Biomedical Research (Tokyo) 35 (1) 47-59, 2014

Purification of the goldfish membrane progestin α (mPRα) ex- pressed in yeast Pichia pastoris

1 2 1 1, 2 Takayuki OSHIMA , Ryo NAKAYAMA , Shimi Rani ROY , and Toshinobu TOKUMOTO 1 Integrated Bioscience Section, Graduate School of Science and Technology, National University Corporation Shizuoka University, Ohya 836, Suruga-ku, Shizuoka 422-8529, Japan; 2 Biological Science Course, Graduate School of Science, National University Corpo- ration Shizuoka University, Ohya 836, Suruga-ku, Shizuoka 422-8529, Japan (Received 22 November 2013; and accepted 12 December 2013)

ABSTRACT Membrane progestin receptors (mPRs) are key mediators of rapid, nongenomic actions of proges- tins on plasma membranes. We established a procedure for the expression and purification of re- combinant goldfish mPRα using the methylotropic yeast Pichia pastoris. In P. pastoris, the recombinant , which carried C-terminal histidine and c-Myc tags, was expressed in an ac- tive form as the receptor for maturation-inducing steroids of fish. Expressed were bound

reversibly with a high affinity (Kd = 9.4 nM) at a single binding site that could be saturated. After solubilization of mPRα with n-dodecyl-β-D-maltoside (DDM) from yeast membranes, the recom- binant protein was purified using three different columns: first it was affinity-purified over nickel- nitrilotriacetic acid (Ni-NTA), then bound to a cellulose resin with free amino groups and finally to a column with affinity for the c-Myc epitope. The identity of the purified protein was verified by MALDI-TOF/MS analysis and its capacity to bind progestin remained. Expression and purifi- cation of mPRα protein in its functional form will enable the screening of ligands and the deter- mination of its three dimensional structure.

Progestins are important steroids regulating final trihydroxy-4-pregnen-3-one (20β-S) was identified maturation at reproductive tissues in male and fe- in Atlantic croaker (Micropogonias undulatus) and male vertebrates. They can be divided into natural spotted seatrout (Cynoscion nebulous) (22). The and synthetic types. Natural type is found in humans structural differences between human progestin, pro- and certain animals (32). Natural progestin in human gesterone, and fish progestins are that fish progestins is . In the medical fields, a variety of have multiple hydroxyl groups on the side chain at synthetic progestins are used. In teleost fishes, two positions 17, 20 and 21. In fish, MIS is produced by distinct progestins have been identified as naturally the follicular envelope in response to luteinizing hor- occurring maturation inducing steroids (MIS), which mone (LH) from the pituitary (15). In salmon, LH control oocyte maturation. 17α,20β-dihydroxy-4- stimulates a number of signaling cascades leading to pregnen-3-one (17,20β-DHP) was identified in ama- the production of 17α-hydroxyprogesterone (17α-HP) go salmon (Oncorhynchus rhodurus) and 17α,20β,21- from progesterone and 17α-hydroxypregnenolone in the theca cells. Then, 17α-HP is converted to 17,20β-DHP by 20β-hydroxysteroid dehydrogenase Address correspondence to: Toshinobu Tokumoto, PhD, (20β-HSD) in the granulosa cells (18, 22). The Integrated Bioscience Section, Graduate School of Sci- ence and Technology, National University Corporation membrane progestin receptor (mPR) mediates the Shizuoka University, Ohya 836, Suruga-ku, Shizuoka nongenomic actions of progestins on the plasma 422-8529, Japan membrane. It was identified in teleost, a ray-finned Tel & Fax: +81-54-238-4778 fish, at first and subsequently identified in other -ver E-mail: [email protected] tebrates including human (44, 45). Research into the 48 T. Oshima et al. functions of mPR have revealed that this receptor cytes, suggesting that the mPRα and β proteins me- mediates the action of steroids that play a role in, diate the action of maturation inducing steroids. for example, oocyte maturation in fish (13, 21, 35, Membranes prepared from breast cancer cells trans- 37, 42). Genome wide phylogenetic analyses have fected with mPRα exhibited single binding sites that shown that mPRs can be classified into a new pro- were specific for a steroid that induced oocyte matu- tein family, the progestin and adipoQ receptor ration in goldfish, 17,20β-DHP (38). We found that (PAQR) family and they have three subtypes named treatment of oocytes with diethylstilbestrol (DES), mPRα, β and γ (corresponding to PAQR7, 8 and 5, an endocrine disrupting chemical, induced their mat- respectively) (33, 36). The mRNAs of mPRs in hu- uration in both goldfish and zebrafish (39). DES man show differential expression in reproductive tis- also bound with high affinity to membranes from sues such as the ovary, testis, placenta, uterus and mPRα-transfected breast cancer cells. These results nonreproductive tissues such as brain, kidney and suggested that mPR proteins are possibly targeted intestinal tissues, suggesting they mediate different by chemicals that affect the endocrine system. nongenomic actions of progestins (35). In goldfish, Recombinant mPR proteins have been expressed mPRs are also expressed in reproductive tissues, in Escherichia coli and several human cancer cell brain, kidney, eye, gill and spleen. The wide tissue lines (38). Although large-scale culturing in E. coli distribution of mPRs suggests mPRs mediates the is possible, post-translational modification required multiple progestin actions in a wide variety of target to obtain an active form of the recombinant protein tissues (37). Recently, based on the analyses of pro- will not occur. On the other hand, expression of teins expressed in yeast, mPRδ and ε (PAQR6 and 9) mPRs in mammalian cell culture, which would re- were proposed to form novel mPR subtypes (31). sult in proteins that are naturally post-translationally Also human orthologs of them, when expressed in modified, will only reach low levels, which is not human breast cancer cells, were found to have pro- suitable for the purification of the amounts of pro- gestin binding activity (24). Therefore, five mPR tein needed for biochemical and structural analysis. subtypes have been described at present. We therefore selected the methanol-utilizing yeast Rapid nongenomic effects of progestin and mPR Pichia pastoris as a host for the expression of mPRα. expression have been investigated in a wide variety P. pastoris has been used for the expression of G of tissues of fishes and mammals. In addition to protein coupled receptors (GPCRs) and can express mPRs, there are two receptor candidates for mediat- proteins in their functional form in large-scale puri- ing progestin signaling from the cell surface: nucle- fications, as has been described in more than 100 ar progestin receptors (nPRs) and progestin receptor reports. For example, mammalian GPCRs expressed membrane components (PGRMCs). Since their tis- in P. pastoris exhibited a natural ligand binding ac- sue-specific expression overlaps, it seems likely that tivity (19, 43). In this way, P. pastoris has been their functions do as well. Thus, physiological roles used to isolate proteins in amounts that enabled of receptors mediating progestin actions might be structural and biochemical studies of GPCRs (3). In complicated and difficult to unravel (12, 17, 46). 2011 and 2012, the structure for two human GPCRs

For example, in addition to its function as a nuclear (the histamine H1 and the adenosine A2a receptor) transcription factor, nPR modulates cell signaling could be determined for the first time after being pathways by activating c-Src and downstream purified from P. pastoris (14, 30). MAPK outside the nucleus (5, 27). In Xenopus oo- In the present study, we established a procedure cytes, like mPRβ, nPR was shown to mediate the for the expression and purification of mPR, a new nongenomic action of progestin during oocyte matu- member of the GPCR family. Recombinant goldfish ration, suggesting a potential interaction between mPRα with progestin binding activity could be ex- mPRs and nPRs (46). Also, transactivation of nPRs pressed in P. pastoris as a fusion-protein with the by the activation of mPRs was demonstrated in hu- pre-pro signal sequence of α-factor from Saccharo- man myometrium (16). Although the details about myces cerevisiae to ensure localization to the plas- the signaling pathway through mPRs still remain ma membrane. The protein was biochemically active unclear, several pathways different from those in- with respect to the binding of a typical mPRα-ligand volving nPRs have been suggested (8, 11, 47). and could be purified by sequential column chroma- We cloned and characterized four mPR subtypes, tography after solubilization from the cell-mem- mPRα, β, γ-1 and γ-2 from goldfish (37, 40). Micro- brane. The described procedure can be used to injection of antisense oligonucleotides targeting generate sufficient amounts of goldfish mPRα pro- mPRα and β blocked the maturation of goldfish oo- tein that would allow the determination of its struc- Purification of recombinant mPRα 49 ture, the analysis of intermolecular interactions, and medium in a 500 mL baffled flask which was incu- the production of monoclonal antibodies. bated at 30°C for 1 day. For induction of mPRα- expression by methanol, cells were harvested by centrifugation, and taken up in 1 L of BMMY medi- MATERIALS AND METHODS um (1% yeast extract, 2% bactopeptone, 100 mM P. pastoris strains expressing goldfish mPRα. For potassium phosphate, pH 6.0, 1.34% yeast nitrogen expression in P. pastoris, the cDNA for goldfish base without amino acids, 4 × 10−5% biotin, 0.5% mPRα, isolated from a mammalian expression vec- methanol) to an OD600 of 1.0–3.0. The medium was tor (38), was fused to the pre-pro secretion signal of placed in a 2 L baffled flask and incubated at 20°C α-factor from S. cerevisiae in the P. pastoris expres- for one day, after which the cells were pelleted at sion vector pPICZαC (Invitrogen). The ORF region 3,000 × g for 5 min, frozen with liquid nitrogen and of the resulting expression cassette (Fig. 1A) was stored at −80°C until use. verified by DNA sequencing. Three P. pastoris strains GS115, KM71H and Membrane preparation and solubilization of mem- X-33 (Invitrogen), were transformed by electropora- brane proteins. Frozen cell pellets were thawed and tion with the mPRα-expression construct as de- resuspended in ice-cold breaking buffer (50 mM so- scribed in the manual of the EasySelect Pichia dium phosphate, 1 mM PMSF, 1 mM EDTA, 5% Expression Kit (Invitrogen). Prior to transformation, glycerol, pH 7.4) containing 0.1% digitonin (Sigma the plasmid was linearized by PmeI. Electroporation Aldrich) with an equal volume of zirconium beads. was performed at 1500 V, 25 μF and 800 Ω using a Cells were broken with six rounds of vigorous shak- Pulser (Bio-Rad). ing for 5 min using a Micro Homogenizing System Recombinant clones were selected on YPD plates MS-100 (TOMY Seiko) followed by 5 min incuba- (1% yeast extract, 2% peptone, 2% dextrose, 2% tion on ice. Intact cells and cell debris were separat- agar) containing 100 μg/mL Zeocin and genomic in- ed from the fraction containing the membranes by a tegration of the mPRα-expression construct was ver- low speed centrifugation step (1,000 × g, 4°C, 7 min), ified by PCR using Ex Taq polymerase (Takara Bio, after which the pellet was resuspended in ice-cold Siga, Japan) and primers (5’ AOX1; 5’-GACTGGTT breaking buffer for another round of cell-breakage. CCAATTGACAAGC and 3’ AOX; 5’-GCAAATGG In all, the supernatants of three subsequent rounds CATTCTGACATCC) that bind to the AOX1 pro- of beat-beating were combined and the membrane moter and terminator regions flanking the ORF fraction was recovered by centrifugation at 20,000 × (Fig. 1A). Several Zeocin-resistant clones were test- g, 4°C, for 20 min. This fraction was used for the ed for the production of recombinant protein. Clones measurement of receptor-activity by a binding assay with the best expression out of ten were kept and (see below) or formed the starting material from stored on MD plates (1.34% yeast nitrogen base, which membrane proteins were solubilized by resus- 4 × 10−5% biotin, 2% dextrose, 1.5% agar) at 4°C. pending the membrane pellet in lysis buffer (50 mM

NaH2PO4, 300 mM NaCl, 10 mM Imidazole, 1 mM Expression of goldfish mPRα in P. pastoris. In order PMSF, 10% Glycerol) to which a detergent was to achieve optimal expression of goldfish mPPα in P. added. As an initial test, eight detergents (MEGA- pastoris, we tested the protein production in three 10, FOS-choline (FOS), n-octyl-β-D-thioglucoside different strains, GS115, KM71H and X-33, and (OTG), Tween-20, Tween-80, n-dodecyl-β-D- varied the conditions of growth. Transformants of maltoside (DDM), 4,4’-Dithiodibutyric acid (DDA), X-33 showed the highest level of mPRα expression n-octyl-β-D-glucoside (OG)) were, at various final (data not shown), which could be optimized by concentrations of 0.01 or 0.1%, incubated with a growing these cells in buffered media (see Materials membrane preparation for 20 min at room tempera- and Methods) and induce expression at 20°C for ture, after which the solubilized fractions were sepa- 24 h. For the production of the mPRα receptor, cells rated from the insoluble material by centrifugation were pre-cultured in 30 mL BMGY medium (1% (20,000 × g, 4°C, 20 min) and analyzed by Western yeast extract, 2% bactopeptone, 100 mM potassium blot analysis using anti-His-tag antibodies. For the phosphate, pH 6.0, 1.34% yeast nitrogen base with- preparation of solubilized proteins that had to be out amino acids, 4 × 10−5% biotin, 1% glycerol) in further purified, the most effective detergent, DDM, 200 mL baffled flasks at 30°C using a shaking incu- was added to the membrane samples to a final con- bator (250 rpm) for 1 day. Then, 4 mL of the pre- centration of 0.1%. After an incubation at 4°C for culture was used to inoculate 100 mL BMGY 30 min insoluble material was removed by centrifu- 50 T. Oshima et al. gation (20,000 × g, 4°C, 20 min) and the supernatant concentrated with a Centriprep YM-3 spin column containing the solubilized proteins was frozen with (Millipore), which was also used to replace the buf- liquid nitrogen and stored at −80°C until further fer to 50 mM Tris-HCl, pH 8.0 containing 0.01% use. DDM and 1 mM PMSF (buffer B). In the second purification step, the sample was loaded onto a Radiolabeled ligand binding assays. For ligand bind- 5 mL Cellfine Amino (JNC, Tokyo, Japan) column ing assays, frozen pellets of membrane fractions (φ1.6 × 10 cm) that had been equilibrated with buf- were resuspended in HEAD buffer (25 mM HEPES, fer B. The column was washed with 15 mL buffer B 10 mM NaCl, 1 mM Dithiothreitol, 1 mM EDTA, and eluted with a 100 mL gradient of 0–0.5 M NaCl pH 7.6) containing 0.1% digitonin, or, when testing in buffer B. The fractions that contained recombi- purified recombinant protein, pellets were taken up nant mPRα as shown by western analysis using anti- in HEAD buffer without digitonin. Progestin recep- His-tag antibodies were collected and concentrated tor binding was measured following procedures es- with a Centriprep YM-3 spin column, and the buffer tablished previously (38). For saturation analyses was replaced with PBS buffer before the sample and Scatchard plots 17α-Hydroxyprogesterone (Sig- was purified with anti-c-Myc-tag beads (part of the ma Aldrich) was enzymatically converted to radio­ “c-Myc tagged protein mild purification kit”; MBL, labeled 17,20β-DHP by 3α,20β-hydroxysteroid Nagoya, Japan). Recombinant mPRα was eluted dehydrogenase (Sigma) as described previously (29). from the beads by washing with PBS buffer contain- Various concentrations (3–12 nM) of [3H]-17,20β- ing 1 mg/mL c-Myc-epitope peptides. The purified DHP (specific activity, 40 Ci/mmol) were added to protein was analyzed by SDS-PAGE and visualized the assay tubes with or without 100-fold molar ex- by silver staining using the 2D-Silver Stain II kit cess cold 17,20β-DHP for the measurement of the (Daiichi Pure Chemicals, Tokyo, Japan). number of nonspecific and total interactions, respec- tively. Binding reactions were done at 4°C for SDS-PAGE and Western blot analysis. Proteins were 30 min and stopped by filtration over Whatman GF/ separated by SDS-polyacrylamide gel electrophoresis B filters, that had been presoaked in wash buffer (SDS-PAGE) on a 10% polyacrylamide gel under (25 mM HEPES, 10 mM NaCl, 1 mM EDTA, denaturing conditions by the method of Laemmli, pH 7.4) containing 2.5% Tween 80 only when ana- and transferred to Immobilon membranes (Milli- lyzing membrane fractions. The filters were washed pore). Membranes were processed using the SNAP three times with 5 mL of wash buffer at 4°C and i.d. protein detection system (Millipore) and blocked bound radioactivity was measured by scintillation in 5% non-fat powdered milk in 20 mM Tris-buff- counting. Binding reactions with purified protein ered saline, pH 7.6 (TBS) containing 0.1% Tween samples were done in the presence of BSA (final 20 (TTBS) for 1–2 h at room temperature, and then concentration 10 mg/mL) and with 1.5 nM [3H]- incubated with primary antibodies (500-fold diluted 17,20β-DHP in the presence or absence of 1 μM in TBS buffer), i.e. polyclonal anti-His-tag antibody 17,20β-DHP. Linear and nonlinear regression analy- or anti-Myc-tag antibody (MBL, Nagoya, Japan), ses for all receptor binding assays and calculations and with anti-rabbit antibody conjugated with per- of Kd and binding capacity (Bmax) were done using oxidase (Invitrogen) as the secondary antibody GraphPad Prism for Macintosh (version 4.0c; www. (2,000-fold diluted in TBS buffer). Proteins were vi- graphpad.com). The results were shown on Scatchard sualized by enhanced chemiluminescence using an plots. ECL detection kit (PerkinElmer). Signals were digi- tized using a CCD camera system, the Luminescent Purification of mPRα. To purify the recombinant Image Analyzer LAS-4000 mini (Fujifilm). mPR, the solubilised proteins were thawed on ice and loaded onto a 25 mL nickel-nitrilotriacetic acid Kex2 digest. Purified proteins were digested by (Ni-NTA) agarose (Qiagen) column (φ2.6 × 20 cm) Kex2 protease. Proteins were incubated in buffer that had been equilibrated with buffer A (50 mM containing 200 mM Bistris, 1 mM CaCl2, 0.01% Tri- NaH2PO4, 300 mM NaCl, 10 mM imidazole, 1 mM tonX-100, 0.5% dimethyl sulfoxide, pH 7.0 with PMSF, 0.01% DDM, pH 8.0). The proteins were Kex2 (final concentration 0.1 mg/mL; Abcam, Cam- eluted with a 500 mL gradient of 10–250 mM imid- bridge, MA) in a total volume of 20 μL at 30°C for azole in buffer A. The fractions that contained re- 1 h. Digested mPRα protein was detected by anti- combinant mPRα as revealed by western analysis His-tag antibodies. (using anti-His-tag antibodies) were collected and Purification of recombinant mPRα 51

MALDI-TOF/MS analysis. After tryptic digestion of molecular weights were consistent with molecular SDS-PAGE gel-slices containing purified recombi- masses predicted for mPRα with and without the nant protein (stained with CBBR), peptides were re- pre-pro α-factor signal peptide, 54 kDa and 44 kDa, covered and purified using a ZipTip (Millipore) and respectively. These two bands were major bands, eluted in 2 μL of a solution containing 5 mg/mL of only detected after induction by methanol (data not CHCA (Bruker Daltonics), 60% acetonitrile, and shown). On the same blot, a 50 kDa protein was de- 0.1% TFA. The samples were loaded onto a 384- tected in the membrane fraction isolated from the P. well plate, that was double layered with CHCA dis- pastoris host strain that did not have the mPRα ex- solved in acetone, and air-dried. The peptide mass pression cassette (Fig. 1C), but this protein did not spectrum was obtained on a MALDI-TOF/MS Auto- co-purify with mPRα (see below). Thus, we con- flex (Bruker Daltonics, Billerica, USA) in positive cluded that these two bands were goldfish mPPα ion mode. All MALDI-TOF/MS spectra were exter- proteins expressed from the integrated cDNA. nally calibrated using a molecular weight standard mixture (Bruker Daltonics). The peptide fingerprint Specific [ 3H]-17,20β-DHP binding to plasma mem- was analyzed with MASCOT software (Matrix Sci- branes from mPRα-expressing P. pastoris ence, London, UK) searching against peptides from For the detection of progestin (17,20β-DHP) binding ray-finned fishes or fungi in the NCBInr database, activity of expressed recombinant mPRα protein using the following parameters: trypsin digest (zero present in the P. pastoris cell membranes, digitonin miss cleavage), cysteines modified by carbamido- was present during preparation of the cell-mem- methylation, and mass tolerance 0.4 Da. The identi- branes (see Materials and Methods) as this glyco- fication was based on probability-based MOWSE side is known to facilitate receptor-access for this scores. steroid (26). As has been previously reported for Pi- chia cells that had to be permeabilized (1), we Statistical analysis. All experiments were repeated found that a final concentration of 0.1% digitonin three times. One-way analysis of variance (ANOVA) was optimal to facilitate steroid binding as measured was calculated using GraphPad Prism (San Diego, by a filter-binding assay established previously (38) CA). A P value < 0.05 was considered statistically (Materials and Methods) whereas cholate did not significant. (Fig. 2A). After treatment with 0.1% digitonin, spe- cific 17,20β-DHP binding activity was detected in plasma membranes that were prepared from P. pas- RESULTS toris cells expressing the goldfish mPRα but not in Expression of recombinant goldfish mPRα membranes from untransformed host cells (Fig. 2B). In order to purify goldfish mPRα using the yeast P. Saturation analysis showed that progestin binding pastoris, the cDNA of goldfish mPRα was fused to to the cell membranes of mPRα-expressing cells the pre-pro secretion signal of the α-factor from S. was saturable and of a limited capacity (Bmax = cerevisiae in an expression cassette (Fig. 1A) that 0.024 nM). Scatchard analysis indicated the pres- was integrated into the host yeast genome by ho- ence of a single class of high-affinity binding sites mologous recombination (Materials and Methods). (Kd = 9.4 nM) in these cell membranes (Fig. 2C). Insertion of the cassette was verified by PCR using Therefore, these results suggested that heterologously primers that amplified both the inserted gene-fusion produced, recombinant goldfish mPRα is active. as well as the endogenous P. pastoris gene, AOX1 (Fig. 1B), whose promoter and terminator regions Receptor solubilization control transcription of heterologous mPRα gene-fu- In order to obtain pure mPRα that could be used for sion (Fig. 1A). structural studies and biochemical assays, the pro- Expression of the fusion-protein, that also carried tein has to be solubilized and separated from the an epitope for c-Myc and a histidine tag at its C-ter- membrane fractions. We found that out of eight de- minal end, was induced by the presence of 0.5% tergents tested, 0.1% n-Dodecyl-β-D-maltoside methanol in the medium. When the synthesis of re- (DDM) solubilized the largest amount of mPRα combinant protein was analyzed by western blotting (Fig. 3), which was used to prepare the protein frac- using cell membranes prepared from P. pastoris tion. The recombinant receptor could be purified cells carrying the expression cassette, two immuno- from this fraction by column chromatography. reactive bands of around 50 kDa were detected us- ing anti-His-tag antibodies (Fig. 1C). The apparent 52 T. Oshima et al.

Fig. 1 Expression of mPRα in Pichia pastoris. (A) Schematic representation of the mPRα-expression cassette. Transcrip- tion of the fusion peptide formed by α-factor signal sequence (α-factor) linked to mPRα that carries C-terminal histidine (6xHis) and c-Myc epitopes is controlled by the methanol-inducible AOX1 promoter (pAOX1) and the AOX1 terminator (AOX1TT). Cleavage by endogenous peptidases Kex2 and Ste13 results in the removal of the signal peptide during intra- cellular transport. The 5’ AOX1 and 3’ AOX1 primer binding are indicated by black bars above and below the cassette, re- spectively. Remaining AOX1 gene (about 2.1 kbp) after insertion of transformed gene is also depicted. (B) Evaluation of gene insertion by PCR amplification. PCR was performed with total DNAs extracted from untransformed cells (X-33), or cells transformed with the goldfish mPRα cassette (mPRα-X-33), the goldfish mPRα inserted vector (mPRα-pPICZαC) or the empty vector (pPICZαC). The bands derived from endogenous AOX1, recombinant mPRα and empty expression-vector pPICZαC are indicated. (C) Western blot analysis of mPRα expression. Proteins of about 54 kDa and 44 kDa (arrows) were reactive to anti-His-tag antibodies in membrane preparations from X-33 cells expressing goldfish mPRα (mPRα-X-33). As visualized in the membrane fraction from unstransformed X-33 cells (X-33), an endogenous protein of about 50 kDa (aster- isk) was reactive as well.

Purification of solubilized mPRα by column chroma- ing anti-His-tag antibodies. As a first purification tography step, the DDM-solubilized membrane proteins were Recombinant mPRα was prepared from large-scale passed over a Ni-NTA column, from which two cultures grown in buffered medium containing 0.5% bands that cross-reacted with anti-His-tag antibodies methanol from which membrane proteins had been were eluted after inclusion of over 10 mM imidaz- isolated by solubilization with 0.1% DDM as de- ole in the buffer (Fig. 5A). The bulk of the protein, scribed above and in the Materials and Methods. which also included the endogenous 50 kDa protein Recombinant mPRα was purified from the fraction that could be detected with anti-His-tag antibodies of solubilized membrane proteins by three steps of (Fig. 1C), washed off the column at lower imidazole column chromatography. Proteins were eluted from concentrations (Fig. 5A). As the putative recombi- the columns by linear salt-gradients and the recov- nant mPRα protein eluted over multiple fractions, ered fractions were assessed by immunoblotting us- the protein had to be concentrated and the imidazole Purification of recombinant mPRα 53

Fig. 2 Production of active recombinant mPRα in P. pastoris X-33. (A) Specific binding activity of [3H]-17,20β-DHP to membranes prepared from X-33 cells expressing mPRα (mPRα-X-33) was determined in the presence of various concen- trations of digitonin or sodium deoxycholate. (B) Specific binding activity of3 [ H]-17,20β-DHP to membranes prepared from untransformed host cells (X-33) or from X-33 cells expressing mPRα (mPRα-X-33). (C) Saturation curves and Scatchard plots of specific [3H]-17,20β-DHP binding to membranes prepared from mPRα-X-33 cells. removed. For this we used a column made with a record ligand-binding activity of the membrane frac- cellulose-based resin with free amino groups (Cell- tions (Fig. 2). Specific 3[ H]-17,20β-DHP binding ac- fine Amino; JNC, Tokyo, Japan), which was the tivity was detected for the purified proteins (Fig. 6C). most effective resin out of 17 tested that was able to The purified proteins were cleaved by Kex2 prote- retain mPRα (Fig. 4). mPRα was eluted by salt solu- ase (Fig. 6D). Immunoblotting with anti-His-tag an- tion (Fig. 5B) and further purified by affinity chro- tibodies showed a decrease in intensity of 54 kDa matography on a column containing anti-c-Myc tag band while the intensity of 44 kDa band was in- beads and eluted by washing with a c-Myc-epitope creased. This result suggested that recombinant peptide. The purified proteins could be visualized by mPRα with pre-pro α-factor signal peptide was protein staining (Fig. 6A) and were reactive to anti- cleaved at the Kex2 cleavage site (Fig. 1A) and sup- bodies specific for both the His- and c-Myc epitope, ported that these two bands were recombinant gold- indicative for enrichment of recombinant goldfish fish mPRα proteins. mPRα (Fig. 6B). The identity of the two purified proteins with ap- parent molecular weights that were close to those pre- Characterization of purified recombinant mPRα dicted for mPRα with (54 kDa) or without (44 kDa) Steroid binding activity of the purified proteins was the pre-pro α-factor peptide, was established by measured by a similar filter-binding assay as used to MALDI-TOF/MS analysis. The 54 kDa protein was 54 T. Oshima et al.

Fig. 3 Efficacy of various detergents to solubilize mPRα. Western blot analysis with anti-His-tag antibodies on solubilized (sup) and insoluble, precipitated material (ppt) after treatment with MEGA-10, FOS-choline (FOS), n-octyl-β-D-thioglucoside (OTG), Tween-20, Tween-80, n-dodecyl-β-D-maltoside (DDM), 4,4’-Dithiobutyric acid (DDA), n-octyl-β-D-glucoside (OG). Bands of approximately 54 kDa and 44 kDa representing recombinant mPRα are indicated (arrows).

Fig. 4 Efficacy of various lesins to purify mPRα. Western blot analysis with anti-His-tag antibodies on unadsorbed (T) and adsorbed proteins (E) after treatment with various lesins are represented as following; lane 1. Gigapite, 2. Blue Sepharose CL-6B, 3. Butyl Sepharose 4B, 4. Cellufine Amino, 5. Matrex gel green A, 6. Bio-Gel HT, 7. Phenyl Sepharose, 8. Poly (U) Sepharose 4B, 9. Red Sepharose CL-6B, 10. DEAE Trisacryl PlusM, 11. 5’ AMP Sepharose 4B, 12. Q Sepharose XL, 13. Con A Sepharose, 14. Glutathion Sepharose, 15. Protein A Sepharose CL-4B, 16. Ni-NTA Agarose, 17. DE52. Lanes M and C represent marker (M) and control sample without treatment (C), respectively. Bands of approximately 54 kDa and 44 kDa representing recombinant mPRα are indicated (arrows). confirmed to be goldfish mPRα containing the sig- fied. Proteins, heterologously expressed in P. pastoris, nal sequence of α-factor as peptide mass finger print can either remain intracellular or be targeted for se- analysis resulted in 6 matched peptides and 18% of cretion by use of the secretion signal sequence of sequence coverage for goldfish mPRα, and a peptide α-factor from S. cerevisiae (7, 28). We identified for the N-terminus of pre-pro α-factor (Fig. 7). Pep- two proteins that co-purified with the plasma-mem- tide mass analysis of fragments derived from the brane that carried the epitopes attached to the C-ter- 44 kDa band revealed that these matched those of minus of recombinant mPRα. The larger form of the goldfish mPRα peptides from the 54 kDa band these proteins turned out to be goldfish mPRα that but no fragments matching α-factor were retrieved. still had the α-factor signal peptide attached, sug- Therefore, we concluded that the 54 kDa band rep- gesting that during the intra-cellular transport the resents the unprocessed form of the α-factor-mPRα removal of this sequence did not occur very effi- fusion protein while the 44 kDa band is recombinant ciently. When we tried to compare the expression mPRα from which the signal sequence has been levels of goldfish mPRα when expressed with or cleaved off, probably during intra-cellular transport without α-factor sequence fused to its N-terminus, to the plasma membrane. we found that only the version targeted to the cell membrane by means of the pre-pro peptide of α-factor was detectable, suggesting that the receptor DISCUSSION needs a membrane environment for stability. No In this study, engineered goldfish mPRα was ex- mPRα was found in the medium, thus the recombi- pressed in yeast P. pastoris and subsequently puri- nant receptor was not secreted. Accordingly, steroid Purification of recombinant mPRα 55

Fig. 5 Purifiation of mPRα over Ni-NTA and amino cellulose. Chromatograms and Western blot analysis of fractions eluted from a Ni-NTA column (A) and the Cellufine Amino columnB ( ). Elution profiles were monitored by absorbance at 280 nm. Horizontal bars represent fractions analyzed by immunoblotting with anti-His-tag antibodies and used in subsequent purifi- cation steps. binding activity could be detected in membranes hydrophobic residues (26). prepared from P. pastoris expressing recombinant In order to purify recombiant mPRα from cell- mPRα but only after treatment with digitonin. For membranes, the seven-transmembrane receptor had ligand binding assays with bovine membrane frac- to be solubilized. For this, DDM was used, as has tions, digitonin has been found to enhance the bind- been described before (4) and found to be the only ing of progesterone (26), while progestin binding to detergent that was effective (Fig. 3). When the solu- a membrane fraction isolated from human sperm bilized membrane proteins were fractionated on a could only be detected in the presence of digitonin Ni-NTA column, the His-tagged recombinant mPRα (2). Digitonin supposedly forms a complex with was separated from the bulk of the proteins, al- progesterone-like steroids and is thought to support though it eluted in all fractions upon increasing the steroid binding by interaction with the surrounding imidazole concentration above 10 mM. To recover 56 T. Oshima et al. the receptor from these fractions, we found a cellu- lose resin with free amino groups that retained mPRα, but purification was not very specific. There- fore, a third step of affinity chromatography, beads having the ability to bind the protein via its C-ter- minal c-Myc tag, was required and yielded very pure mPRα (Fig. 6B) which was able to bind pro- gestin (Fig. 6C). The purified protein was confirmed as goldfish mPRα protein by MALDI-TOF/MS anal- ysis, which also revealed that the signal sequence of α-factor was still attached to the higher-molecular weight form of the isolated receptor. Thus, we suc- ceeded in expressing and purifying a recombinant goldfish mPR which was active for the first time. The potential cellular roles of progestin binding to mPRs have been investigated widely, and it was found these interactions triggered major events dur- ing reproduction, breast cancer cell physiology and neuroendocrine responses. During the reproductive cycle in mammals, fish and frogs mPRs, like mPRα, mediate the induction of oocyte maturation and sperm motility, which could involve the activation of a G-protein, upon binding of progestins (34, 37, 41, 45). An immunological response linked to repro- duction is the up-regulation of the leukemia inhibi- tory factor by progesterone, which is essential for the implantation and development of the embryo and inhibits several macrophage functions (25). Pro- gesterone, possibly via interaction with mPRα, mPRβ or mPRγ, suppresses the proliferation of human T cells, which might attack the fetus during pregnan- 2+ cy, by regulating changes in the [Ca ]i and pHi in- side the T cells (6). Furthermore, progesterone might modulate the activation of bovine T lympho- cytes and their proliferation in the corpus luteum by Fig. 6 Analysis of purified recombinant αmPR . (A) SDS- binding to mPRs (23). PAGE analysis of representative fractions after solubiliza- tion of the membrane preparation (sample), column Progesterone interactions with mPRα are associat- chromatography over Ni-NTA and amino cellulose (Cellufine ed with changes in breast cancer cells. An increase Amino). Protein bands were detected by CBBR staining. (B) in the levels of mRNAs coding for mPRs within SDS-PAGE and Western blot analysis of the purified frac- breast cancer cells has been observed (8), which fits tion. Representative fraction after column chromatography recent findings that show how a cell-signaling cas- over beads conjugated with c-Myc-tag antibodies. Protein bands were detected by silver staining (Ag Stain). Western cade in breast cancer cells originates from proges- blot analysis (Immuno) was conducted using anti-His-tag terone binding to mPRα in the cell membrane (47). antibody (α-His) and anti-Myc antibody (α-Myc). Two forms Other, recent data suggest that, as part of the cellu- of mPRα are indicated by arrows. (C) Specific binding of 3 lar response to the binding of progesterone to [ H]-17,20β-DHP to purified recombinant mPRα or in a mPRα, apoptosis is inhibited in breast cancer cells mock sample lacking the purified receptor (control). (D) Western blot analysis of purified recombinant αmPR after (8–10). proteolytic cleavage with Kex2. Proteins (sample) were in- In the membranes of neuroendocrine cells, mPRs cubated with (Kex2) or without (control) Kex2 protease. appear to interact with G proteins and have a role in the release of hormones and maintain oocyte matu- ration by regulating gonadotropin releasing hormone diated by mPRα. Interestingly, this receptor is only (GnRH) secretion (24). Progesterone is known to present in neurons in basal condition, while it is in- have neuroprotective effects, which possibly is me- duced in glial cells after traumatic brain injury (20). Purification of recombinant mPRα 57

Fig. 7 MALDI-TOF mass analysis of purified recombinant αmPR . Amino acid sequence of recombinant goldfish mPRα from construct produced in this study is represented. Peptides matched in peptide mass finger print analysis are underlined (upper and lower lines correspond to 54 and 44 kDa bands). A peptide fragment derived from α-factor signal sequence was only detected in the analysis of 54 kDa band.

Since mPRs are potential intermediaries of the REFERENCES various cellular responses to progesterone described 1. Alamae T and Jarviste A (1995) Permeabilization of the above, they are expected to be targets for new drugs methylotrophic yeast Pichia pinus for intracellular enzyme that would address fertility problems, suppress can- analysis: a quantitative study. J Microbiol Methods 22, 193– cer or fit in therapies for the treatment of encephali- 205. tis and brain injury. The establishment of a procedure 2. Ambhaikar M and Puri C (1998) Cell surface binding sites for progesterone on human spermatozoa. Mol Hum Reprod 4, by which large amounts of pure, functional mPR 413–421. proteins can be produced will facilitate the screen- 3. Asada H, Uemura T, Yurugi-Kobayashi T, Shiroishi M, ing of ligands for mPRs. We previously reported Shimamura T, Tsujimoto H, Ito K, Sugawara T, Nakane T, that the binding of endocrine disrupting chemicals Nomura N, Murata T, Haga T, Iwata S and Kobayashi T (2011) Evaluation of the Pichia pastoris expression system to mPRs could induce or inhibit the maturation of for the production of GPCRs for structural analysis. Microb fish oocytes (39). The recombinant protein purified Cell Fact 10, 24. in this study could be used in a high throughput as- 4. Barrera NP, Di Bartolo N, Booth PJ and Robinson CV (2008) say to identify such endoctrine disrupting chemicals Micelles protect membrane complexes from solution to vacu- by their affinity for membrane-receptors. um. Science 321, 243–246. 5. Boonyaratanakornkit V, Bi Y, Rudd M and Edwards DP (2008) The role and mechanism of activation Acknowledgements of extra-nuclear signaling pathways in regulating gene tran- scription and cell cycle progression. Steroids 73, 922–928. This work was supported by JSPS KAKENHI Grant 6. Chien CH, Lai JN, Liao CF, Wang OY, Lu LM, Huang MI, Number 25440154 (to TT) and a grant from The Lee WF, Shie MC and Chien EJ (2009) acts as progesterone antagonist of non-genomic responses but inhib- Naito Foundation (to TT), Japanese environmental its phytohemagglutinin-induced proliferation in human T agency (ExTEND2005 FS to TT) and a SUNBOR cells. Hum Reprod 24, 1968–1975. grant from Suntory Institute For Bioorganic Re- 7. Cregg JM, Vedvick TS and Raschke WC (1993) Recent ad- search (to TT). The funders had no role in study vances in the expression of foreign in Pichia pastoris. Biotechnology (N Y) 11, 905–910. design, data collection and analysis, decision to pub- 8. Dressing GE and Thomas P (2007) Identification of mem- lish, or preparation of the manuscript. 58 T. Oshima et al.

brane progestin receptors in human breast cancer cell lines membrane progestin receptor (mPR) beta, are expressed in and biopsies and their potential involvement in breast cancer. the rainbow trout oocyte and exhibit similar ovarian expres- Steroids 72, 111–116. sion patterns. Reprod Biol Endocrinol 4, 6. 9. Dressing GE, Pang Y, Dong J and Thomas P (2010) Proges- 22. Nagahama Y and Yamashita M (2008) Regulation of oocyte tin signaling through mPRalpha in Atlantic croaker granulosa/ maturation in fish. Dev Growth Differ 50 Suppl 1, S195– theca cell cocultures and its involvement in progestin inhibi- 219. tion of apoptosis. Endocrinology 151, 5916–5926. 23. Ndiaye K, Poole DH, Walusimbi S, Cannon MJ, Toyokawa K, 10. Dressing GE, Alyea R, Pang Y and Thomas P (2012) Mem- Maalouf SW, Dong J, Thomas P and Pate JL (2012) Proges- brane progesterone receptors (mPRs) mediate progestin in- terone effects on lymphocytes may be mediated by mem- duced antimorbidity in breast cancer cells and are expressed brane progesterone receptors. J Reprod Immunol 95, 15–26. in human breast tumors. Horm Cancer 3, 101–112. 24. Pang Y, Dong J and Thomas P (2013) Characterization, 11. Dressing GE, Goldberg JE, Charles NJ, Schwertfeger KL and Neurosteroid Binding and Brain Distribution of Human Mem- Lange CA (2011) Membrane progesterone receptor expres- brane Progesterone Receptors delta and {epsilon} (mPRdelta sion in mammalian tissues: a review of regulation and physi- and mPR{epsilon}) and mPRdelta Involvement in Neuros- ological implications. Steroids 76, 11–17. teroid Inhibition of Apoptosis. Endocrinology 154, 283–295. 12. Guennoun R, Meffre D, Labombarda F, Gonzalez SL, 25. Piccinni MP, Scaletti C, Maggi E and Romagnani S (2000) Deniselle MC, Stein DG, De Nicola AF and Schumacher M Role of hormone-controlled Th1- and Th2-type cytokines in (2008) The membrane-associated progesterone-binding pro- successful pregnancy. J Neuroimmunol 109, 30–33. tein 25-Dx: expression, cellular localization and up-regula- 26. Rae MT, Menzies GS and Bramley TA (1998) Bovine ovari- tion after brain and spinal cord injuries. Brain Res Rev 57, an non-genomic progesterone binding sites: presence in fol- 493–505. licular and luteal cell membranes. J Endocrinol 159, 13. Hanna RN and Zhu Y (2009) Expression of membrane pro- 413–427. gestin receptors in zebrafish (Danio rerio) oocytes, testis and 27. Scarpin KM, Graham JD, Mote PA and Clarke CL (2009) pituitary. Gen Comp Endocrinol 161, 153–157. Progesterone action in human tissues: regulation by proges- 14. Hino T, Arakawa T, Iwanari H, Yurugi-Kobayashi T, Ikeda- terone receptor (PR) isoform expression, nuclear positioning Suno C, Nakada-Nakura Y, Kusano-Arai O, Weyand S, and coregulator expression. Nucl Recept Signal 7, e009. Shimamura T, Nomura N, Cameron AD, Kobayashi T, 28. Scorer CA, Buckholz RG, Clare JJ and Romanos MA (1993) Hamakubo T, Iwata S and Murata T (2012) G-protein-cou- The intracellular production and secretion of HIV-1 envelope pled receptor inactivation by an allosteric inverse-agonist an- protein in the methylotrophic yeast Pichia pastoris. Gene tibody. Nature 482, 237–240. 136, 111–119. 15. Jalabert B and Finet B (1986) Regulation of oocyte matura- 29. Scott AP, Sheldrick EL and Flint AP (1982) Measurement of tion in the rainbow trout,Salmo gairdneri: role of cyclic AMP 17 alpha, 20 beta-dihydroxy-4-pregnen-3-one in plasma of in the mechanism of action of the maturation inducing ste- trout (Salmo gairdneri Richardson): seasonal changes and re- roid (MIS), 17alpha-hydroxy, 20beta-dihydroprogesterone. sponse to salmon pituitary extract. Gen Comp Endocrinol 46, Fish Physiol Biochem 2, 65–74. 444–451. 16. Karteris E, Zervou S, Pang Y, Dong J, Hillhouse EW, Randeva 30. Shimamura T, Shiroishi M, Weyand S, Tsujimoto H, Winter G, HS and Thomas P (2006) Progesterone signaling in human Katritch V, Abagyan R, Cherezov V, Liu W, Han GW, myometrium through two novel membrane G protein-coupled Kobayashi T, Stevens RC and Iwata S (2011) Structure of receptors: potential role in functional progesterone withdraw- the human histamine H1 receptor complex with doxepin. Na- al at term. Mol Endocrinol 20, 1519–1534. ture 475, 65–70. 17. Kelder J, Azevedo R, Pang Y, de Vlieg J, Dong J and 31. Smith JL, Kupchak BR, Garitaonandia I, Hoang LK, Maina Thomas P (2010) Comparison between steroid binding to AS, Regalla LM and Lyons TJ (2008) Heterologous expres- membrane progesterone receptor alpha (mPRalpha) and to sion of human mPRalpha, mPRbeta and mPRgamma in yeast nuclear progesterone receptor: correlation with physicochemi- confirms their ability to function as membrane progesterone cal properties assessed by comparative molecular field analy- receptors. Steroids 73, 1160–1173. sis and identification of mPRalpha-specific agonists. Steroids 32. Stanczyk FZ (2003) All progestins are not created equal. Ste- 75, 314–322. roids 68, 879–890. 18. Kohli G, Clelland E and Peng C (2005) Potential targets of 33. Tang YT, Hu T, Arterburn M, Boyle B, Bright JM, Emtage transforming growth factor-beta1 during inhibition of oocyte PC and Funk WD (2005) PAQR proteins: a novel membrane maturation in zebrafish. Reprod Biol Endocrinol 3, 53. receptor family defined by an ancient 7-transmembrane pass 19. Lundstrom K, Wagner R, Reinhart C, Desmyter A, Cherouati motif. J Mol Evol 61, 372–380. N, Magnin T, Zeder-Lutz G, Courtot M, Prual C, Andre N, 34. Thomas P, Tubbs C and Garry VF (2009) Progestin functions Hassaine G, Michel H, Cambillau C and Pattus F (2006) in vertebrate gametes mediated by membrane progestin re- Structural genomics on membrane proteins: comparison of ceptors (mPRs): Identification of mPRalpha on human sperm more than 100 GPCRs in 3 expression systems. J Struct and its association with sperm motility. Steroids 74, 614–621. Funct Genomics 7, 77–91. 35. Thomas P, Pang Y, Zhu Y, Detweiler C and Doughty K (2004) 20. Meffre D, Labombarda F, Delespierre B, Chastre A, De Multiple rapid progestin actions and progestin membrane re- Nicola AF, Stein DG, Schumacher M and Guennoun R (2013) ceptor subtypes in fish. Steroids 69, 567–573. Distribution of membrane progesterone receptor alpha in the 36. Thomas P, Pang Y, Dong J, Groenen P, Kelder J, de Vlieg J, male mouse and rat brain and its regulation after traumatic Zhu Y and Tubbs C (2007) Steroid and G protein binding brain injury. Neuroscience 231C, 111–124. characteristics of the seatrout and human progestin mem- 21. Mourot B, Nguyen T, Fostier A and Bobe J (2006) Two un- brane receptor alpha subtypes and their evolutionary origins. related putative membrane-bound progestin receptors, proges- Endocrinology 148, 705–718. terone membrane receptor component 1 (PGMRC1) and 37. Tokumoto M, Nagahama Y, Thomas P and Tokumoto T (2006) Purification of recombinant mPRα 59

Cloning and identification of a membrane progestin receptor in Atlantic croaker (Micropogonias undulatus) gonads: role in goldfish ovaries and evidence it is an intermediary inoo- in gamete maturation. Gen Comp Endocrinol 165, 144–154. cyte meiotic maturation. Gen Comp Endocrinol 145, 101– 43. Zeder-Lutz G, Cherouati N, Reinhart C, Pattus F and Wagner 108. R (2006) Dot-blot immunodetection as a versatile and high- 38. Tokumoto T, Tokumoto M and Thomas P (2007) Interactions throughput assay to evaluate recombinant GPCRs produced of diethylstilbestrol (DES) and DES analogs with membrane in the yeast Pichia pastoris. Protein Expr Purif 50, 118–127. progestin receptor-alpha and the correlation with their nonge- 44. Zhu Y, Bond J and Thomas P (2003) Identification, classifi- nomic progestin activities. Endocrinology 148, 3459–3467. cation, and partial characterization of genes in humans and 39. Tokumoto T, Tokumoto M, Horiguchi R, Ishikawa K and other vertebrates homologous to a fish membrane progestin Nagahama Y (2004) Diethylstilbestrol induces fish oocyte receptor. Proc Natl Acad Sci USA 100, 2237–2242. maturation. Proc Natl Acad Sci USA 101, 3686–3690. 45. Zhu Y, Rice CD, Pang Y, Pace M and Thomas P (2003) 40. Tokumoto T, Tokumoto M, Oshima T, Shimizuguchi K, Cloning, expression, and characterization of a membrane pro- Fukuda T, Sugita E, Suzuki M, Sakae YT, Akiyama Y, gestin receptor and evidence it is an intermediary in meiotic Nakayama R, Roy SR, Saydur Rahman M, Pang Y, Dong J maturation of fish oocytes. Proc Natl Acad Sci USA 100, and Thomas P (2012) Characterization of multiple membrane 2231–2236. progestin receptor (mPR) subtypes from the goldfish ovary 46. Zhu Y, Hanna RN, Schaaf MJ, Spaink HP and Thomas P and their roles in the induction of oocyte maturation. Gen (2008) Candidates for membrane progestin receptors—past Comp Endocrinol 177, 168–176. approaches and future challenges. Comp Biochem Physiol C 41. Tubbs C and Thomas P (2009) Progestin signaling through Toxicol Pharmacol 148, 381–389. an olfactory G protein and membrane progestin receptor- 47. Zuo L, Li W and You S (2010) Progesterone reverses the alpha in Atlantic croaker sperm: potential role in induction of mesenchymal phenotypes of basal phenotype breast cancer sperm hypermotility. Endocrinology 150, 473–484. cells via a membrane progesterone receptor mediated path- 42. Tubbs C, Pace M and Thomas P (2010) Expression and go- way. Breast Cancer Res 12, R34. nadotropin regulation of membrane progestin receptor alpha