Analysis of a putative voltage-gated prokaryotic potassium channel.

Most of the completely sequenced prokaryotic genomes contain genes of potassium channel homologues, but there is still not much known about the role of these proteins in prokaryotes. Here we describe the large-scale overproduction and purification of a prokaryotic voltage-gated potassium channel homologue, Kch, from Escherichia coli. After successful overproduction of the protein, a specific increase in the potassium permeability of the cells was found. Kch could be purified in large amounts using classical purification methods to prevent aggregation of the protein. The physiological state of the protein was revealed to be a homotetramer and the protein was shown to be localized to the cytoplasmic membrane of the cells. In the course of the localization studies, we found a specific increase in the density of the cytoplasmic membrane on Kch production. This was linked to the observed increase in the protein to lipid ratio in the membranes. Another observed change in the membrane composition was an increase in the cardiolipin to phosphatidylglycerol ratio, which may indicate a specific cardiolipin requirement of Kch. On the basis of some of our results, we discuss a function for Kch in the maintenance of the membrane potential in E. coli.

Solubilization and purification of the human ETB endothelin receptor produced by high-level fermentation in Pichia pastoris.

In the present report, the successful solubilization and purification of the ETB receptor heterologously produced in the methylotrophic yeast P. pastoris is described for the first time. In comparison to the baculovirus system where successful production, solubilization and purification have already been reported, handling and up-scaling of recombinant P. pastoris cells was much easier and less time consuming. Recombinant P. pastoris clones producing two different ETB receptor constructs were grown in a fermenter to a density of about 360 g/l. After induction with methanol, a production level of maximally 45 pmol/mg was obtained, a value which is in the range of that reported for baculovirus-infected insect cells. A method for the large-scale preparation of membranes was established. Solubilization of the recombinant ETB receptor was achieved with the detergent n-dodecyl-/beta-D-maltopyranoside. The stability of the solubilized and ligand-bound receptor was examined in detail. Subsequently, two purification methods for two different receptor constructs were tested and a large-scale procedure for isolation of recombinant receptor was established. In general, the purification methods described herein will be adaptable to other G protein-coupled receptors heterologously produced in heterologous expression systems including P. pastoris.

The human ET(B) endothelin receptor heterologously produced in the methylotrophic yeast Pichia pastoris shows high-affinity binding and induction of stacked membranes.

The human endothelin B receptor (ET(B) receptor) was produced in the methylotrophic yeast Pichia pastoris under transcriptional control of the highly inducible alcohol oxidase 1 (AOX1) gene promoter. In the expression plasmids pPIC9KFlagET(B)Bio and pPIC9KFlag deltaGPET(B)Bio the ET(B) receptor coding region was fused in frame to the Saccharomyces cerevisiae alpha-factor prepropeptide and the FLAG-tag. In both constructs, the receptor was also fused to a biotinylation domain. Additionally, in pPIC9KFlag deltaGPET(B)Bio the putative N-glycosylation site and a protease site have been deleted by site directed mutagenesis. Crude membranes prepared from recombinant P. pastoris revealed specific and saturable binding of [125I]ET-1 with a K(D) of about 42 pM. Receptor levels of 60 pmol/mg and 35 pmol/mg for the Flag deltaGPET(B)Bio and the FlagET(B)Bio construct, respectively, were determined. The pharmacological profile for ET-1, ET-2 and ET-3 were as expected for a subtype B endothelin (ET) receptor. Immunoblot analysis showed an apparent molecular mass of 55 kDa for the Kex2-processed and about 74 kDa for the Kex2-unprocessed receptor. Contrary to the Flag deltaGPET(B)Bio construct, the FlagET(B)Bio construct was not correctly processed by the internal Kex2 endopeptidase. As was detected by ultrastructural analysis of recombinant yeast cells, high-level production of the receptor resulted in the formation of stacked membranes.

Residues within transmembrane helices 2 and 5 of the human gonadotropin-releasing hormone receptor contribute to agonist and antagonist binding.

To understand the ligand binding properties of the human GnRH receptor (hGnRH-R), 24 site-specific mutants within transmembrane helices (TMH) 1, 2, and 5 and the extracellular loop 2 (E2) were generated. These mutants were analyzed by using a functional reporter gene assay, monitoring receptor signaling via adenylate cyclase to a cAMP-responsive element fused to Photinus pyralis luciferase. The functional behavior of 14 receptor mutants, capable of G-protein coupling and signaling, was studied in detail with different well described agonistic and antagonistic peptide ligands. Furthermore, the binding constants were determined in displacement binding experiments with the antagonist [125I]Cetrorelix. The substitution of residues K36, Q204, W205, H207, Q208, F20, F213, F216, and S217 for alanine had no or only a marginal effect on ligand binding and signaling. In contrast, substitution of N87, Eg9, D9, R179, W206, Y211, F214, and T215 for alanine resulted in receptor proteins neither capable of ligand binding nor signal transduction. Within those mutants affecting ligand binding and signaling to various degrees, W101A, N102A, and N212Q differentiate between agonists and antagonists. Thus, in addition to N102 already described, the residues W101 in TMH2 and N212 in TMH5 are important for the architecture of the ligand-binding pocket. Based on the experimental data, three-dimensional models for binding of the superagonist D-Trp6-GnRH (Triptorelin) and the antagonist Cetrorelix to the hGnRH-R are proposed. Both decapeptidic ligands are bound to the receptor in a bent conformation with distinct interactions within the binding pocket formed by all TMHs, E2, and E3. The antagonist Cetrorelix with bulky hydrophobic N-terminal amino acids interacts with quite different receptor residues, a hint at the failure to induce an active, G protein-coupling receptor conformation.

Engineering of a proteolytically stable human beta 2-adrenergic receptor/maltose-binding protein fusion and production of the chimeric protein in Escherichia coli and baculovirus-infected insect cells.

The hydrophobic human beta 2 adrenergic receptor was produced in fusion to the hydrophilic maltose-binding protein (MalE) in Escherichia coli. Photoaffinity labeling with the adrenergic ligand [125I]cyanopindolole-diazirine indicated that the majority of the protein was proteolyzed in the intergenic region between the fusion partners after production in E. coli. The simple and fast genetics of the bacterium enabled us to engineer a linker with an increased proteolytic stability. The fusion protein produced in E. coli was fully functional with respect to binding of adrenergic ligands and coupling to stimulatory GTP-binding protein. The production level with 3 pmol receptor fusion protein per mg membrane protein in a crude membrane preparation was significantly higher than those reported for other beta 2 adrenergic receptor constructs in E. coli. After solubilization with dodecanoyl sucrose, the fusion protein was purified to near homogeneity by affinity chromatography on immobilized Ni2+ ions (binding to a C-terminal His6-tag) and on crosslinked amylose (binding to the MalE). In order to achieve higher production levels, the fusion protein preceded by an insect signal peptide was produced in baculovirus-infected insect cells. As expected, the production level with about 17 pmol receptor per mg membrane protein was higher in the insect cells than in E. coli. The receptor fusion protein produced in the insect cells bound adrenergic ligands and activated heterotrimeric GTP-binding proteins with biochemical properties comparable to that of the unfused receptor.

High-level secretion of biologically active recombinant human macrophage inflammatory protein-1alpha by the methylotrophic yeast Pichia pastoris.

The human CC chemokine macrophage inflammatory protein-1alpha (MIP-1alpha) was produced at a high level in Pichia pastoris under transcriptional control of the highly inducible alcohol oxidase 1 promoter. To ensure proper folding and secretion of the recombinant polypeptide, the MIP-1alpha gene had been fused to the Saccharomyces cerevisiae alpha-factor prepropeptide. As was revealed by analysis of the cell culture supernatant of recombinant Pichia pastoris, MIP-1alpha was efficiently secreted. Immunoblot analysis of secreted proteins from recombinant clones using a polyclonal antibody directed against MIP-1alpha revealed an apparent molecular mass of 8 kDa for the recombinant polypeptide. Up to 70 mg of MIP-1alpha was purified from 1 liter of yeast culture supernatant by a single chromatography step. Biological activity of recombinant MIP-1alpha was shown in a chemotaxis assay. Here, the polypeptide specifically induced migration of U937 cells expressing the CCR1 (MIP-1alpha receptor). Also, in competition binding assays the recombinant MIP-1alpha displayed high affinity binding.

Production of G-protein-coupled receptors in yeast.

Yeasts combine the advantages of fast and easy handling with the potential to perform eukaryotic post-translational modifications and are for this reason interesting hosts for heterologous production of G-protein-coupled receptors. The possibility to connect foreign receptors to a yeast internal MAP kinase pathway was used to establish yeast-based systems for high-throughput screening of compound libraries. In addition, yeasts have the potential for high level production of G-protein-coupled receptors. In this field, non-Saccharomyces yeasts seems to be interesting alternatives to S. cerevisiae, as well as to systems based on higher eukaryotic cells.

Characterization of the human gonadotropin-releasing hormone receptor heterologously produced using the baculovirus/insect cell and the Semliki Forest virus systems.

1. Two eukaryotic viral systems, the baculovirus/insect cell and the Semliki Forest virus systems, were tested for heterologous expression of human gonadotropin-releasing hormone receptor (GnRHR) cDNA. 2. An unmodified as well as a c-myc epitope-tagged human GnRH receptor was produced in two insect cell lines (Spodoptera frugiperda, Trichoplusia ni) after infection with the respective recombinant baculoviruses. In both insect cell lines, the receptor was identified by immunoblot analysis as a triplet of bands between 35 and 40 kDa. After deglycosylation of the receptor the molecular mass decreased to 35 kDa. The GnRH receptor was localized in membrane compartments within the infected insect cells. However, only in membranes of infected Trichoplusia ni insect cells could approximately 2000 receptors per cell be detected. 3. Production of the GnRH receptor in BHK cells using the Semliki Forest virus system resulted in approximately 50,000 receptors per cell. A maximal yield of 0.42 pmol/mg membrane protein was obtained 24 hr after electroporation of BHK cells with in vitro synthesized RNA. Binding of the antagonist [125I]Cetrorelix was saturable with a KD of 1.3 nM. The receptor produced in the BHK cells was further characterized by ligand displacement studies. The rank order of agonist and antagonist affinities was Cetrorelix > Triptorelin > Antide > GnRH.

Expression and initial characterization of a soluble glycine binding domain of the N-methyl-D-aspartate receptor NR1 subunit.

Glycine is an essential co-agonist of the excitatory N-methyl-D-aspartate (NMDA) receptor, a subtype of the ionotropic glutamate receptor family. The glycine binding site of this hetero-oligomeric ion channel protein is formed by two distinct extracellular regions, S1 and S2, of the NR1 subunit, whereas the homologous domains of the NR2 subunit mediate glutamate binding. Here, segments S1 and S2 of the NR1 polypeptide were fused via a linker peptide followed by N- and C-terminally tagging with Flag and His6 epitopes, respectively. Infection of High Five insect cells with a recombinant baculovirus containing this glycine binding site construct resulted in efficient secretion of a soluble fusion protein of about 53 kDa. After affinity purification to near-homogeneity, the fusion protein bound the competitive glycine site antagonist [3H]MDL105,519 with high affinity (Kd = 5.22 +/- 0. 13 nM) similar to that determined with rat brain membrane fractions. This high affinity binding could be competed by the glycine site antagonist 7-chlorokynurenic acid as well as the agonists glycine and D-serine but not by L-glutamate. This indicates that the S1 and S2 domains of the NR1 subunit are sufficient for the formation of a glycine binding site that displays pharmacological properties similar to those of the NMDA receptor in vivo.

Comparative biochemical and pharmacological characterization of the mouse 5HT5A 5-hydroxytryptamine receptor and the human beta2-adrenergic receptor produced in the methylotrophic yeast Pichia pastoris.

Over the last few years, Pichia pastoris has been developed into a powerful expression system for a multitude of foreign genes. Here, we demonstrate that the P. pastoris expression system has similar power to the baculovirus expression system in high-level production of two G-protein-coupled receptors, the mouse 5HT5A 5-hydroxtryptamine receptor and the human beta2-adrenergic receptor. Different expression plasmids were constructed in which the cDNAs of the two receptors were cloned under the transcriptional control of the highly inducible promoter of the P. pastoris alcohol oxidase 1 (AOX1) gene. In three expression plasmids, the receptors were fused to the Saccharomyces cerevisiae alpha-factor prepropeptide and also to the c-myc tag or the FLAG tag to permit immunological detection of the receptors. After transformation into P. pastoris strains KM71 and SMD 1163, recombinant clones were selected and tested for the production of the 5HT5A receptor and the beta2-adrenergic receptor by radioligand binding using [N-methyl-3H]lysergic acid diethylamide and [5,7-3H](-)CGP-12177 respectively. The production level of the 5HT5A receptor was improved by a factor of three by fusion with the alpha-factor prepropeptide. Also, the higher gene dosage resulting from multiple insertions of the expression cassette led to an improvement in production by a factor of two for both receptors. The addition of the adrenergic antagonist alprenolol to the culture medium had a positive effect on the number of specific binding sites detectable in clones producing the beta2-adrenergic receptor. For the 5HT5A receptor the addition of yohimbine resulted in a similar but smaller effect. Binding assays revealed that approx. 25 pmol of beta2-adrenergic receptor and approx. 40 pmol of 5HT5A receptor per mg of membrane protein in crude membrane preparations were produced. The pharmacological profiles for the heterologously produced receptors, estimated by ligand-displacement analysis using certain adrenergic and serotoninergic agonists and antagonists, were comparable with those reported for the receptors expressed in mammalian systems. Immunoblot analysis of the 5HT5A receptor revealed an apparent molecular mass about 20 kDa higher than expected from the amino acid sequence. Here, the Kex2 endopeptidase failed to process the alpha-factor leader correctly. Blocking glycosylation in vivo by tunicamycin or in vitro deglycosylation of membranes by endoglycosidase H resulted in correct processing. In contrast, the beta2-adrenergic receptor fusion to the alpha-factor leader was correctly processed by the internal Kex2 endopeptidase. The Kex2-processed beta2-adrenergic receptor was not glycosylated. In conclusion, the high-level production of the two receptors in P. pastoris will allow their purification in quantities sufficient for subsequent biophysical and structural studies.

Characterization of the human dopamine transporter heterologously expressed in BHK-21 cells.

1. cDNA of the human dopamine transporter (hDAT) was cloned into a cloning vector based on the Semliki Forest virus. Electroporation of in vitro transcribed mRNA from this plasmid into BHK-21 cells resulted in production of the transporter as measured by [3H]dopamine uptake (Km = 2.0 +/- 0.4 microM), which was specifically inhibited in the presence of cocaine. 2. The recombinant transporter protein exhibited an apparent molecular mass of 56 kDa, which was reduced to 50 kDa after tunicamycin treatment of the producing BHK-21 cells. Tunicamycin treatment of the electroporated cells also resulted in a decrease in transport activity with no change in the Km value (2.1 +/- 0.4 microM). 3. The localization of the heterologously produced transporter in the BHK cells either with or without tunicamycin treatment was studied by electron microscopic immunogold staining. The glycosylated transporter was found to be localized at the plasma membrane, whereas in the case of the unglycosylated transporter, transport to the plasma membrane was blocked.

Lipid composition of Spodoptera frugiperda (Sf9) and Trichoplusia ni (Tn) insect cells used for baculovirus infection.

The lipid composition of two different insect cell lines from Spodoptera frugiperda (Sf9) and Trichoplusia ni (Tn) which are established cell lines for infection with recombinant baculovirus was analyzed by high-performance liquid chromatography and gas-liquid chromatography. The major phospholipids found were phosphatidylcholine and phosphatidylethanolamine, the major mono-unsaturated fatty acids were oleic acid and palmitoleic acid, the major saturated fatty acid was stearic acid. The cholesterol to phospholipid ratio was demonstrated to be lower than in mammalian cell lines. Infection with a recombinant baculovirus Autographa californica resulted in increased levels of phosphatidylcholine in the insect cells. The baculovirus/insect cell system has become a popular system for heterologous protein production. Functional changes of membrane proteins produced in these two cell lines might be correlated to a different lipid profile of their cellular membranes.

Probing human beta1- and beta2 -adrenoceptors with domain-specific fusion protein antibodies.

In order to generate antibodies suitable for immunological studies on beta-adrenoceptors constitutively expressed at low levels in cells or tissues we have produced fusion proteins of the amino- and carboxy-terminus, and the second extracellular loop of the human beta1- or beta2-adrenoceptors with bacterial glutathione-S-transferase in E. coli. Rabbit antibodies raised against these fusion proteins strongly reacted with intact human beta1- or beta2-adrenoceptors in a subtype- and domain-specific manner. Antibodies directed against the second extracellular loop of the beta1-adrenoceptor reacted stronger with non-denatured receptors and decreased the affinity of the 3H-labelled antagonist (-)-4-(3-t-butylamino-2-hydroxypropoxy)-[5,7-3H]benzimidazol-2-one ([3H]CGP 12 177), indicating a specific interaction with the native receptor. In contrast, antibodies directed against carboxy- and amino-terminal receptor domains reacted strongly both with denatured and non-denatured receptors but did not interfere with binding of [3H]CGP 12 177. Affinity purified antibodies were used for detecting the beta1- or the beta2-adrenoceptor subtype heterologously produced in Sf9 cells by enzyme-linked immunosorbent assay, Western blotting, immunoprecipitation, and indirect immunofluorescence microscopy. Moreover, we could demonstrate that avidity, titers, and specificity of these antibodies were high enough for studying beta-adrenoceptors constitutively expressed in human A431 cells, where we observed a patched membrane distribution of the receptors.

Engineering the folding pathway of insect cells: generation of a stably transformed insect cell line showing improved folding of a recombinant membrane protein.

The baculovirus-insect cell expression system has proven to be a valuable tool for the high level production of a multitude of recombinant proteins. However, production of membrane proteins in infected insect cells is often hampered by incorrect folding and processing which results in the accumulation of non-functional protein. Here, we report the construction of a Sf9 insect cell line stably transformed with the ninaA gene from D. melanogaster (Sfn cell line). The ninaA protein is a membrane bound cyclophilin which acts as a peptidyl-prolyl cis/trans isomerase during the folding process of rhodopsin 1 in D. melanogaster rhabdomere. Engineered Sfn insect cells infected with a recombinant baculovirus bearing the human dopamine transporter gene under the control of the polyhedrin promoter showed a > or = 5 times enhanced uptake of [3H]dopamine in comparison to similarly infected Sf9 cells. This increase in specific transport activity was not due to an altered Km value in the Sfn cell line. The uptake in infected Sfn cells was blocked by the peptidyl-prolyl cis/trans isomerase inhibitor cyclosporin A which had no effect on infected Sf9 cells. From these results we conclude that the prolyl-cis/trans isomerase activity of the ninaA in the stably transformed Sfn cell line was responsible, directly or indirectly, for the improved folding of the heterologously produced human dopamine transporter.

Characterization of gonadotropin-releasing hormone analogs based on a sensitive cellular luciferase reporter gene assay.

A novel cellular assay for the functional characterization of agonistic and antagonistic analogs of gonadotropin-releasing hormone (GnRH) was developed. This assay is based on a fusion of the c-fos immediate-early gene promoter to Photinus pyralis luciferase (Luc) as a reporter gene, stably transfected in a recombinant cell line expressing the human GnRH receptor. Transcription of endogenous c-fos and fos-Luc fusion gene are transiently induced quite similar by fetal calf serum or the superagonistic analog [D-Trp6] GnRH in a selected cell line. The reporter gene was therefore used to monitor agonist-induced signaling via the human GnRH receptor. Whereas Luc activity was induced in a dose-dependent manner by GnRH or [D-Trp6] GnRH, different antagonistic peptides completely inhibited this stimulation. The antagonistic potency (IC50) of various peptides with Cetrorelix and Antarelix as lead compounds in general correlated well with the binding affinity (KD) as determined from ligand binding experiments. The specificity of an inhibitory effect was confirmed by GnRH receptor-independent stimulation with the phorbol ester 12-O-tetradecanoylphorbol 13-acetate or basic fibroblast growth factor. Since this new reporter gene assay is sensitive and simple and can be performed in a microtiter plate, it will significantly facilitate screening and functional characterization of GnRH analogs.

Glycosylation, palmitoylation, and localization of the human D2S receptor in baculovirus-infected insect cells.

In order to evaluate the baculovirus expression system as a means for high-yield production of homogeneous D2S receptor, we have expressed various D2S receptor constructs in two Spodoptera frugiperda cell lines, a Trichoplusia ni and a Mammestra brassicae cell line. To improve expression yield, the environment of the polyhedrin gene translational initiation site was retained by fusing the first 12 codons of the polyhedrin gene to the 5'-end of the D2S receptor coding sequence. The pharmacological profile of the expressed D2S receptor was similar to that reported for neuronal D2 receptors. Sf9 and Tn cells were best suited for overexpression, yielding about 2 x 10(6) and 4 x 10(6) receptors/cell, respectively, corresponding to 6 pmol/mg of cell protein in Sf9 cells and 10 pmol/mg of cell protein in Tn cells. We have developed a D2 receptor-specific anti-peptide antibody to study glycosylation, palmitoylation, and localization of the heterologously produced receptor. Immunoprecipitation of digitonin/cholate-solubilized receptor from control and tunicamycin-treated Sf9, Tn, and Mb cells revealed an apparent molecular mass of 47-48 kDa for the glycosylated receptor and of 39-40 kDa for the unglycosylated receptor. Although pulse-chase studies showed that glycosylation occurred rapidly and efficiently, the glycosylated receptor only constituted a small fraction of the overall produced receptor protein, which was mainly located intracellularly. The glycosylation of the receptor was of the high-mannose-type in contrast to the complex-type glycosylation found in native tissue. The glycosylated D2S receptor was palmitoylated. Glycosylation, however, was not a prerequisite for palmitoylation which was insensitive to tunicamycin, brefeldin A, and monensin. NH2-terminal addition of the signal sequence of prepromelittin to the D2S receptor increased expression levels 2-3-fold and significantly enhanced membrane insertion and processing, resulting in increased targeting of the synthesized receptor to the plasma membrane.

In vivo reconstitution of dopamine D2S receptor-mediated G protein activation in baculovirus-infected insect cells: preferred coupling to Gi1 versus Gi2.

Agonist binding of the human D2S receptor overexpressed in baculovirus-infected Sf9 insect cells was of low affinity and GppNHp-insensitive, yet, dopaminergic agonists were able to partly inhibit forskolin-stimulated cAMP accumulation. In order to prove full functionality of the receptor, we used an "in vivo" reconstitution system, which is based on coinfection of Sf9 cells with the appropriate receptor and G protein encoding baculoviruses. In cells coexpressing the D2S receptor and either Gi1 or Gi2, the dopaminergic agonist apomorphine effectively stimulated [35S]GTP gamma S binding and GTPase activity. Agonist-stimulated [35S]GTP gamma S binding was dependent on the ratio of G protein to receptor. Expression levels of receptor and G protein influenced each other reciprocally. G protein activation could be optimized by varying the multiplicity of infection of the receptor and G protein encoding baculoviruses. Coexpression of either Gi1 or Gi2 led to the appearance of GppNHp-sensitive high-affinity agonist binding. Detailed agonist competition binding analysis revealed that the percentage of high-affinity agonist binding sites was significantly higher in D2S receptor-expressing cells coinfected with Gi1 viruses than when coinfected with Gi2 viruses. Moreover, the coexpressed Gi proteins seemed to modulate the affinity of agonists for the high-affinity form of the receptor. In cells coexpressing Gi1, agonist high affinity was 2-4-fold higher than in cells coexpressing Gi2. Na+ increased the dissociation constant of apomorphine for the high-affinity site by 2-4-fold without affecting the percentage of high-affinity sites or the preference for Gi1. In some dopamine competition experiments with coinfected cells, displacement data were best fit assuming three noninteracting classes of sites in the absence and two independent classes of sites in the presence of GppNHp. Dopamine competition curves with cells highly overexpressing the D2S receptor or with membranes from such cells were best fit assuming two independent classes of sites which were insensitive to GppNHp and might reflect abnormal compartimentalization and/or different states of aggregation.