [24] [25] [26] [27]
A.G. Gilman, Annu. Rev. Biochem. 56 (1987) 615–649. S.R. Neves, P.T. Ram, R. Iyengar, Science 296 (2002) 1636–1639. Y. Landry, J.P. Gies, Mini. Rev. Med. Chem. 2 (2002) 361–372. T.M. Cabrera-Vera, J. Vanhauwe, T.O. Thomas, M. Medkova, A. Preininger, M.R. Mazzoni, H.E. Hamm, Endocr. Rev. 24 (2003) 765– 781. S.M. Lanier, Biol. Cell. 96 (2004) 369–372. A.M. Spiegel, L.S. Weinstein, Annu. Rev. Med. 55 (2004) 27–39. J. Lyons, C.A. Landis, G. Harsh, L. Vallar, K. Grunewald, H. Feichtinger, Q.Y. Duh, O.H. Clark, E. Kawasaki, H.R. Bourne, et al., Science 249 (1990) 655–659. E.A. Williamson, P.G. Ince, D. Harrison, P. Kendall-Taylor, P.E. Harris, Eur. J. Clin. Invest. 25 (1995) 128–131. B.B. Lerman, B. Dong, K.M. Stein, S.M. Markovitz, J. Linden, D.F. Catanzaro, J. Clin. Invest. 101 (1998) 2862–2868. N. Galeotti, C. Ghelardini, M. Zoppi, E. Del Bene, L. Raimondi, E. Beneforti, A. Bartolini, J. Rheumatol. 28 (2001) 2298–2304. N. Galeotti, C. Ghelardini, M. Zoppi, E. Del Bene, L. Raimondi, E. Beneforti, A. Bartolini, Cephalalgia 21 (2001) 38–45. K.L. Laugwitz, A. Allgeier, S. OVermans, K. Spicher, J. Van Sande, J.E. Dumont, G. Schultz, Proc. Natl. Acad. Sci. USA 93 (1996) 116– 120. S.A. Akhter, L.M. Luttrell, H.A. Rockman, G. Iaccarino, R.J. Lefkowitz, W.J. Koch, Science 280 (1998) 574–577. C. Höller, M. Freissmuth, C. NanoV, Cell. Mol. Life. Sci. 55 (1999) 257–270. B. Nürnberg, W. Tögel, G. Krause, R. Storm, E. Breitweg-Lehmann, W. Schunack, Eur. J. Med. Chem. 34 (1999) 5–30. W.W. Ja, R.W. Roberts, Trends Biochem. Sci. 30 (2005) 318–324. Y. Odagaki, N. Nishi, T. Koyama, Life Sci. 62 (1998) 1537–1541. C. Lesckhe, R. Storm, E. Breitweg, T. Exner, B. Nürnberg, W. Schunack, J. Med. Chem. 40 (1997) 3130–3139. E. Breitweg-Lehmann, C. Czupalla, R. Storm, O. Kudlacek, W. Schunack, M. Freissmuth, B. Nürnberg, Mol. Pharmacol. 61 (2002) 628–636. D. Manetti, L. Di Cesare Mannelli, S. Dei, N. Galeotti, C. Ghelardini, M.N. Romanelli, S. Scapecchi, E. Teodori, A. Pacini, A. Bartolini, F. Gualtieri, J. Med. Chem. 48 (2005) 6491–6503. L. Di Cesare Mannelli, A. Pacini, A. Toscano, M. Fortini, D. Berti, C. Ghelardini, N. Galeotti, P. Baglioni, A. Bartolini, 47 (2006) 303–310. C. Wittpoth, K. Scholich, Y. Yigzaw, T.M. StringWeld, T.B. Patel, Proc. Natl. Acad. Sci. USA 96 (1999) 9551–9556. C.W. Dessauer, M. Chen-Goodspeed, J. Chen, J. Biol. Chem. 277 (2002) 28823–28829. M.E. Linder, I.H. Pang, R.J. Duronio, J.I. Gordon, P.J. Sternweis, A.G. Gilman, J. Biol. Chem. 266 (1991) 4654–4659. T.B. Patel, C. Wittpoth, A.J. Barbier, Y. Yigzaw, K. Scholich, Methods Enzymol. 345 (2002) 160–187. M.J. Hope, M.B. Bally, G. Webb, P.R. Cullis, Biochim. Biophys. Acta 812 (1985) 55–65. C. Oppi, T. Wagner, A. Crisari, B. Camerini, G.P. Tocchini Valentini, Proc. Natl. Acad. Sci. USA 89 (1992) 8268–8272.
160
L.Di Cesare Mannelli et al. / Archives of Biochemistry and Biophysics 453 (2006) 151–160
[28] R. Taussig, J.A. Iniguez-Lluhi, A.G. Gilman, Science 261 (1993) 218– 221. [29] S.M. Mumby, R.O. Heuckeroth, J.I. Gordon, A.G. Gilman, Proc. Natl. Acad. Sci. USA 87 (1990) 728–732. [30] S.M. Mumby, M.E. Linder, Meth. Enzymol. 237 (1994) 254–268.
[31] C.E. Stebbins, W.G.T. Jr Kaelin, N.P. Pavletich, Science 284 (1999) 455–461. [32] J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular Cloning: A Laboratory Manual, 2nd ed., Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1989.
