Patent Application
20080166748
The present invention relates to a method of identifying protein CAMs and the use thereof.
The G protein-coupled receptors (GPCRs) constitute the largest family of membrane receptors with a common evolutionary origin. They include receptors which respond to environmental ligands (odorants, flavors) or radiations (light of various wavelengths) and to innumerable internal signals (hormones, bioactive amines, neuropeptides, arachidonic acid metabolites, purines, etc . . . ). This extreme diversity contrasts with their stereotyped structure (seven transmembrane alpha helices, three extracellular loops, three intracellular loops, amino terminus outside and carboxyl terminus inside the cell) and with the limited number of downstream regulatory cascades they control.
The GPCR interacts with an intracellular heterotrimeric G protein consisting of αβγ subunits. Upon binding of the receptor's ligand, the α-subunit dissociates from the β-and γ-subunits, and hydrolyses GTP to GDP. Both Gα and Gβγ can then activate downstream transduction effectors or regulate other receptors (Zwick et at., 1999). In haploid Saccharomyces cerevisiae cells, GPCRs are regulating the mating process. The receptor (Step 2 or Step 3) detects the presence of cells of the opposite mating type (through binding of peptide mating pheromones), and activates intracellular heterotnmeric G proteins, thus initiating the mating process. Gpa1 (α subunit) dissociates from the βγ (Step4-Ste18) complex which activates downstream elements of the pheromone response pathway which includes a well-characterized mitogen-activated protein kinase (MAP kinase) cascade. The transcription factor Ste12 can then initiate the transcription of several mating factor-inducible genes such as FUS1.
Reports of mammalian GPCRs expressed in yeast indicate that these heterologous proteins can be reliably expressed in yeast and properly inserted into yeast membranes (Tate and Grisshammer, 1996). Reports, e.g. (Price et al., 1995), demonstrate that a large number of heterologous GPCRs interact with the yeast heterotrimeric G protein with sufficient efficacy to induce a growth-promoting signal. In case the GPCR under investigation does not couple to Gpa1, it is co-expressed together with a chimera where the C-terminal part of Gpa1 is replaced by the corresponding amino acids of a given human Gα subunit (Brown et al., 2000). A reporter construct (such as pFUS1-HIS3 or pFUS1-lacZ) is then expected to produce a detectable response upon receptor activation.
Increasingly, it is being appreciated that endogenous receptors, and in particular those of the G protein-coupled receptor family, may possess some level of constitutive activity even in the absence of activating mutation. A potentially important physiological ramification of the constitutive activity of such receptors is that the ability of different receptor subtypes (for the same ligand) to spontaneously isomerize to the active state might well differ. Such receptor subtypes would vary substantially in their properties, thus best suiting them for one or another physiological context (Lefkowitz et al., 1993).
The discovery of constitutive GPCR activity presents a theoretical approach to the identification of ligands for orphan receptors. The basic premise for this idea is that different tertiary conformations (i.e. different allosteric change states) of the receptor protein will display different binding domains for ligands, or different binding affinities for the same ligand. Since the mutation of a receptor sequence can only affect the physico-chemical properties of the receptor, but not those of ligands, a change of affinity of a ligand for a receptor ought to be of a similar magnitude for all ligands and not proportional to the ligand's efficacy (Lefkowitz et al., 1993).
The notion that constitutive activation of G protein-coupled receptors could be responsible for hereditary diseases came first from the study of patients suffering of retinis pigmentosa (Robinson et al. 1992). Since then, several other human pathologies have been linked to constitutive activity or aberrant receptors
(Dhanasekaran et al., 1995) (Rao and Oprian, 1996) (Duprez et al., 1997) (Jensen et al., 2000). It has now been recognized that very valuable information relevant to treatment of diseases caused by constitutively active receptors (for instance TSH and LH receptors (Spiegel, 1996)) can be directly obtained by identifying compounds which act as inverse agonists to constitutively activated forms of the receptor.
Additionally, some of the therapeutic effects of presently used receptor antagonists may be related to their inverse agonist properties. Recent results (Varma et al., 1999) show that almost all β-adrenergic antagonists (with the exception of pindolol) have inverse agonist properties in the heart of reserpine treated rats.
Down regulation and desensitization of GPCRs is elicited by agonists or as a consequence of spontaneous activity. Thus, inverse agonists upregulate heptahelical receptors by decreasing spontaneous downregulation (Daeffler and Landry, 2000), offering new approaches to tolerance and dependence to drugs. Amino acid residues can be mutated and lead to ligand-independent activation of the receptor and constitutive activation of signaling pathway (Lefkowitz et al., 1993) (Rao and Oprian, 1996) (Sommers et al., 2000) (Konopka et al., 1996) (Alewijnse et al., 2000).
Several techniques have been applied to the discovery or the study of constitutively active receptors. For instance, manipulation of the stoichiometry of receptors and G proteins (mainly over-expression of the receptor) can create a constitutive active receptor system (Chen et al., 2000) (Samama et al., 1997) or site directed mutagenesis on residues such as the highly conserved DRY motif were found to be involved in stabilizing intramolecular interactions (Alewijnse et al., 2000). To date, random mutagenesis has been used in many works to identify CAMs: in a random saturation mutagenesis of a critical region of the Calcium-sensing receptor (Jensen et al., 2000) or in a systematic screening in a mammalian cell based bioassay of a random mutant library of the angiotensin II AT1A receptor (Parnot et al., 2000).
The ease of genetic manipulation of yeast and the availability of an assay that allows detection of a signaling activity made it possible to search through large random mutational libraries to study the spectrum of mutations capable of causing constitutive activation. Ma & al. (Ma et al., 1987) described a fast and reliable method for plasmid construction (Gap Repair) that is based on the efficient repair of a linearized plasmid by recombination with a homologous DNA restriction fragment during yeast transformation.
Additionally, S. cerevisiae does not show such a rapid desensitization process comparable to the ligand-dependent phosphorylation of receptors followed by receptor interaction with arresting, disruption of the interaction receptor-G protein, and in some case sequestration (Tsao et al., 2001). This makes the identification of a constitutive activity much easier.
Previously, a random mutagenesis strategy combined with a yeast based in vivo sub-cloning/screening has been applied successfully on the amino-terminal and transmembrane regions (approximately the first 300 out of 431 residues) of the yeast Step 2 G protein-coupled receptor (Sommers et al., 2000) and on the second intracellular loop of the V2 Vasopressin receptor (Erlenbach et al., 2001). In contrast, the present method allows the systematic identification of activating mutations over the whole open reading frame, without the need of focusing on some regions. Although people usually choose to mutagenise only a part of the coding sequence because they believe that only this region is involved in the mechanism studied (Erlenbach et al., 2001), no doubt would persist and sometimes new prospects on structure-activity would appear.
WO 00/12705 discloses methods for improving the function of heterologous G protein-coupled receptors.
Random mutagenesis on human GPCRs and functionally studied in mammalian cells, were described by (Parnot et al., 2000)) (CAM discovery of Angiotensin II 1A receptor, full length) and (Jensen et al., 2000) (Functional Importance of the Ala116-Pro136 Region in the Calcium-sensing Receptor).
Random mutagenesis of a yeast GPCR and functionally studied in yeast cells, in particular CAM discovery of Step 2 (α-factor receptor), random mutagenesis of amino terminal and transmembrane regions, including Gap Repair were described by (Sommers et al., 2000) and (Sommers and Dumont, 1997)).
Random mutagenesis on a human GPCR functionally studied in yeast cells, in particular coupling properties study of V2 vasopressin receptor, oligonucleotide-directed random mutagenesis of the intracellular loop 2 (228 bp), including Gap repair were described by (Erlenbach et al., 2001)).
CAMs and methods of using them are also disclosed in WO 00/121987. WO 00/06597 discloses endogenous constitutively activated G protein-coupled orphan receptors. WO 00/22129 and WO 00/22131 disclose non-endogenous constitutively activated human G protein-coupled orphan receptors (site directed mutagenesis of GPCRs to generate constitutively activated mutants) and WO 97/21731 an assay for and uses of peptide hormone receptor ligands.
The discovery of constitutively activated mutants (CAMs) is usually the result of a long process of genetic manipulations and assays in mammalian cell culture. Researchers usually choose site directed mutagenesis because of its more straight forward and fast principle (Egan et al., 1998; Alewijnse et al., 2000).
It was a task of the present invention to provide an easy and fast method for identifying CAMs of proteins, e.g. for GPCRs, ion-channel, enzymes.
The present invention provides a method for identifying protein CAMs (constitutively active mutants), wherein
a) a library of mutated sequences of a protein is generated,
b) yeast cells are transformed with such library and
c) the respective protein CAM is identified.
Examples for proteins for which CAMs can be identified are GPCRs, ion-channels, enzymes, e.g. kinases, proteases, transcription factors.
Preferably, protein CAMs of mammalian proteins are identified, e.g. CAMs of human proteins.
The present invention provides a method of identifying protein CAMs (constitutively active mutants) wherein
The present invention provides a method of identifying protein CAMs (constitutively active mutants) wherein
In a special embodiment of the invention low fidelity PCR is applied on a full length sequences of a particular protein, e.g. a GPCR, preferably a mammalian protein sequence. The PCR products were co-transformed with a linearized expression vector (e.g. containing at each end short sequences homologous to the end of the PCR product) into an engineered yeast strain. The transformed yeast cells were first selected for the repair of the plasmid (e.g. selection by colony forming on a selective medium). The colonies previously selected were replicated an another medium, selective for the activity of the protein, e.g. the receptor (e.g. by the use of a survival reporter gene expressed only upon receptor signaling). Preferably, three or more identical and independent experiments were done to avoid the PCR's bias. The protein CAMs (“mutants”) have an increased basal signaling activity and the same maximum of stimulation than the wild type protein.
A yeast based in vivo discovery of random active mutants can be applied to the entire coding sequence of a protein, e.g. a human receptor. This was done by screening for constitutive mutations of the human sphingosine 1-phosphate receptor EDG5 (Endothelial Differentiation Gene 5) (An et al., 2000) (HIa, 2001).
To obtain a whole set of single point mutants directly (and also to avoid excessive secondary sub-cloning work to find out the activity conferred by each single point mutation), the PCR protocol was optimized to induce an average of less than one point mutation per copy of the gene. Indeed, the high throughput potential of an in vivo subcloning/screening strategy allows us to increase the size of the library without consuming more time/money.
In another embodiment the present invention relates to engineered yeast cells comprising a library of mutants (e.g. GPCR CAMs) and the use of such engineered yeast cells.
For such engineering for example Saccharomyces cerevisiae, Schizosaccharomyces pombe and Candida albicans cells can be used.
The use of such an engineered yeast cell should bring three major improvements at the same time:
All three in the same engineered yeast cell.
Further advantages are, that the yeast is a powerful tool for the study of mammalian GPCRs and their transduction characteristics because of the high homology between these eukaryotic cells (Price et al., 1995; Hadcock and Pausch, 1999; Botstein et al., 1997); Yeast has a high rate of homologous recombination and the genetic manipulations of yeast are easy (Ma et al., 1987; Oldenburg et al., 1997); Yeast allows in vivo selection of a receptors activity (Chambers et al., 2000); In vivo screen allows the direct recovery of the plasmid carrying the mutant of interest (CAM) from the microorganism. Yeast is cheaper to cultivate and engineer than mammalian cells. The technology used to sub-clone and detect the mutants' activity in mammalian cells is far more expensive and qualitative selection and quantification of the mutants' activity can both be done in the same yeast system.
The present method of identifying protein CAMs presents a low cost, fast and powerful method to systematically identify activating mutations along the whole coding sequence of a protein. In contrast to previous work (Parnot et al., 2000), the cloning step is simplified to a simple transformation in yeast and the selection of active mutants is not more than picking growing colonies.
The transposition of the method into mammalian cells confirmed very well the constitutive activity of the mutants screened and selected in yeast. This proves that the method is a suitable alternative to mutant screening in mammalian systems.
Another big advantage of the method is that it immediately discriminates between a moderately active and a highly active mutant (a too high basal activity would not be suitable for agonist discovery, but appropriate for inverse-agonist screening). The growth speed of the colonies on agar selective medium is well correlated to the different “intensities” of constitutive activity observed in a liquid reporter assay.
De-orphaning can also be achieved with this method, e.g. the method can be applied to orphan GPCRs. Therefore, a low fidelity PCR product was co-transfected with the linearized vector into a panel of yeast strains expressing different humanized Gα protein subunits. On selective medium, mutants were selected only from the yeast strain expressing the Gα specific for its coupling. β-Galactosidase detection after growth in a selective medium showed an increased basal activity of the receptor mutant (i.e. an increased expression level of lacZ, controlled by a FUS1 promoter).
The method of identifying protein CAMs, of e.g. GPCR CAMs can be used for:
A NheI restriction site was created at position 157 bp of the coding sequence of EDG5. Three nucleotides where exchanged by site directed mutagenesis to create the site. This was necessary to conserve, after linearization of the plasmid, only the first 157 bp and the last 101 bp of the open reading frame for homologous recombination.
To allow in vivo recombination, the p416GPD-Edg5 was linearized by double digestion NheI-Xmal before co-transformation with the low-fidelity PCR amplification product.
Three colonies of each yeast transformation were tested for the growth and for the β-Galactosidease activity on selective plate:
After 24 hours of growth of the different mutants in selective liquid medium, in the presence of an increasing concentration of Sphingosine 1-Phosphate, β-Galactosidase activity was measured in a calorimetric assay by adding the substrate CPRG, incubating 2 hours and measuring the absorbance at 574 nm.
Luciferase activity measured in triplicates after 24 hours of stimulation by a serial dilution of Sphingosine 1-Phosphate.
A modified PCR protocol (Svetlov and Cooper, 1998) was used: initial denaturation at 95° C. for 3 min, 30 cycles of denaturation at 95° C. for 5 s, annealing at 50° C. for 5 s, and primer extension at 72° C. for 5 s, and final extension at 72° for 5 min, performed on the Cycler PTC-200 (MJ-Research).
The reaction was carried out with 2.5 U of Taq polymerase (Promega) using standard reaction buffer (10 mM Tris-HCl, pH 8.3, 1.5 mM MgCl2, 50 mM KCl) supplemented with 0.5 mM MnSO4. An equimolar mix of dNTPs (Amersham Pharmacia Biotech Inc) was used to provide 500 μM of each nucleotide triphosphate in a 100 μl reaction volume. 10 ng of the p416GPD-Edg5 plasmid were used as template. The following oligonucleotides (30 pmol of each) were used as primers for the PCR amplification: EDG5 fwd CAR (SEQ ID NO. 1: 5′-ATG GGC AGC TTG TAC TCG GAG T-3′) and EDG5 rev CAR (SEQ ID NO. 2: 5′-TCA GAA CAC CGT GTT GCC CTC-3′). They correspond exactly to the first 22 and last 21 nucleotides of the receptors sequence, thus the PCR amplifies exactly the open reading frame.
The PCR product (about 5 μg) was purified by electrophoresis through a 1% agarose-TBE gel followed by elution into 40 μl sterile water (QIAquick Gel Extraction Kit, Qiagen). The DNA final concentration was about 0.1 μg/μl.
2 μl of the PCR product were cloned into the TA Topo Cloning Vector (Invitrogen) for further qualitative and quantitative analysis of the randomly induced mutations.
The remaining volume of purified PCR product (3-5 μg in 38 μl) was co-transformed with 1 μg of p416GPD-Edg5 (linearized by double digestion with NheI-Xmal) into about 109 cells of the yeast strain (W303 MATa far1::hisG, sst2::ura3FOA, fus1::HIS3, ΔSte2::KanR, mfa2-fus1-lacZ::ura3FOA) according to a modified Lithium acetate method (Ito et al., 1983).
The transformed yeast cells were plated on 10 plates SC/Glucose -Ura medium to select for the cells with a “repaired plasmid” (about 108 yeast cells per plate).
After 36 hours incubation at 30° C. (when very small colonies were visible), the transformation plates were replica-plated onto selective medium: SC/Glucose -Ura -His, pH 6.8, containing 2 mM 3-Aminotriazol (3-AT).
After 48 hours incubation at 30° C., the colonies still growing were picked and restreaked as patches on a new selective plate (SC/Glucose-Ura-His, pH 6.8, containing 2 mM of 3-AT).
To eliminate false positive mutants (plasmid independent activity), the following steps were performed:
The plasmid from each selected clone was recovered by a Zymolase /SDS treatment protocol adapted from H. Ma & al. (Ma et al., 1987).
After ethanol purification, each plasmid was transformed into E. coli DH5α electro-competent cells. Individual bacterial transformants, one for each mutant, were grown in mini culture for plasmid preparation (QIAprep Spin Miniprep Kit, Qiagen).
The purified DNA was then transformed again into the same yeast strain and each mutant assayed.
From a 16 hours culture in SC/glucose -Ura medium, about 3×105 cells of each mutant (in triplicates) were spotted onto three different plates:
SC/Glucose -Ura as a control (to make sure that every spot contains roughly the same number of cells);
SC/Glucose -Ura -His, pH 6.8, containing 2 mM 3-AT;
SC/Glucose -Ura, pH 7, containing 100 μg/ml X-Gal (5-Bromo4-chloro-3-indolyl, β-D-galactopyranoside).
The two first plates were analyzed after 48 hours of growth at 30° C.; the third one was kept for 2 or 3 additional days at 4° C. to develop the blue coloration due to β-Galactosidase activity.
We selected the clones which grew on selective medium (SC/Glucose -Ura -His, pH 6.8, containing 2 mM 3-AT) and gave rise to a blue colored patch on X-Gal medium (SD -Ura, pH 7, containing 100 μg/ml X-Gal). These were candidates for constitutive activity (
The same 16 hours culture was diluted 200 times in selective medium (SC/Glucose -Ura -His, pH 6.8, containing 2 mM 3-AT) and 90 μl were dispensed into the 8 wells of a microtiter plate column already containing 10 μl of a serial dilution of the ligand Spingosine 1-Phosphate (Matreya) (solubilized and diluted in water from 10−3 to 10−9 M).
After 18 to 24 hours of stimulation/growth in a shaking incubator (700 rpm, 30° C.), the β-Galactosidase activity was detected with the substrate Chlorophenolred-β-D-galactopyranoside (CPRG, Boehringer).
The wild type Edg5 receptor and 3 of the 22 CAMs were sub-cloned into the mammalian expression vector pcDNA3.1(+) to be tested in cell culture (
This adherent cell line was grown under normal conditions (37° C., 5% CO2, humid atmosphere) in DMEM-Glutamax (Gibco BRL)+1% Penicillin/Streptomycin +10% Fetal Bovine Serum.
Day 1-40.000 cells/well were plated in a white 96-well plate
Day 2—the cells were rinsed with 200 μl Opti-MEM (Gibco BRL) and each well received 100 μl of a transfection mix containing: 0.5 μg receptor plasmid +0.25 μg CMV-β Gal (Promega)+1.2 μl Lipofectamine (Life Technologies) in Opti-MEM.
After 5 hours of incubation with this mix, the wells were emptied and received 180 μl of the normal culture medium (DMEM-Glutamax+1% Penicillin/Streptomycin) containing only 0.5% of Fetal Bovine Serum.
Day 3-20 μl of a 10-fold concentrated serial dilution of Sphingosine 1-Phosphate (from 10−4 to 10−10 M) was added to each well.
Day 4—the wells were emptied, rinsed with 200 μl phosphate buffer (without calcium and magnesium) and received 50 μl of Glo Lysis Buffer (Promega), after 5 minutes at room temperature, they received 50 μl of Steady-Glo Luciferase Reagent, and the measurement was achieved 5 minutes later in a Luminoskan (Labsystem), 15 seconds integration of the signal.
To normalize the results of the assay, β-Galactosidase activity was measured, from the same plate, after 5 minutes incubation with 25 μl of Gal-Screen Reagent (Tropix), 5 seconds integration of the signal. The luciferase numbers were then divided by the β-Galactosidase numbers.
The analysis of 38 randomly sequenced clones revealed 28 nucleotide mutations: 5 silent mutations, 1 STOP codon and 22 amino acid substitutions. These results suggest that under the experimental conditions the probability for an amino acid substitution to occur in the 354 residues of the wild-type sequence is 0.61.
The solid phase assay gives a confirmation of the first selection done after replicaplating the gap repair plates on selection plates. After being grown again on selective plate as patches (for confirmation), the plasmid DNA carrying the active mutant was purified, amplified in E. coli and re-transformed into the same yeast strain.
Three colonies of each yeast transformation were tested for growth and for β-Galactosidase activity on selective plates. The
To further characterize the mutants, their activity was tested in a liquid assay. Triplicates of each mutant (re-transformed in yeast) were grown to saturation in a pre-culture. These cell suspensions were diluted 200 times into a medium lacking histidine, permitting the growth of only activated receptors (HIS3 gene under the control of the FUS1 promoter) and distributed in a 96-well microtiterplate together with an increasing concentration of Sphingosine 1-Phosphate. After 24 hours of incubation and shaking at 30° C., the presence of β-Galactosidase activity was measured in a colorimetric assay by adding the substrate CPRG, incubating 2 hours at 30° C. and measuring the absorbance at 574 nm.
Out of three independent screens, 22 mutations have been found to confer constitutive activity to the Edg5 receptor (increased basal activity and the ability to be further stimulated by Sphingosine I-Phosphate) (Table 1). Interestingly, the mutation Ser159Pro was isolated in each of the three screens, and the mutations Ala82Val and Phe242Leu were isolated in two of the three screens.
A Serum Responsive Element (SRE)—Luciferase reporter assay in HEK 293 was chosen to verify in mammalian cells the activity of the CAMs selected with the yeast system. A stable HEK 293 cell line carrying the 6SRE-Luciferase construct was transfected with the wild type Edg5 or the mutants Ala82Val, Ser159Pro and Val238Glu. After 24 hours of stimulation, the measurement of Luciferase reflected the receptor's activity.
This assay (
To validate a screen, the whole process must be repeated in the same conditions. Indeed, the PCR principle can create an important bias introducing an activating (i.e. the Ala 82->Val mutation found 68 times in the third screen) or inactivating mutation in an early stage of the reaction. This mutation is then present in a high percentage of clones and can mask other interesting point mutations. This has to be circumvented. The best and fastest way would be doing at least three low-fidelity PCRs at the time and all the following steps in parallel.
| Number | Date | Country | Kind |
|---|---|---|---|
| 02000733.2 | Jan 2002 | EP | regional |
| Number | Date | Country | |
|---|---|---|---|
| 60360328 | Feb 2002 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 10340447 | Jan 2003 | US |
| Child | 11684459 | US |
